RFC 2274
User-based Security Model (USM) for version 3 of the Simple Network Management Protocol (SNMPv3)
RFC 2274: User-based Security Model (USM) for SNMPv3 🎀
okay bestie, this is the spec for USM, the security model that keeps SNMPv3 messages safe 🔐✨ here's the title block so you know who's behind it:
- Series: Network Working Group 🌸
- Authors: U. Blumenthal and B. Wijnen, both at IBM T. J. Watson Research 💅
- Category: Standards Track 🔥
- Date: January 1998 📅
- Obsoletes: RFC 2264 🦋
Status of this Memo ✨
okay so this document is an Internet standards track protocol for the Internet community 🌈 it also asks people to talk about it and send over ideas for making it better 💬 if you want to know where this protocol stands officially, check the current "Internet Official Protocol Standards" (STD 1) 📚 and the memo can be shared as widely as you like, no limits 🫶
Copyright Notice 🌷
Copyright (C) The Internet Society (1998). All Rights Reserved.
IANA Note 🧚
so there was a clerical mix-up when the snmpModules got their numbers in this memo 😭 this RFC fixes that with the corrected number assignment for the protocol 💖 and yep, it obsoletes RFC 2264 🫶
Abstract 🎀
this doc describes the User-based Security Model (USM) for SNMP version 3, built for the SNMP architecture [RFC2271] ✨ it lays out the Elements of Procedure for SNMP message level security 🔐 and it also comes with a MIB so you can remotely monitor and manage the configuration parameters for this Security Model 🦋
Table of Contents 🌸
1. Introduction 3
1.1. Threats 4
1.2. Goals and Constraints 5
1.3. Security Services 6
1.4. Module Organization 7
1.4.1. Timeliness Module 7
1.4.2. Authentication Protocol 8
1.4.3. Privacy Protocol 8
1.5. Protection against Message Replay, Delay and Redirection 8
1.5.1. Authoritative SNMP engine 8
1.5.2. Mechanisms 9
1.6. Abstract Service Interfaces. 10
1.6.1. User-based Security Model Primitives for Authentication 11
1.6.2. User-based Security Model Primitives for Privacy 11
2. Elements of the Model 12
2.1. User-based Security Model Users 12
2.2. Replay Protection 13
2.2.1. msgAuthoritativeEngineID 13
2.2.2. msgAuthoritativeEngineBoots and msgAuthoritativeEngineTime 14
2.2.3. Time Window 15
2.3. Time Synchronization 15
2.4. SNMP Messages Using this Security Model 16
2.5. Services provided by the User-based Security Model 17
2.5.1. Services for Generating an Outgoing SNMP Message 17
2.5.2. Services for Processing an Incoming SNMP Message 19
2.6. Key Localization Algorithm. 21
3. Elements of Procedure 21
3.1. Generating an Outgoing SNMP Message 22
3.2. Processing an Incoming SNMP Message 25
4. Discovery 30
5. Definitions 31
6. HMAC-MD5-96 Authentication Protocol 45
6.1. Mechanisms 45
6.1.1. Digest Authentication Mechanism 46
6.2. Elements of the Digest Authentication Protocol 46
6.2.1. Users 46
6.2.2. msgAuthoritativeEngineID 47
6.2.3. SNMP Messages Using this Authentication Protocol 47
6.2.4. Services provided by the HMAC-MD5-96 Authentication Module 47
6.2.4.1. Services for Generating an Outgoing SNMP Message 47
6.2.4.2. Services for Processing an Incoming SNMP Message 48
6.3. Elements of Procedure 49
6.3.1. Processing an Outgoing Message 49
6.3.2. Processing an Incoming Message 50
7. HMAC-SHA-96 Authentication Protocol 51
7.1. Mechanisms 51
7.1.1. Digest Authentication Mechanism 51
7.2. Elements of the HMAC-SHA-96 Authentication Protocol 52
7.2.1. Users 52
7.2.2. msgAuthoritativeEngineID 52
7.2.3. SNMP Messages Using this Authentication Protocol 53
7.2.4. Services provided by the HMAC-SHA-96 Authentication Module 53
7.2.4.1. Services for Generating an Outgoing SNMP Message 53
7.2.4.2. Services for Processing an Incoming SNMP Message 54
7.3. Elements of Procedure 54
7.3.1. Processing an Outgoing Message 55
7.3.2. Processing an Incoming Message 55
8. CBC-DES Symmetric Encryption Protocol 56
8.1. Mechanisms 56
8.1.1. Symmetric Encryption Protocol 57
8.1.1.1. DES key and Initialization Vector. 57
8.1.1.2. Data Encryption. 58
8.1.1.3. Data Decryption 59
8.2. Elements of the DES Privacy Protocol 59
8.2.1. Users 59
8.2.2. msgAuthoritativeEngineID 59
8.2.3. SNMP Messages Using this Privacy Protocol 60
8.2.4. Services provided by the DES Privacy Module 60
8.2.4.1. Services for Encrypting Outgoing Data 60
8.2.4.2. Services for Decrypting Incoming Data 61
8.3. Elements of Procedure. 61
8.3.1. Processing an Outgoing Message 61
8.3.2. Processing an Incoming Message 62
9. Intellectual Property 62
10. Acknowledgements 63
11. Security Considerations 64
11.1. Recommended Practices 64
11.2. Defining Users 66
11.3. Conformance 67
12. References 67
13. Editors' Addresses 69
A.1. SNMP engine Installation Parameters 70
A.2. Password to Key Algorithm 71
A.2.1. Password to Key Sample Code for MD5 71
A.2.2. Password to Key Sample Code for SHA 72
A.3. Password to Key Sample Results 73
A.3.1. Password to Key Sample Results using MD5 73
A.3.2. Password to Key Sample Results using SHA 74
A.4. Sample encoding of msgSecurityParameters 74
B. Full Copyright Statement 761. Introduction ✨
okay so the Architecture for describing Internet Management Frameworks [RFC2271] says an SNMP engine is made up of four things 👑
1) a Dispatcher
2) a Message Processing Subsystem,
3) a Security Subsystem, and
4) an Access Control Subsystem.applications get to use the services these subsystems offer 💅
it's really important to get the SNMP architecture and its vocab down so you can see where the Security Model in this doc fits in and how it talks to the other subsystems 🧠 the reader is expected to already understand the SNMP architecture as defined in [RFC2271], so make sure you've read that first bestie 📖
this memo [RFC2274] describes the User-based Security Model as it's used inside the SNMP Architecture 🎀 the big idea is that we use the classic idea of a user (identified by a userName) and attach security information to that user 🫶
this memo also covers HMAC-MD5-96 and HMAC-SHA-96 as the authentication protocols, and CBC-DES as the privacy protocol 🔐 but USM isn't locked into these: it can use other protocols instead of them or alongside them 🌈 that's why HMAC-MD5-96, HMAC-SHA-96 and CBC-DES each get their own section, since they're self-contained and can be swapped out or supplemented later on 🦋
the key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be read as described in [RFC2119] 🍓
1.1. Threats 🎀
Okay so, a bunch of the classic network protocol threats totally apply to network management, which means they'd apply to any SNMP Security Model too 😭. Other threats just aren't relevant to network management at all. This section covers the principal threats, the secondary ones, and the ones that are lowkey less important. 💅
The principal threats this SNMP Security Model should protect against are:
- Modification of Information The modification threat is the danger that some unauthorized entity may alter in-transit SNMP messages generated on behalf of an authorized user in such a way as to effect unauthorized management operations, including falsifying the value of an object. 🙈
- Masquerade The masquerade threat is the danger that management operations not authorized for some user may be attempted by assuming the identity of another user that has the appropriate authorizations. 🎭
Two secondary threats are also identified. The Security Model defined in this memo provides limited protection against:
- Disclosure The disclosure threat is the danger of eavesdropping on the exchanges between managed agents and a management station. Protecting against this threat may be required as a matter of local policy. 👀
- Message Stream Modification The SNMP protocol is typically based upon a connection-less transport service which may operate over any sub-network service. The re-ordering, delay or replay of messages can and does occur through the natural operation of many such sub-network services. The message stream modification threat is the danger that messages may be maliciously re-ordered, delayed or replayed to an extent which is greater than can occur through the natural operation of a sub-network service, in order to effect unauthorized management operations. 🌀
There are at least two threats that an SNMP Security Model need not protect against. The security protocols defined in this memo do not provide protection against:
- Denial of Service This SNMP Security Model does not attempt to address the broad range of attacks by which service on behalf of authorized users is denied. Indeed, such denial-of-service attacks are in many cases indistinguishable from the type of network failures with which any viable network management protocol must cope as a matter of course. 🌧️
- Traffic Analysis This SNMP Security Model does not attempt to address traffic analysis attacks. Indeed, many traffic patterns are predictable (devices may be managed on a regular basis by a relatively small number of management applications), and therefore there is no significant advantage afforded by protecting against traffic analysis. 📊
1.2. Goals and Constraints 🎯
Based on everything we just covered about threats in the SNMP network management environment, the goals of this SNMP Security Model are as follows. 🧚
- Provide for verification that each received SNMP message has not been modified during its transmission through the network.
- Provide for verification of the identity of the user on whose behalf a received SNMP message claims to have been generated.
- Provide for detection of received SNMP messages, which request or contain management information, whose time of generation was not recent.
- Provide, when necessary, that the contents of each received SNMP message are protected from disclosure.
Besides the main goal of supporting secure network management, the design of this SNMP Security Model is also shaped by the following constraints. 🧱
- When the requirements of effective management in times of network stress are inconsistent with those of security, the design should prefer the former.
- Neither the security protocol nor its underlying security mechanisms should depend upon the ready availability of other network services (e.g., Network Time Protocol (NTP) or key management protocols).
- A security mechanism should entail no changes to the basic SNMP network management philosophy.
1.3. Security Services 🔐
The security services needed to support the goals of this SNMP Security Model are as follows. ✨
- Data Integrity is the provision of the property that data has not been altered or destroyed in an unauthorized manner, nor have data sequences been altered to an extent greater than can occur non-maliciously.
- Data Origin Authentication is the provision of the property that the claimed identity of the user on whose behalf received data was originated is corroborated.
- Data Confidentiality is the provision of the property that information is not made available or disclosed to unauthorized individuals, entities, or processes.
- Message timeliness and limited replay protection is the provision of the property that a message whose generation time is outside of a specified time window is not accepted. Note that message reordering is not dealt with and can occur in normal conditions too.
For the protocols specified in this memo, it is not possible to assure the specific originator of a received SNMP message. Instead, it's the user on whose behalf the message was originated that gets authenticated. 🪪
For these protocols, it's not possible to get data integrity without data origin authentication, and it's not possible to get data origin authentication without data integrity. Also, there's no provision for data confidentiality without both data integrity and data origin authentication. They really come as a package deal 💕.
The security protocols used in this memo are considered acceptably secure at the time of writing. However, the procedures allow for new authentication and privacy methods to be specified at a future time if the need arises. 🔄
1.4. Module Organization 🧩
The security protocols defined in this memo are split into three different modules. Each one has its own responsibilities, and together they realize the goals and security services described above. It's giving teamwork 🫶.
- The authentication module MUST provide for:
- Data Integrity,
- Data Origin Authentication
- The timeliness module MUST provide for:
- Protection against message delay or replay (to an extent greater than can occur through normal operation)
- The privacy module MUST provide for:
- Protection against disclosure of the message payload.
The timeliness module is fixed for the User-based Security Model, while there is provision for multiple authentication and/or privacy modules. Each of those implements a specific authentication or privacy protocol, respectively. 🔀
1.4.1. Timeliness Module ⏰
Section 3 (Elements of Procedure) uses the timeliness values in an SNMP message to do timeliness checking. The timeliness check is only performed if authentication is applied to the message. Since the complete message is checked for integrity, we can assume that the timeliness values in a message that passes the authentication module are trustworthy. 🥹
1.4.2. Authentication Protocol 🪄
Section 6 describes the HMAC-MD5-96 authentication protocol, which is the first authentication protocol that MUST be supported with the User-based Security Model. Section 7 describes the HMAC-SHA-96 authentication protocol, which is another authentication protocol that SHOULD be supported with the User-based Security Model. In the future, additional or replacement authentication protocols may be defined as new needs arise. 🆕
The User-based Security Model prescribes that, if authentication is used, then the complete message is checked for integrity in the authentication module.
For a message to be authenticated, it needs to pass the authentication check by the authentication module and the timeliness check, which is a fixed part of this User-based Security Model. Both are required, no skipping 🚫.
1.4.3. Privacy Protocol 🤫
Section 8 describes the CBC-DES Symmetric Encryption Protocol, which is the first privacy protocol to be used with the User-based Security Model. In the future, additional or replacement privacy protocols may be defined as new needs arise. 🗝️
The User-based Security Model prescribes that the scopedPDU is protected from disclosure when a message is sent with privacy.
The User-based Security Model also prescribes that a message needs to be authenticated if privacy is in use. 🔒
1.5. Protection against Message Replay, Delay and Redirection 🛡️
1.5.1. Authoritative SNMP engine 👑
In order to protect against message replay, delay and redirection, one of the SNMP engines involved in each communication is designated to be the authoritative SNMP engine. When an SNMP message contains a payload which expects a response (for example a Get, GetNext, GetBulk, Set or Inform PDU), then the receiver of such messages is authoritative. When an SNMP message contains a payload which does not expect a response (for example an SNMPv2-Trap, Response or Report PDU), then the sender of such a message is authoritative. The whole thing hinges on who's the authority for each message type, so keep that straight, bestie 🫵.
1.5.2. Mechanisms ✨
okay so here's how the security model actually keeps the bad guys out 🕵️♀️ there are three main mechanisms, and they're lowkey genius:
- Timeliness checks 🕰️ To stop messages from being delayed or replayed (in a way that goes beyond normal network lag), every message carries a set of timeliness indicators from the authoritative SNMP engine. Receivers check those indicators to see if a message is recent. An SNMP engine MAY check them to make sure a message is at least as recent as the last one it got from that same source. A non-authoritative SNMP engine uses messages that pass authentication to move forward its idea of the remote authoritative engine's timeliness indicators.
An SNMP engine MUST also have a way to match incoming Responses with the Requests it's waiting on, and it MUST drop any Responses that don't match an outstanding Request. For example, a msgID can be stuck into every message to handle this.
These mechanisms catch authenticated messages that weren't generated recently. But here's the tea ☕: this protection against delay and replay doesn't imply or provide any protection against unauthorized deletion or suppression of messages. An SNMP engine also might not notice reordering if every message involved is sent inside the Time Window interval. Other mechanisms that live outside the security protocol can be used to catch reordered, deleted, or suppressed messages that contain Set operations (for example, the MIB variable snmpSetSerialNo [RFC1907]). 🔍
- Keeping engines apart 🚧 Messages sent to or from one authoritative SNMP engine can't be replayed to or pretend to be from another authoritative SNMP engine. How? Every message includes an identifier that's unique to the authoritative SNMP engine that's either sending it or meant to receive it.
Now, a Report, Response, or Trap sent from an authoritative SNMP engine to one non-authoritative SNMP engine can get replayed to another non-authoritative SNMP engine. That second engine might (if it knows the same userName with the same secrets at the authoritative engine) update its idea of the authoritative engine's timeliness indicators, but that's not considered a threat 🙅♀️
In this situation, a Report or Response gets discarded by the Message Processing Model, since there shouldn't be an outstanding Request for it. A Trap might get accepted, though. Again, not a threat, because the communication was authenticated and timely. It's basically as if the authoritative engine had been configured to start sending Traps to the second engine, which can theoretically happen without the second engine knowing about it anyway. Anyway, the second SNMP engine might not expect this Trap, but it's allowed to see the management information inside it. 👀
- Spotting messages that aren't fresh 🥀 A set of time indicators gets included in the message, showing when it was generated. Messages without recent time indicators are not considered authentic. On top of that, an SNMP engine MUST drop any Responses that don't match an outstanding request. This part, however, is the responsibility of the Message Processing Model. 💅
This memo lets the same user be defined on multiple SNMP engines. 🧑🤝🧑
Each SNMP engine keeps a value called snmpEngineID, which uniquely identifies that engine. This value is included in every message sent to or from the engine when it's the authoritative one (see section 1.5.1). When a message comes in, an authoritative SNMP engine checks this value to make sure it's the intended recipient. A non-authoritative SNMP engine uses the value to make sure the message gets processed with the correct state information. 🎯
Each SNMP engine also keeps two values, snmpEngineBoots and snmpEngineTime. Together they give an indication of time at that engine. Both get included in an authenticated message sent to or received from that engine. When one arrives, the values are checked to make sure the timeliness value is within a Time Window of the current time. The Time Window is an administrative upper bound on how long protocol messages are allowed to take to arrive. ⏳
So, for an SNMP engine to generate a message that an authoritative SNMP engine will accept as authentic, and to verify that a message from that authoritative engine is authentic, the engine must first achieve timeliness synchronization with the authoritative SNMP engine. See section 2.3. 🫶
1.6. Abstract Service Interfaces ✨
Abstract service interfaces describe the conceptual interfaces between the different subsystems inside an SNMP entity. There's also a set of abstract service interfaces defined within the User-based Security Model (USM), which describe the conceptual interfaces between the generic USM services and the self-contained authentication and privacy services. 🧩
These abstract service interfaces are defined by a set of primitives. The primitives describe the services provided and the abstract data elements that must be passed when the services get invoked. This section lists the primitives defined for the User-based Security Model. 📋
1.6.1. User-based Security Model Primitives for Authentication
The User-based Security Model has these internal primitives for passing data back and forth between the Security Model itself and the authentication service 🔐
statusInformation =
authenticateOutgoingMsg(
IN authKey -- secret key for authentication
IN wholeMsg -- unauthenticated complete message
OUT authenticatedWholeMsg -- complete authenticated message
)
statusInformation =
authenticateIncomingMsg(
IN authKey -- secret key for authentication
IN authParameters -- as received on the wire
IN wholeMsg -- as received on the wire
OUT authenticatedWholeMsg -- complete authenticated message
)1.6.2. User-based Security Model Primitives for Privacy
The User-based Security Model has these internal primitives for passing data back and forth between the Security Model itself and the privacy service 🙈
statusInformation =
encryptData(
IN encryptKey -- secret key for encryption
IN dataToEncrypt -- data to encrypt (scopedPDU)
OUT encryptedData -- encrypted data (encryptedPDU)
OUT privParameters -- filled in by service provider
)
statusInformation =
decryptData(
IN decryptKey -- secret key for decrypting
IN privParameters -- as received on the wire
IN encryptedData -- encrypted data (encryptedPDU)
OUT decryptedData -- decrypted data (scopedPDU)
)2. Elements of the Model 🎀
This section has the definitions you need to actually realize the security model defined by this memo. 📖
2.1. User-based Security Model Users 👩💻
Management operations using this Security Model rely on a defined set of user identities. For any user on whose behalf management operations are authorized at a particular SNMP engine, that engine must have knowledge of that user. An SNMP engine that wants to talk to another SNMP engine must also have knowledge of a user known to that engine, including the applicable attributes of that user. 🫠
Here's what a user and its attributes look like 💅
userName: a string representing the name of the user.securityName: a human-readable string representing the user, in a format that's independent of the Security Model.authProtocol: says whether messages sent on behalf of this user can be authenticated, and if so, which authentication protocol is used. Two are defined in this memo: the HMAC-MD5-96 authentication protocol and the HMAC-SHA-96 authentication protocol.authKey: if messages sent for this user can be authenticated, this is the (private) authentication key used with the authentication protocol. A user's authentication key will normally be different at different authoritative SNMP engines. TheauthKeyisn't accessible via SNMP. The length requirements forauthKeyare defined by theauthProtocolin use. 🔑authKeyChangeandauthOwnKeyChange: the only way to remotely update the authentication key. They do this in a secure manner, so the update can be completed without needing privacy protection. 🔄privProtocol: says whether messages sent for this user can be protected from disclosure, and if so, which privacy protocol is used. One is defined in this memo: the CBC-DES Symmetric Encryption Protocol. 🔒privKey: if messages sent for this user can be en/decrypted, this is the (private) privacy key used with the privacy protocol. A user's privacy key will normally be different at different authoritative SNMP engines. TheprivKeyisn't accessible via SNMP. The length requirements forprivKeyare defined by theprivProtocolin use. 🗝️privKeyChangeandprivOwnKeyChange: the only way to remotely update the encryption key. They do this in a secure manner, so the update can be completed without needing privacy protection. 💌
2.2. Replay Protection 🛡️
Each SNMP engine maintains three objects:
snmpEngineID, which (at least within an administrative domain) uniquely and unambiguously identifies an SNMP engine.snmpEngineBoots, which counts how many times the SNMP engine has re-booted or re-initialized sincesnmpEngineIDwas last configured; and,snmpEngineTime, which is the number of seconds since thesnmpEngineBootscounter was last incremented.
Each SNMP engine is always authoritative for these objects in its own SNMP entity. It's the responsibility of a non-authoritative SNMP engine to synchronize with the authoritative SNMP engine, as appropriate. 🤝
An authoritative SNMP engine is required to maintain the values of its snmpEngineID and snmpEngineBoots in non-volatile storage. 💾
2.2.1. msgAuthoritativeEngineID
The msgAuthoritativeEngineID value in an authenticated message is what defeats attacks where messages from one SNMP engine to another get replayed to a different SNMP engine. It represents the snmpEngineID at the authoritative SNMP engine involved in the message exchange. 🚫
When an authoritative SNMP engine is first installed, it sets its local snmpEngineID value according to an enterprise-specific algorithm (see the definition of the Textual Convention for SnmpEngineID in the SNMP Architecture document [RFC2271]). 🌸
2.2.2. msgAuthoritativeEngineBoots and msgAuthoritativeEngineTime 🦋
Okay so these two values are literally the anti-replay bodyguards 💅 The msgAuthoritativeEngineBoots and msgAuthoritativeEngineTime values sitting inside an authenticated message exist to stop replay attacks, where someone sneaks in an old message after it's stopped being valid 🙈 They're basically a copy of the snmpEngineBoots and snmpEngineTime values at the authoritative SNMP engine in that exchange ✨
Here's the lowkey genius part: an SNMP engine doesn't need a non-volatile clock that keeps ticking while it's powered off 😭 Instead, every time an engine reboots, it grabs its snmpEngineBoots, bumps it up, saves it back to non-volatile storage, and resets snmpEngineTime to zero ⏰ Iconic.
When an engine is first installed, it sets both of its local values to zero 🍓 If snmpEngineTime ever hits its max value (2147483647), then snmpEngineBoots goes up as if the engine rebooted, snmpEngineTime resets to zero, and it starts counting again 🔁
Every time an authoritative engine reboots, any engine holding that authoritative engine's snmpEngineBoots and snmpEngineTime values needs to re-synchronize before it sends correctly authenticated messages to it (see Section 2.3 for (re-)synchronization procedures). Ngl, there's a twist 👀 The procedures still provide for a notification to be accepted as authentic by a receiving engine when it was sent by an authoritative engine that has rebooted since the receiver last (re-)synchronized 💖
If an authoritative engine is ever unable to determine its latest snmpEngineBoots value, then it must set its snmpEngineBoots value to 2147483647 😱
Whenever the local snmpEngineBoots value is 2147483647, it latches at that value, and an authenticated message always causes a notInTimeWindow authentication failure 🚫
To reset an engine whose snmpEngineBoots has reached 2147483647, manual intervention is required 🛠️ Someone has to physically visit the engine and reconfigure it, either with a new snmpEngineID value or with new secret values for the authentication and privacy protocols of all users known to that engine 🧑🔧 Even if an engine rebooted once a second, it would still take approximately 68 years to reach that max value of 2147483647 🌈
2.2.3. Time Window ⏳
The Time Window is the window of time in which a message generated on behalf of any user is valid 💕 This memo specifies that the same Time Window value, 150 seconds, is used for all users 🎀
2.3. Time Synchronization 🕰️
Time synchronization is what a non-authoritative SNMP engine needs before it can proceed with authentic communications ✨ It's considered done once that engine has obtained a local notion of the authoritative engine's snmpEngineBoots and snmpEngineTime values from the authoritative engine itself 💫 Those values must be (and remain) within the authoritative engine's Time Window, so the local copy has to be kept loosely synchronized with the values stored at the authoritative engine 🧚 On top of its local copy of snmpEngineBoots and snmpEngineTime, a non-authoritative engine must also keep one more local variable, latestReceivedEngineTime 👑 This records the highest snmpEngineTime it has received from the authoritative engine, and it's used to eliminate the possibility of replaying messages that would stop the non-authoritative engine's notion of snmpEngineTime from advancing 🙅♀️
A non-authoritative engine must keep local notions of these values for each authoritative engine it wishes to communicate with 📚 Since each authoritative engine is uniquely and unambiguously identified by its snmpEngineID, the non-authoritative engine can use that value as a key to cache its local notions 🗝️
Time synchronization occurs as part of the procedure for receiving an SNMP message (Section 3.2, step 7b) 🫶 So no explicit time synchronization procedure is required by a non-authoritative engine 🙌 Note, though, that whenever the local value of snmpEngineID is changed (e.g., through discovery), or when secure communications are first established with an authoritative engine, the local values of snmpEngineBoots and latestReceivedEngineTime should be set to zero 🔄 That makes time synchronization happen when the next authentic message is received 💜
2.4. SNMP Messages Using this Security Model 📨
The syntax of an SNMP message using this Security Model follows the message format defined in the version-specific Message Processing Model document (for example [RFC2272]) 💌
The field msgSecurityParameters in SNMPv3 messages has a data type of OCTET STRING 🧾 Its value is the BER serialization of the following ASN.1 sequence:
USMSecurityParametersSyntax DEFINITIONS IMPLICIT TAGS ::= BEGIN
UsmSecurityParameters ::=
SEQUENCE {
-- global User-based security parameters
msgAuthoritativeEngineID OCTET STRING,
msgAuthoritativeEngineBoots INTEGER (0..2147483647),
msgAuthoritativeEngineTime INTEGER (0..2147483647),
msgUserName OCTET STRING (SIZE(1..32)),
-- authentication protocol specific parameters
msgAuthenticationParameters OCTET STRING,
-- privacy protocol specific parameters
msgPrivacyParameters OCTET STRING
}
ENDThe fields of this sequence are:
msgAuthoritativeEngineIDspecifies thesnmpEngineIDof the authoritative SNMP engine involved in the exchange of the message. 🆔msgAuthoritativeEngineBootsspecifies thesnmpEngineBootsvalue at the authoritative SNMP engine involved in the exchange of the message. 🔢msgAuthoritativeEngineTimespecifies thesnmpEngineTimevalue at the authoritative SNMP engine involved in the exchange of the message. ⏱️msgUserNamespecifies the user (principal) on whose behalf the message is being exchanged. 🧑🎤msgAuthenticationParametersare defined by the authentication protocol in use for the message, as defined by theusmUserAuthProtocolcolumn in the user's entry in theusmUserTable. 🔐msgPrivacyParametersare defined by the privacy protocol in use for the message, as defined by theusmUserPrivProtocolcolumn in the user's entry in theusmUserTable). 🕵️♀️
See appendix A.4 for an example of the BER encoding of field msgSecurityParameters 📖
2.5. Services provided by the User-based Security Model 🎁
This section describes the services provided by the User-based Security Model, with their inputs and outputs 💖
The services are described as primitives of an abstract service interface, and the inputs and outputs are described as abstract data elements as they are passed in these abstract service primitives 🧚
2.5.1. Services for Generating an Outgoing SNMP Message ✨
okay so when the Message Processing (MP) Subsystem wants the User-based Security module to lock down an outgoing SNMP message, it must use the right service that the Security module offers. there are two of them, and they're both giving main character energy 💅
- a service to generate a Request message. the abstract service primitive looks like this:
statusInformation = -- success or errorIndication
generateRequestMsg(
IN messageProcessingModel -- typically, SNMP version
IN globalData -- message header, admin data
IN maxMessageSize -- of the sending SNMP entity
IN securityModel -- for the outgoing message
IN securityEngineID -- authoritative SNMP entity
IN securityName -- on behalf of this principal
IN securityLevel -- Level of Security requested
IN scopedPDU -- message (plaintext) payload
OUT securityParameters -- filled in by Security Module
OUT wholeMsg -- complete generated message
OUT wholeMsgLength -- length of generated message
)- a service to generate a Response message. the abstract service primitive looks like this:
statusInformation = -- success or errorIndication
generateResponseMsg(
IN messageProcessingModel -- typically, SNMP version
IN globalData -- message header, admin data
IN maxMessageSize -- of the sending SNMP entity
IN securityModel -- for the outgoing message
IN securityEngineID -- authoritative SNMP entity
IN securityName -- on behalf of this principal
IN securityLevel -- Level of Security requested
IN scopedPDU -- message (plaintext) payload
IN securityStateReference -- reference to security state
-- information from original
-- request
OUT securityParameters -- filled in by Security Module
OUT wholeMsg -- complete generated message
OUT wholeMsgLength -- length of generated message
)the abstract data elements passed as parameters in these primitives work like this 🫶
statusInformation: tells you whether encoding and securing the message worked. if it didn't, it tells you what went wrong.messageProcessingModel: the SNMP version number for the message being generated. the User-based Security module doesn't use this one.globalData: the message header, meaning its administrative info. the User-based Security module doesn't use this one either.maxMessageSize: the maximum message size, as it's included in the message. the User-based Security module doesn't use this one.securityParameters: the security parameters. the User-based Security module fills these in 🔐securityModel: the securityModel in use, which should be the User-based Security Model. the User-based Security module doesn't use this one.securityName: together with the snmpEngineID, this identifies a row in the usmUserTable that gets used to secure the message. the securityName has a format that doesn't depend on the Security Model. for a response, this parameter gets ignored and the value from the cache is used instead.securityLevel: the Level of Security. the User-based Security module uses it to decide whether the message needs to be protected from disclosure and whether it needs to be authenticated. for a response, this parameter gets ignored and the value from the cache is used instead.securityEngineID: the snmpEngineID of the authoritative SNMP engine that a Request message is headed to. for a response, it's implied to be the processing SNMP engine's own snmpEngineID, so if you do pass it in, it gets ignored 🙈scopedPDU: the message payload. it's opaque as far as the User-based Security Model is concerned, so it just rides along 🎀securityStateReference: a handle or reference to cachedSecurityData, used when securing an outgoing Response message. it's the exact same handle that the User-based Security module generated while it was processing the incoming Request this Response is answering 🦋wholeMsg: the fully encoded and secured message, ready to go out on the wire.wholeMsgLength: the length of the encoded and secured message (wholeMsg).
when the process is done, the User-based Security module hands back a statusInformation. if it went well, you get the completed message, with privacy and authentication applied if the specified securityLevel asked for them. if it didn't go well, you get an errorIndication instead 💀
2.5.2. Services for Processing an Incoming SNMP Message 🌸
when the Message Processing (MP) Subsystem asks the User-based Security module to check that an incoming message is secure, it must use the service that's provided for incoming messages. the abstract service primitive is:
statusInformation = -- errorIndication or success
-- error counter OID/value if error
processIncomingMsg(
IN messageProcessingModel -- typically, SNMP version
IN maxMessageSize -- of the sending SNMP entity
IN securityParameters -- for the received message
IN securityModel -- for the received message
IN securityLevel -- Level of Security
IN wholeMsg -- as received on the wire
IN wholeMsgLength -- length as received on the wire
OUT securityEngineID -- authoritative SNMP entity
OUT securityName -- identification of the principal
OUT scopedPDU, -- message (plaintext) payload
OUT maxSizeResponseScopedPDU -- maximum size of the Response PDU
OUT securityStateReference -- reference to security state
) -- information, needed for responsethe abstract data elements passed as parameters in this primitive are like so 💖
statusInformation: says whether the process worked or not. if it didn't, the statusInformation includes the OID and the value of the error counter that got incremented.messageProcessingModel: the SNMP version number, as received in the message. the User-based Security module doesn't use this one.maxMessageSize: the maximum message size, as it's included in the message. the User-based Security module uses this value to work out the maxSizeResponseScopedPDU 🧮securityParameters: the security parameters, exactly as they were received in the message.securityModel: the securityModel in use, which should be the User-based Security Model. the User-based Security module doesn't use this one.securityLevel: the Level of Security. the User-based Security module uses it to decide whether the message needs to be protected from disclosure and whether it needs to be authenticated.wholeMsg: the whole message, exactly as it came in.wholeMsgLength: the length of the message as it came in (wholeMsg).securityEngineID: the snmpEngineID that got pulled out of the msgAuthoritativeEngineID field. it's also the one used to look up the secrets in the usmUserTable 🔑securityName: the security name for the user on whose behalf the message was received. like before, its format doesn't depend on the Security Model.scopedPDU: the message payload. again, opaque to the User-based Security Model, so hands off 🙌maxSizeResponseScopedPDU: the largest scopedPDU that could fit in a possible Response message. the User-base Security module works this out from the mms (as received in the message) and the room the header needs for that Response message, including the securityParameters 📏securityStateReference: a handle or reference to cachedSecurityData, for when securing an outgoing Response message. when the Message Processing Subsystem asks the User-based Security module to generate a response to this incoming message, it must pass this handle back in 🦋
when the process finishes, the User-based Security module returns a statusInformation and, if everything went right, the extra data elements needed to keep processing the message. if it didn't, you get an errorIndication, possibly along with an OID and value pair for whichever error counter got incremented 😭
2.6. Key Localization Algorithm. 🌈
okay, this part is lowkey so clever. a localized key is a secret key shared between a user U and one authoritative SNMP engine E. even if a user only has one password (and so one key for the whole network), the actual secrets shared between that user and each authoritative SNMP engine are different. that's the magic of key localization [Localized-key] ✨
first up, if a user uses a password, that password gets turned into a key Ku, using one of the two algorithms described in Appendices A.2.1 and A.2.2.
to turn key Ku into a localized key Kul for user U at the authoritative SNMP engine E, you stick the snmpEngineID of that authoritative SNMP engine onto the key Ku, then stick the key Ku on again after it. so the snmpEngineID ends up wrapped in two copies of the user's key Ku, like a little burrito 🌯 then you run a secure hash function on it. which hash function you use depends on the authentication protocol defined for this user U at authoritative SNMP engine E. this document defines two authentication protocols, with their associated algorithms based on MD5 and SHA. the output of the hash function is the localized key Kul for user U at the authoritative SNMP engine E 💜
3. Elements of Procedure 🎀
this section describes the security related procedures that an SNMP engine follows when it processes SNMP messages according to the User-based Security Model. let's get into it, bestie 💅
3.1. Generating an Outgoing SNMP Message ✨
Okay so this section walks through what an SNMP engine does every time it builds a message with a management operation (a request, response, notification, or report) for a user at a particular securityLevel. Let's gooo 💅
- Check the state first 🔍
- a) If any
securityStateReferenceis passed (that's a Response message), the user info gets pulled fromcachedSecurityData. ThesecurityEngineIDand thesecurityLevelget extracted from there too. After that, thecachedSecurityDatacan be discarded 🗑️
- a) If any
Otherwise,
- b) based on the
securityName, the user's info at the destinationsnmpEngineID(the one named bysecurityEngineID) gets pulled from the Local Configuration Datastore (LCD,usmUserTable). If the user isn't in the LCD at all, an error indication (unknownSecurityName) gets returned to the calling module 🙈
- Privacy check 🔐 If the securityLevel says the message has to be protected from disclosure, but the user doesn't support both an authentication and a privacy protocol, then the message can't be sent. An error indication (
unsupportedSecurityLevel) gets returned to the calling module. Not it 💀 - Auth check 🛡️ If the securityLevel says the message has to be authenticated, but the user doesn't support an authentication protocol, then the message can't be sent either. An error indication (
unsupportedSecurityLevel) gets returned to the calling module. - Encryption time 🎀
- a) If the securityLevel says the message has to be protected from disclosure, the octet sequence that represents the serialized scopedPDU gets encrypted with the user's privacy protocol. To do that, we call the privacy module that implements the user's privacy protocol, using this abstract primitive:
statusInformation = -- success or failure
encryptData(
IN encryptKey -- user's localized privKey
IN dataToEncrypt -- serialized scopedPDU
OUT encryptedData -- serialized encryptedPDU
OUT privParameters -- serialized privacy parameters
)statusInformationtells us whether the encryption worked or not 🙌encryptKeyis the user's localized private privKey, the secret key the encryption algorithm can use 🔑dataToEncryptis the serialized scopedPDU, the data that's getting encrypted 📦encryptedDatais the encryptedPDU (the encrypted scopedPDU), encoded as an OCTET STRING 🫶privParametersare the privacy parameters, encoded as an OCTET STRING ✨
If the privacy module says failure, the message can't be sent and an error indication (encryptionError) gets returned to the calling module 😭
If the privacy module says success, the returned privParameters get put into the msgPrivacyParameters field of the securityParameters, and the encryptedPDU becomes the payload of the message being prepared 🎉
Otherwise,
- b) If the securityLevel says the message does not need to be protected from disclosure, then the NULL string gets encoded as an OCTET STRING and put into the
msgPrivacyParametersfield of the securityParameters. The plaintext scopedPDU becomes the payload of the message being prepared 🌸
- Engine ID 🆔 The snmpEngineID gets encoded as an OCTET STRING into the
msgAuthoritativeEngineIDfield of the securityParameters. Here's the tea: an empty (zero length) snmpEngineID is totally fine for a Request message, because the remote (authoritative) SNMP engine will send back a Report PDU with the proper snmpEngineID included in themsgAuthoritativeEngineIDin the securityParameters of that Report PDU 🧚 - Boots and time ⏰
- a) If the securityLevel says the message has to be authenticated, then the current values of snmpEngineBoots and snmpEngineTime for the snmpEngineID from the LCD are used.
Otherwise,
- b) If this is a Response message, then the current values of snmpEngineBoots and snmpEngineTime for the local snmpEngineID from the LCD are used.
Otherwise,
- c) If this is a Request message, then a zero value is used for both snmpEngineBoots and snmpEngineTime. That zero value gets used if snmpEngineID is empty 0️⃣
The values get encoded as INTEGERs into the msgAuthoritativeEngineBoots and msgAuthoritativeEngineTime fields of the securityParameters 💖
- Username 👤 The userName gets encoded as an OCTET STRING into the
msgUserNamefield of the securityParameters. - Auth stamp ✍️
- a) If the securityLevel says the message has to be authenticated, the message gets authenticated with the user's authentication protocol. We do that by calling the authentication module that implements the user's authentication protocol, using this abstract service primitive:
statusInformation =
authenticateOutgoingMsg(
IN authKey -- the user's localized authKey
IN wholeMsg -- unauthenticated message
OUT authenticatedWholeMsg -- authenticated complete message
)statusInformationtells us whether the authentication worked 💫authKeyis the user's localized private authKey, the secret key the authentication algorithm can use 🗝️wholeMsgis the complete serialized message that needs authenticating 📨authenticatedWholeMsgis the same as the input given to the authenticateOutgoingMsg service, but withmsgAuthenticationParametersproperly filled in ✅
If the authentication module says failure, the message can't be sent and an error indication (authenticationFailure) gets returned to the calling module 💀
If the authentication module says success, the msgAuthenticationParameters field gets put into the securityParameters, and the authenticatedWholeMsg represents the serialized authenticated message being prepared 🦋
Otherwise,
- b) If the securityLevel says the message does not need to be authenticated, then the NULL string gets encoded as an OCTET STRING into the
msgAuthenticationParametersfield of the securityParameters. The wholeMsg is now serialized and represents the unauthenticated message being prepared 🌈
- Send it off 🚀 The completed message, along with its length, gets returned to the calling module with the statusInformation set to success.
3.2. Processing an Incoming SNMP Message 🦋
This section is the other side of the coin: what an SNMP engine does whenever it receives a message with a management operation for a user at a particular securityLevel. Pay attention, bestie 👀
To keep the elements of procedure simple, the release of state information isn't always spelled out. The general rule is: if state information exists when a message gets discarded, the state information should also be released 🧹 Also, when an error indication comes back with an OID and value for an incremented counter, the available information (like securityStateReference) must be passed back to the caller so it can generate a Report PDU 📣
- Parse check 🧐 If the received securityParameters isn't the serialization (following the conventions of [RFC1906]) of an OCTET STRING formatted according to the UsmSecurityParameters defined in section 2.4, then the snmpInASNParseErrs counter [RFC1907] gets incremented, and an error indication (
parseError) gets returned to the calling module. Note that we return without the OID and value of the incremented counter, because there isn't enough info in this case to generate a Report PDU 😭 - Unpack the goods 📦 The values of the security parameter fields get extracted from the securityParameters. The securityEngineID to be returned to the caller is the value of the
msgAuthoritativeEngineIDfield. The cachedSecurityData gets prepared, and a securityStateReference gets prepared to reference this data. The values to be cached are:msgUserNamesecurityEngineIDsecurityLevel
- Unknown engine? 👽 If the value of the
msgAuthoritativeEngineIDfield in the securityParameters is unknown, then:- a) a non-authoritative SNMP engine that performs discovery may optionally create a new entry in its Local Configuration Datastore (LCD) and keep processing;
or
- b) the
usmStatsUnknownEngineIDscounter gets incremented, and an error indication (unknownEngineID) gets returned to the calling module, together with the OID and value of the incremented counter 🚫
- User lookup 🔎 Info about the values of the
msgUserNameandmsgAuthoritativeEngineIDfields gets pulled from the Local Configuration Datastore (LCD,usmUserTable). If there's no info for the user, theusmStatsUnknownUserNamescounter gets incremented, and an error indication (unknownSecurityName) gets returned to the calling module, together with the OID and value of the incremented counter 🙅♀️ - Level check 🎚️ If the user info says the user doesn't support the securityLevel the caller asked for, then the
usmStatsUnsupportedSecLevelscounter gets incremented, and an error indication (unsupportedSecurityLevel) gets returned to the calling module, together with the OID and value of the incremented counter 💅 - Verify the signature 🔏 If the securityLevel says the message has to be authenticated, then the message gets authenticated with the user's authentication protocol. We call the authentication module that implements that protocol, using this abstract service primitive:
statusInformation = -- success or failure
authenticateIncomingMsg(
IN authKey -- the user's localized authKey
IN authParameters -- as received on the wire
IN wholeMsg -- as received on the wire
OUT authenticatedWholeMsg -- checked for authentication
)statusInformationtells us whether authentication worked or not ✨authKeyis the user's localized private authKey, the secret key the authentication algorithm can use 🗝️wholeMsgis the complete serialized message that needs authenticating 📨authenticatedWholeMsgis the same as the input given to the authenticateIncomingMsg service, but after authentication has been checked 🫶
If the authentication module says failure, the message can't be trusted, so the usmStatsWrongDigests counter gets incremented, and an error indication (authenticationFailure) gets returned to the calling module, together with the OID and value of the incremented counter 😱
If the authentication module says success, the message is authentic and can be trusted, so processing keeps going 🎀
- If the securityLevel says the message is authenticated, the engine pulls out the local snmpEngineBoots and snmpEngineTime that match the msgAuthoritativeEngineID from the Local Configuration Datastore (LCD) 🔐
- a) We're the authoritative engine 👑 If the extracted msgAuthoritativeEngineID matches this engine's snmpEngineID, then the message is considered outside of the Time Window if ANY of these are true:
- the local snmpEngineBoots is
2147483647; - the msgAuthoritativeEngineBoots field differs from the local snmpEngineBoots; or,
- the msgAuthoritativeEngineTime field differs from the local snmpEngineTime by more than +/- 150 seconds.
- the local snmpEngineBoots is
- a) We're the authoritative engine 👑 If the extracted msgAuthoritativeEngineID matches this engine's snmpEngineID, then the message is considered outside of the Time Window if ANY of these are true:
If the message is outside the Time Window, the usmStatsNotInTimeWindows counter goes up 📈, and an error indication (notInTimeWindow) plus the OID and value of that counter gets sent back to the calling module 🙈
- b) Someone else is the authoritative engine 🧚 If the extracted msgAuthoritativeEngineID does NOT match this engine's snmpEngineID, then:
- if at least one of these is true:
- the extracted msgAuthoritativeEngineBoots is greater than the local snmpEngineBoots; or,
- the extracted msgAuthoritativeEngineBoots equals the local snmpEngineBoots, and the extracted msgAuthoritativeEngineTime is greater than latestReceivedEngineTime,
- if at least one of these is true:
then the LCD entry for the extracted msgAuthoritativeEngineID gets updated 🎀: the local snmpEngineBoots is set to msgAuthoritativeEngineBoots, the local snmpEngineTime is set to msgAuthoritativeEngineTime, and latestReceivedEngineTime is set to msgAuthoritativeEngineTime.
- if ANY of these is true, the message is outside the Time Window 💀:
- the local snmpEngineBoots is
2147483647; - msgAuthoritativeEngineBoots is less than the local snmpEngineBoots; or,
- msgAuthoritativeEngineBoots equals the local snmpEngineBoots and msgAuthoritativeEngineTime is more than 150 seconds less than the local snmpEngineTime.
- the local snmpEngineBoots is
If it's outside the Time Window, an error indication (notInTimeWindow) goes back to the calling module.
Lowkey important: this means a message that's way too old (maybe replayed) got caught and is treated as not authentic 🚫
Also, msgAuthoritativeEngineBoots is allowed to be greater than the local snmpEngineBoots. This lets messages from an authoritative engine that rebooted since the receiver last (re-)synchronized still be accepted as authentic 🔄
Be aware though: this procedure does NOT do automatic time sync when the non-authoritative engine is really out of sync, meaning the authoritative engine is more than 150 seconds behind the non-authoritative one. That's a real out-of-sync situation and it doesn't fix itself 😭
- a) If the securityLevel says the message was protected from disclosure, the OCTET STRING holding the encryptedPDU gets decrypted with the user's privacy protocol to get back an unencrypted serialized scopedPDU 🔓 To do that, the engine calls the privacy module that implements the user's privacy protocol, using this abstract primitive:
statusInformation = -- success or failure
decryptData(
IN decryptKey -- the user's localized privKey
IN privParameters -- as received on the wire
IN encryptedData -- encryptedPDU as received
OUT decryptedData -- serialized decrypted scopedPDU
)Here's what each piece means 🧠
- statusInformation tells you if the decryption worked or not.
- decryptKey is the user's localized private privKey, the secret key the decryption algorithm can use.
- privParameters is the msgPrivacyParameters, encoded as an OCTET STRING.
- encryptedData is the encryptedPDU (the encrypted scopedPDU), encoded as an OCTET STRING.
- decryptedData is the serialized scopedPDU if decryption works.
If the privacy module returns failure, the message can't be processed 😩 so the usmStatsDecryptionErrors counter goes up, and an error indication (decryptionError) plus the OID and value of that counter goes back to the calling module.
If the privacy module returns success, the decrypted scopedPDU becomes the message payload that gets handed back to the calling module ✨
Otherwise, if the message wasn't encrypted:
b) The scopedPDU is assumed to already be plain text, and it becomes the message payload handed back to the calling module 📝
- The maxSizeResponseScopedPDU gets calculated 🧮 This is the biggest scopedPDU allowed for a possible Response message. It leaves room for a message header that allows the same securityLevel as the received Request.
- The securityName for the user gets retrieved from the usmUserTable 🙋♀️
- The security data gets cached as cachedSecurityData 💾 That way, a possible response to this message can use the very same authentication and privacy secrets, the same securityLevel, and the same msgAuthoritativeEngineID. Here's what gets saved:
msgUserName,
usmUserAuthProtocol, usmUserAuthKey
usmUserPrivProtocol, usmUserPrivKey
securityEngineID, securityLevel- The statusInformation is set to success, and the engine returns to the calling module, passing back the OUT parameters as specified in the processIncomingMsg primitive 🎉
4. Discovery 🔎
The User-based Security Model needs a discovery process to get enough info about other SNMP engines so they can talk to each other 🤝 Discovery means a non-authoritative SNMP engine has to learn the authoritative SNMP engine's snmpEngineID value before any communication can happen. One way to do this is to send a Request message with a securityLevel of noAuthNoPriv, a msgUserName of initial, a msgAuthoritativeEngineID with zero length, and an empty varBindList 🫶 The response will be a Report message with the authoritative engine's snmpEngineID as the value of the msgAuthoritativeEngineID field inside msgSecurityParameters. It also has a Report PDU with the usmStatsUnknownEngineIDs counter in the varBindList 💅
If authenticated communication is required, discovery should also set up time synchronization with the authoritative SNMP engine ⏰ To do that, send an authenticated Request message where msgAuthoritativeEngineID is set to the newly learned snmpEngineID, and msgAuthoritativeEngineBoots and msgAuthoritativeEngineTime are both set to zero. The response will be a Report message with the up-to-date snmpEngineBoots and snmpEngineTime values from the authoritative engine, given as the value of the msgAuthoritativeEngineBoots and msgAuthoritativeEngineTime fields respectively. It also includes the usmStatsNotInTimeWindows counter in the varBindList of the Report PDU. After that, time sync happens automatically through the procedures in section 3.2 step 7b. See also section 2.3 🦋
5. Definitions 💜
Here's the start of the SNMP-USER-BASED-SM-MIB module, with the imports, the module identity, and the first protocol identifiers 🌸
SNMP-USER-BASED-SM-MIB DEFINITIONS ::= BEGIN
IMPORTS
MODULE-IDENTITY, OBJECT-TYPE,
OBJECT-IDENTITY,
snmpModules, Counter32 FROM SNMPv2-SMI
TEXTUAL-CONVENTION, TestAndIncr,
RowStatus, RowPointer,
StorageType, AutonomousType FROM SNMPv2-TC
MODULE-COMPLIANCE, OBJECT-GROUP FROM SNMPv2-CONF
SnmpAdminString, SnmpEngineID,
snmpAuthProtocols, snmpPrivProtocols FROM SNMP-FRAMEWORK-MIB;
snmpUsmMIB MODULE-IDENTITY
LAST-UPDATED "9711200000Z" -- 20 Nov 1997, midnight
ORGANIZATION "SNMPv3 Working Group"
CONTACT-INFO "WG-email: snmpv3@tis.com
Subscribe: majordomo@tis.com
In msg body: subscribe snmpv3
Chair: Russ Mundy
Trusted Information Systems
postal: 3060 Washington Rd
Glenwood MD 21738
USA
email: mundy@tis.com
phone: +1-301-854-6889
Co-editor Uri Blumenthal
IBM T. J. Watson Research
postal: 30 Saw Mill River Pkwy,
Hawthorne, NY 10532
USA
email: uri@watson.ibm.com
phone: +1-914-784-7964
Co-editor: Bert Wijnen
IBM T. J. Watson Research
postal: Schagen 33
3461 GL Linschoten
Netherlands
email: wijnen@vnet.ibm.com
phone: +31-348-432-794
"
DESCRIPTION "The management information definitions for the
SNMP User-based Security Model.
"
::= { snmpModules 15 }
-- Administrative assignments ****************************************
usmMIBObjects OBJECT IDENTIFIER ::= { snmpUsmMIB 1 }
usmMIBConformance OBJECT IDENTIFIER ::= { snmpUsmMIB 2 }
-- Identification of Authentication and Privacy Protocols ************
usmNoAuthProtocol OBJECT-IDENTITY
STATUS current
DESCRIPTION "No Authentication Protocol."
::= { snmpAuthProtocols 1 }
usmHMACMD5AuthProtocol OBJECT-IDENTITY
STATUS current
DESCRIPTION "The HMAC-MD5-96 Digest Authentication Protocol."
REFERENCE "- H. Krawczyk, M. Bellare, R. Canetti HMAC:
Keyed-Hashing for Message Authentication,
RFC2104, Feb 1997.
- Rivest, R., Message Digest Algorithm MD5, RFC1321.
"
::= { snmpAuthProtocols 2 }Okay bestie, the MIB is getting extra dramatic now 💅 so grab your snacks 🍓
Here's the SHA one, which is basically the MD5 one's glow-up with its own key sizes and references 🦋
usmHMACSHAAuthProtocol OBJECT-IDENTITY
STATUS current
DESCRIPTION "The HMAC-SHA-96 Digest Authentication Protocol."
REFERENCE "- H. Krawczyk, M. Bellare, R. Canetti, HMAC:
Keyed-Hashing for Message Authentication,
RFC2104, Feb 1997.
- Secure Hash Algorithm. NIST FIPS 180-1.
"
::= { snmpAuthProtocols 3 }
usmNoPrivProtocol OBJECT-IDENTITY
STATUS current
DESCRIPTION "No Privacy Protocol."
::= { snmpPrivProtocols 1 }
usmDESPrivProtocol OBJECT-IDENTITY
STATUS current
DESCRIPTION "The CBC-DES Symmetric Encryption Protocol."
REFERENCE "- Data Encryption Standard, National Institute of
Standards and Technology. Federal Information
Processing Standard (FIPS) Publication 46-1.
Supersedes FIPS Publication 46,
(January, 1977; reaffirmed January, 1988).
- Data Encryption Algorithm, American National
Standards Institute. ANSI X3.92-1981,
(December, 1980).
- DES Modes of Operation, National Institute of
Standards and Technology. Federal Information
Processing Standard (FIPS) Publication 81,
(December, 1980).
- Data Encryption Algorithm - Modes of Operation,
American National Standards Institute.
ANSI X3.106-1983, (May 1983).
"
::= { snmpPrivProtocols 2 }So the privacy side gets its own little lineup: usmNoPrivProtocol is the "no privacy, bestie" option 🙈 and usmDESPrivProtocol is the CBC-DES one, with a whole stack of FIPS and ANSI references stapled on 📚
Now the textual conventions 🫶 The KeyChange one is lowkey the most intense thing in this whole document 😭
-- Textual Conventions ***********************************************
KeyChange ::= TEXTUAL-CONVENTION
STATUS current
DESCRIPTION
"Every definition of an object with this syntax must identify
a protocol P, a secret key K, and a hash algorithm H
that produces output of L octets.
The object's value is a manager-generated, partially-random
value which, when modified, causes the value of the secret
key K, to be modified via a one-way function.
The value of an instance of this object is the concatenation
of two components: first a 'random' component and then a
'delta' component.
The lengths of the random and delta components
are given by the corresponding value of the protocol P;
if P requires K to be a fixed length, the length of both the
random and delta components is that fixed length; if P
allows the length of K to be variable up to a particular
maximum length, the length of the random component is that
maximum length and the length of the delta component is any
length less than or equal to that maximum length.
For example, usmHMACMD5AuthProtocol requires K to be a fixed
length of 16 octets and L - of 16 octets.
usmHMACSHAAuthProtocol requires K to be a fixed length of
20 octets and L - of 20 octets. Other protocols may define
other sizes, as deemed appropriate.
When a requestor wants to change the old key K to a new
key keyNew on a remote entity, the 'random' component is
obtained from either a true random generator, or from a
pseudorandom generator, and the 'delta' component is
computed as follows:
- a temporary variable is initialized to the existing value
of K;
- if the length of the keyNew is greater than L octets,
then:
- the random component is appended to the value of the
temporary variable, and the result is input to the
the hash algorithm H to produce a digest value, and
the temporary variable is set to this digest value;
- the value of the temporary variable is XOR-ed with
the first (next) L-octets (16 octets in case of MD5)
of the keyNew to produce the first (next) L-octets
(16 octets in case of MD5) of the 'delta' component.
- the above two steps are repeated until the unused
portion of the delta component is L octets or less,
- the random component is appended to the value of the
temporary variable, and the result is input to the
hash algorithm H to produce a digest value;
- this digest value, truncated if necessary to be the same
length as the unused portion of the keyNew, is XOR-ed
with the unused portion of the keyNew to produce the
(final portion of the) 'delta' component.
For example, using MD5 as the hash algorithm H:
iterations = (lenOfDelta - 1)/16; /* integer division */
temp = keyOld;
for (i = 0; i < iterations; i++) {
temp = MD5 (temp || random);
delta[i*16 .. (i*16)+15] =
temp XOR keyNew[i*16 .. (i*16)+15];
}
temp = MD5 (temp || random);
delta[i*16 .. lenOfDelta-1] =
temp XOR keyNew[i*16 .. lenOfDelta-1];
The 'random' and 'delta' components are then concatenated as
described above, and the resulting octet string is sent to
the receipient as the new value of an instance of this
object.
At the receiver side, when an instance of this object is set
to a new value, then a new value of K is computed as follows:
- a temporary variable is initialized to the existing value
of K;
- if the length of the delta component is greater than L
octets, then:
- the random component is appended to the value of the
temporary variable, and the result is input to the
the hash algorithm H to produce a digest value, and
the temporary variable is set to this digest value;
- the value of the temporary variable is XOR-ed with
the first (next) L-octets (16 octets in case of MD5)
of the delta component to produce the first (next)
L-octets (16 octets in case of MD5) of the new value
of K.
- the above two steps are repeated until the unused
portion of the delta component is L octets or less,
- the random component is appended to the value of the
temporary variable, and the result is input to the
hash algorithm H to produce a digest value;
- this digest value, truncated if necessary to be the same
length as the unused portion of the delta component, is
XOR-ed with the unused portion of the delta component to
produce the (final portion of the) new value of K.
For example, using MD5 as the hash algorithm H:
iterations = (lenOfDelta - 1)/16; /* integer division */
temp = keyOld;
for (i = 0; i < iterations; i++) {
temp = MD5 (temp || random);
keyNew[i*16 .. (i*16)+15] =
temp XOR delta[i*16 .. (i*16)+15];
}
temp = MD5 (temp || random);
keyNew[i*16 .. lenOfDelta-1] =
temp XOR delta[i*16 .. lenOfDelta-1];
The value of an object with this syntax, whenever it is
retrieved by the management protocol, is always the zero
length string.
"
SYNTAX OCTET STRINGHere's the vibe in plain terms ✨ the requestor makes a random chunk, does a bunch of hashing and XOR-ing to build a delta, and sends random plus delta together. The receiver runs the same dance backwards to get the new key K 🔥 And the wildest bit is that reading the object always gives back the zero length string 💀 so nobody gets to peek at the secret, periodt 🤫
Next up are the stats counters for the User-based Security Model, which are basically the drama tracker of dropped packets 📊
-- Statistics for the User-based Security Model **********************
usmStats OBJECT IDENTIFIER ::= { usmMIBObjects 1 }
usmStatsUnsupportedSecLevels OBJECT-TYPE
SYNTAX Counter32
MAX-ACCESS read-only
STATUS current
DESCRIPTION "The total number of packets received by the SNMP
engine which were dropped because they requested a
securityLevel that was unknown to the SNMP engine
or otherwise unavailable.
"
::= { usmStats 1 }
usmStatsNotInTimeWindows OBJECT-TYPE
SYNTAX Counter32
MAX-ACCESS read-only
STATUS current
DESCRIPTION "The total number of packets received by the SNMP
engine which were dropped because they appeared
outside of the authoritative SNMP engine's window.
"
::= { usmStats 2 }
usmStatsUnknownUserNames OBJECT-TYPE
SYNTAX Counter32
MAX-ACCESS read-only
STATUS current
DESCRIPTION "The total number of packets received by the SNMP
engine which were dropped because they referenced a
user that was not known to the SNMP engine.
"
::= { usmStats 3 }
usmStatsUnknownEngineIDs OBJECT-TYPE
SYNTAX Counter32
MAX-ACCESS read-only
STATUS current
DESCRIPTION "The total number of packets received by the SNMP
engine which were dropped because they referenced an
snmpEngineID that was not known to the SNMP engine.
"
::= { usmStats 4 }
usmStatsWrongDigests OBJECT-TYPE
SYNTAX Counter32
MAX-ACCESS read-only
STATUS current
DESCRIPTION "The total number of packets received by the SNMP
engine which were dropped because they didn't
contain the expected digest value.
"
::= { usmStats 5 }
usmStatsDecryptionErrors OBJECT-TYPE
SYNTAX Counter32
MAX-ACCESS read-only
STATUS current
DESCRIPTION "The total number of packets received by the SNMP
engine which were dropped because they could not be
decrypted.
"
::= { usmStats 6 }So each counter is a little tally of packets that got bounced for a specific reason: an unsupported security level, a message outside the authoritative engine's window, an unknown user, an unknown snmpEngineID, a wrong digest, or a decryption failure 🙅♀️ Six counters, six different ways to get rejected, iconic 👑
Last stop for this part: the usmUser group and its table, where the users actually live 🎀
-- The usmUser Group ************************************************
usmUser OBJECT IDENTIFIER ::= { usmMIBObjects 2 }
usmUserSpinLock OBJECT-TYPE
SYNTAX TestAndIncr
MAX-ACCESS read-write
STATUS current
DESCRIPTION "An advisory lock used to allow several cooperating
Command Generator Applications to coordinate their
use of facilities to alter secrets in the
usmUserTable.
"
::= { usmUser 1 }
-- The table of valid users for the User-based Security Model ********
usmUserTable OBJECT-TYPE
SYNTAX SEQUENCE OF UsmUserEntry
MAX-ACCESS not-accessible
STATUS current
DESCRIPTION "The table of users configured in the SNMP engine's
Local Configuration Datastore (LCD)."
::= { usmUser 2 }
usmUserEntry OBJECT-TYPE
SYNTAX UsmUserEntry
MAX-ACCESS not-accessible
STATUS current
DESCRIPTION "A user configured in the SNMP engine's Local
Configuration Datastore (LCD) for the User-based
Security Model.
"
INDEX { usmUserEngineID,
usmUserName
}
::= { usmUserTable 1 }
UsmUserEntry ::= SEQUENCE
{
usmUserEngineID SnmpEngineID,
usmUserName SnmpAdminString,
usmUserSecurityName SnmpAdminString,
usmUserCloneFrom RowPointer,
usmUserAuthProtocol AutonomousType,
usmUserAuthKeyChange KeyChange,
usmUserOwnAuthKeyChange KeyChange,
usmUserPrivProtocol AutonomousType,
usmUserPrivKeyChange KeyChange,
usmUserOwnPrivKeyChange KeyChange,
usmUserPublic OCTET STRING,
usmUserStorageType StorageType,
usmUserStatus RowStatus
}The usmUserSpinLock is an advisory lock so several Command Generator Applications can coordinate when they change secrets in the usmUserTable 🔒 The table itself holds the users configured in the engine's Local Configuration Datastore (LCD), indexed by usmUserEngineID and usmUserName 💜 And each row, the UsmUserEntry, bundles up the engine ID, names, protocols, key-change objects, storage type and row status for one user 🧚
okay so this part is all about the columns of the user table 💖 Each one holds a different piece of a user's info: who they are, how they authenticate, and how their messages get protected 🔐
usmUserEngineID OBJECT-TYPE
SYNTAX SnmpEngineID
MAX-ACCESS not-accessible
STATUS current
DESCRIPTION "An SNMP engine's administratively-unique identifier.
In a simple agent, this value is always that agent's
own snmpEngineID value.
The value can also take the value of the snmpEngineID
of a remote SNMP engine with which this user can
communicate.
"
::= { usmUserEntry 1 }
usmUserName OBJECT-TYPE
SYNTAX SnmpAdminString (SIZE(1..32))
MAX-ACCESS not-accessible
STATUS current
DESCRIPTION "A human readable string representing the name of
the user.
This is the (User-based Security) Model dependent
security ID.
"
::= { usmUserEntry 2 }
usmUserSecurityName OBJECT-TYPE
SYNTAX SnmpAdminString
MAX-ACCESS read-only
STATUS current
DESCRIPTION "A human readable string representing the user in
Security Model independent format.
The default transformation of the User-based Security
Model dependent security ID to the securityName and
vice versa is the identity function so that the
securityName is the same as the userName.
"
::= { usmUserEntry 3 }
usmUserCloneFrom OBJECT-TYPE
SYNTAX RowPointer
MAX-ACCESS read-create
STATUS current
DESCRIPTION "A pointer to another conceptual row in this
usmUserTable. The user in this other conceptual
row is called the clone-from user.
When a new user is created (i.e., a new conceptual
row is instantiated in this table), the privacy and
authentication parameters of the new user are cloned
from its clone-from user.
The first time an instance of this object is set by
a management operation (either at or after its
instantiation), the cloning process is invoked.
Subsequent writes are successful but invoke no
action to be taken by the receiver.
The cloning process fails with an 'inconsistentName'
error if the conceptual row representing the
clone-from user is not in an active state when the
cloning process is invoked.
Cloning also causes the initial values of the secret
authentication key and the secret encryption key of
the new user to be set to the same value as the
corresponding secret of the clone-from user.
When this object is read, the ZeroDotZero OID
is returned.
"
::= { usmUserEntry 4 }the usmUserCloneFrom one is giving main character energy ✨ It's basically "copy my settings from this other user" 🦋 And the cloning only happens the first time someone sets it, so later writes are just... vibes, no action 🙈
usmUserAuthProtocol OBJECT-TYPE
SYNTAX AutonomousType
MAX-ACCESS read-create
STATUS current
DESCRIPTION "An indication of whether messages sent on behalf of
this user to/from the SNMP engine identified by
usmUserEngineID, can be authenticated, and if so,
the type of authentication protocol which is used.
An instance of this object is created concurrently
with the creation of any other object instance for
the same user (i.e., as part of the processing of
the set operation which creates the first object
instance in the same conceptual row). Once created,
the value of an instance of this object can not be
changed.
If a set operation tries to set a value for an unknown
or unsupported protocol, then a wrongValue error must
be returned.
"
DEFVAL { usmHMACMD5AuthProtocol }
::= { usmUserEntry 5 }
usmUserAuthKeyChange OBJECT-TYPE
SYNTAX KeyChange -- typically (SIZE (0..32))
MAX-ACCESS read-create
STATUS current
DESCRIPTION "An object, which when modified, causes the secret
authentication key used for messages sent on behalf
of this user to/from the SNMP engine identified by
usmUserEngineID, to be modified via a one-way
function.
The associated protocol is the usmUserAuthProtocol.
The associated secret key is the user's secret
authentication key (authKey). The associated hash
algorithm is the algorithm used by the user's
usmUserAuthProtocol.
When creating a new user, it is an 'inconsistentName'
error for a Set operation to refer to this object
unless it is previously or concurrently initialized
through a set operation on the corresponding value
of usmUserCloneFrom.
"
DEFVAL { ''H } -- the empty string
::= { usmUserEntry 6 }
usmUserOwnAuthKeyChange OBJECT-TYPE
SYNTAX KeyChange -- typically (SIZE (0..32))
MAX-ACCESS read-create
STATUS current
DESCRIPTION "Behaves exactly as usmUserAuthKeyChange, with one
notable difference: in order for the Set operation
to succeed, the usmUserName of the operation
requester must match the usmUserName that
indexes the row which is targeted by this
operation.
The idea here is that access to this column can be
public, since it will only allow a user to change
his own secret authentication key (authKey).
"
DEFVAL { ''H } -- the empty string
::= { usmUserEntry 7 }
usmUserPrivProtocol OBJECT-TYPE
SYNTAX AutonomousType
MAX-ACCESS read-create
STATUS current
DESCRIPTION "An indication of whether messages sent on behalf of
this user to/from the SNMP engine identified by
usmUserEngineID, can be protected from disclosure,
and if so, the type of privacy protocol which is used.
An instance of this object is created concurrently
with the creation of any other object instance for
the same user (i.e., as part of the processing of
the set operation which creates the first object
instance in the same conceptual row). Once created,
the value of an instance of this object can not be
changed.
If a set operation tries to set a value for an unknown
or unsupported protocol, then a wrongValue error must
be returned.
"
DEFVAL { usmNoPrivProtocol }
::= { usmUserEntry 8 }
usmUserPrivKeyChange OBJECT-TYPE
SYNTAX KeyChange -- typically (SIZE (0..32))
MAX-ACCESS read-create
STATUS current
DESCRIPTION "An object, which when modified, causes the secret
encryption key used for messages sent on behalf
of this user to/from the SNMP engine identified by
usmUserEngineID, to be modified via a one-way
function.
The associated protocol is the usmUserPrivProtocol.
The associated secret key is the user's secret
privacy key (privKey). The associated hash
algorithm is the algorithm used by the user's
usmUserAuthProtocol.
When creating a new user, it is an 'inconsistentName'
error for a set operation to refer to this object
unless it is previously or concurrently initialized
through a set operation on the corresponding value
of usmUserCloneFrom.
"
DEFVAL { ''H } -- the empty string
::= { usmUserEntry 9 }
usmUserOwnPrivKeyChange OBJECT-TYPE
SYNTAX KeyChange -- typically (SIZE (0..32))
MAX-ACCESS read-create
STATUS current
DESCRIPTION "Behaves exactly as usmUserPrivKeyChange, with one
notable difference: in order for the Set operation
to succeed, the usmUserName of the operation
requester must match the usmUserName that indexes
the row which is targeted by this operation.
The idea here is that access to this column can be
public, since it will only allow a user to change
his own secret privacy key (privKey).
"
DEFVAL { ''H } -- the empty string
::= { usmUserEntry 10 }
usmUserPublic OBJECT-TYPE
SYNTAX OCTET STRING (SIZE(0..32))
MAX-ACCESS read-create
STATUS current
DESCRIPTION "A publicly-readable value which is written as part
of the procedure for changing a user's secret
authentication and/or privacy key, and later read to
determine whether the change of the secret was
effected.
"
DEFVAL { ''H } -- the empty string
::= { usmUserEntry 11 }the auth and priv key pairs are literally a whole mood 🔥 The "OwnKeyChange" versions let a user change their own key, and the usmUserName has to match the row they're targeting 🫶
usmUserStorageType OBJECT-TYPE
SYNTAX StorageType
MAX-ACCESS read-create
STATUS current
DESCRIPTION "The storage type for this conceptual row.
Conceptual rows having the value 'permanent'
must allow write-access at a minimum to:
- usmUserAuthKeyChange, usmUserOwnAuthKeyChange
and usmUserPublic for a user who employs
authentication, and
- usmUserPrivKeyChange, usmUserOwnPrivKeyChange
and usmUserPublic for a user who employs
privacy.
Note that any user who employs authentication or
privacy must allow its secret(s) to be updated and
thus cannot be 'readOnly'.
"
DEFVAL { nonVolatile }
::= { usmUserEntry 12 }
usmUserStatus OBJECT-TYPE
SYNTAX RowStatus
MAX-ACCESS read-create
STATUS current
DESCRIPTION "The status of this conceptual row.
Until instances of all corresponding columns are
appropriately configured, the value of the
corresponding instance of the usmUserStatus column
is 'notReady'.
In particular, a newly created row cannot be made
active until the corresponding usmUserCloneFrom,
usmUserAuthKeyChange, usmUserOwnAuthKeyChange,
usmUserPrivKeyChange and usmUserOwnPrivKeyChange
have all been set.
The RowStatus TC [RFC1903] requires that this
DESCRIPTION clause states under which circumstances
other objects in this row can be modified:
The value of this object has no effect on whether
other objects in this conceptual row can be modified.
"
::= { usmUserEntry 13 }and now we're at the conformance bit 🎀 Two OBJECT IDENTIFIERs live under usmMIBConformance, one for the compliance statements and one for the groups 🧚
-- Conformance Information *******************************************
usmMIBCompliances OBJECT IDENTIFIER ::= { usmMIBConformance 1 }
usmMIBGroups OBJECT IDENTIFIER ::= { usmMIBConformance 2 }
-- Compliance statementsusmMIBCompliance MODULE-COMPLIANCE
STATUS current
DESCRIPTION "The compliance statement for SNMP engines which
implement the SNMP-USER-BASED-SM-MIB.
"
MODULE -- this module
MANDATORY-GROUPS { usmMIBBasicGroup }
OBJECT usmUserAuthProtocol
MIN-ACCESS read-only
DESCRIPTION "Write access is not required."
OBJECT usmUserPrivProtocol
MIN-ACCESS read-only
DESCRIPTION "Write access is not required."
::= { usmMIBCompliances 1 }
-- Units of compliance
usmMIBBasicGroup OBJECT-GROUP
OBJECTS {
usmStatsUnsupportedSecLevels,
usmStatsNotInTimeWindows,
usmStatsUnknownUserNames,
usmStatsUnknownEngineIDs,
usmStatsWrongDigests,
usmStatsDecryptionErrors,
usmUserSpinLock,
usmUserSecurityName,
usmUserCloneFrom,
usmUserAuthProtocol,
usmUserAuthKeyChange,
usmUserOwnAuthKeyChange,
usmUserPrivProtocol,
usmUserPrivKeyChange,
usmUserOwnPrivKeyChange,
usmUserPublic,
usmUserStorageType,
usmUserStatus
}
STATUS current
DESCRIPTION "A collection of objects providing for configuration
of an SNMP engine which implements the SNMP
User-based Security Model.
"
::= { usmMIBGroups 1 }
END6. HMAC-MD5-96 Authentication Protocol 🦋
Okay so this section is all about the HMAC-MD5-96 authentication protocol 🔐 It's the very first one defined for the User-based Security Model, so it's kind of the OG, no cap ✨ It uses MD5 [MD5] (the message digest algorithm) in HMAC mode [RFC2104], and then it chops the output down to 96 bits. Iconic trim, honestly 💅
This protocol is identified by usmHMACMD5AuthProtocol 🎀
Bestie, just so you know, other authentication protocols may be defined later on 🌈 They could replace this one or sit alongside it. Main character energy for future protocols!
6.1. Mechanisms 🌸
- In support of data integrity, a message digest algorithm is required 🔍 A digest gets calculated over an appropriate portion of an SNMP message, and then it's included as part of the message that gets sent to the recipient.
- In support of data origin authentication and data integrity, a secret value is prepended to the SNMP message before the digest is computed 🤫 The digest is then partially inserted into the SNMP message before it's sent, and the prepended value is never transmitted. That secret value is shared by all SNMP engines that are authorized to originate messages on behalf of the appropriate user. Secret sauce, but make it shared with the right crew only 🫶
6.1.1. Digest Authentication Mechanism 💖
The Digest Authentication Mechanism in this memo gives you two things:
- verification of the integrity of a received message, i.e., the message received is the message sent 🧚
The integrity of the message is protected by computing a digest over an appropriate portion of the message. The originator computes the digest, it gets transmitted with the message, and then the recipient verifies it. Receipts, but for packets ✨
- verification of the user on whose behalf the message was generated 👑
A secret value that only SNMP engines authorized to generate messages on behalf of a user know is used in HMAC mode (see [RFC2104]). It also recommends that the hash-function output used as the Message Authentication Code be truncated. Chop chop, babe ✂️
This protocol uses the MD5 [MD5] message digest algorithm 💜 A 128-bit MD5 digest is calculated in a special (HMAC) way over the designated portion of an SNMP message, and then the first 96 bits of that digest are included as part of the message sent to the recipient. So the digest carried in a message is 12 octets, and the private authentication key (the secret) is 16 octets. For the nitty-gritty details, see section 6.3 🦋
6.2. Elements of the Digest Authentication Protocol 🍓
This section has the definitions you need to realize the authentication module defined in this section of this memo. Lots of setup, but it's worth it 💅
6.2.1. Users 🎀
Authentication using this protocol makes use of a defined set of userNames 🧸 For any user on whose behalf a message must be authenticated at a particular SNMP engine, that SNMP engine must have knowledge of that user. An SNMP engine that wants to talk to another SNMP engine also has to know a user that's known to that engine, including the applicable attributes of that user. No stranger danger here 😌
A user and its attributes are defined as follows:
<userName>
A string representing the name of the user.
<authKey>
A user's secret key to be used when calculating a digest.
It MUST be 16 octets long for MD5.6.2.2. msgAuthoritativeEngineID 🌟
The msgAuthoritativeEngineID value in an authenticated message tells you which authoritative SNMP engine is in charge of that particular message (see the definition of SnmpEngineID in the SNMP Architecture document [RFC2271]) 🧚
The user's (private) authentication key is normally different at each authoritative SNMP engine, so the snmpEngineID is what gets used to select the right key for the authentication process. Key matchmaking, literally 💘
6.2.3. SNMP Messages Using this Authentication Protocol 💌
Messages using this authentication protocol carry a msgAuthenticationParameters field as part of the msgSecurityParameters. For this protocol, the msgAuthenticationParameters field is the serialized OCTET STRING that represents the first 12 octets of the HMAC-MD5-96 output, computed over the wholeMsg ✨
The digest is calculated over the wholeMsg, so if a message is authenticated, that also means every field in the message is intact and hasn't been tampered with. Full integrity check, periodt 🔒
6.2.4. Services provided by the HMAC-MD5-96 Authentication Module 🌈
This section describes the inputs and outputs that the HMAC-MD5-96 Authentication module expects and produces when the User-based Security module calls it for services 🎀
6.2.4.1. Services for Generating an Outgoing SNMP Message 📤
The HMAC-MD5-96 authentication protocol assumes that the caller does the selection of the authKey, and that the caller passes in the secret key to be used 🗝️
When it's done, the authentication module returns statusInformation and, if the message digest was calculated correctly, the wholeMsg with the digest inserted at the proper place. The abstract service primitive looks like this:
statusInformation = -- success or failure
authenticateOutgoingMsg(
IN authKey -- secret key for authentication
IN wholeMsg -- unauthenticated complete message
OUT authenticatedWholeMsg -- complete authenticated message
)The abstract data elements are:
- statusInformation: an indication of whether the authentication process was successful. If not, it's an indication of the problem 🚨
- authKey: the secret key to be used by the authentication algorithm. The length of this key MUST be 16 octets 🔑
- wholeMsg: the message to be authenticated 📨
- authenticatedWholeMsg: the authenticated message (including the inserted digest) on output ✨
Note that the authParameters field is filled in by the authentication module, and this field should already be present in the wholeMsg before the Message Authentication Code (MAC) is generated. Order of operations matters, bestie 🙈
6.2.4.2. Services for Processing an Incoming SNMP Message 📥
The HMAC-MD5-96 authentication protocol assumes that the caller does the selection of the authKey, and that the caller passes in the secret key to be used 🗝️
When it's done, the authentication module returns statusInformation and, if the message digest was calculated correctly, the wholeMsg as it was processed. The abstract service primitive looks like this:
statusInformation = -- success or failure
authenticateIncomingMsg(
IN authKey -- secret key for authentication
IN authParameters -- as received on the wire
IN wholeMsg -- as received on the wire
OUT authenticatedWholeMsg -- complete authenticated message
)The abstract data elements are:
- statusInformation: an indication of whether the authentication process was successful. If not, it's an indication of the problem 🚨
- authKey: the secret key to be used by the authentication algorithm. The length of this key MUST be 16 octets 🔑
- authParameters: the authParameters from the incoming message 📬
- wholeMsg: the message to be authenticated on input, and the authenticated message on output 🔄
- authenticatedWholeMsg: the whole message after the authentication check is complete ✅
6.3. Elements of Procedure 🧩
This section describes the procedures for the HMAC-MD5-96 authentication protocol. Step-by-step time, bestie 💅
6.3.1. Processing an Outgoing Message 📤
This section describes the procedure an SNMP engine follows whenever it has to authenticate an outgoing message using the usmHMACMD5AuthProtocol 🦋
1) The msgAuthenticationParameters field is set to the serialization, according to the rules in [RFC1906], of an OCTET STRING containing 12 zero octets 🌸
2) From the secret authKey, two keys, K1 and K2, get derived:
- a) extend the authKey to 64 octets by appending 48 zero octets, and save it as
extendedAuthKey🧁 - b) obtain IPAD by replicating the octet
0x3664 times 🔁 - c) obtain K1 by XORing
extendedAuthKeywith IPAD 💫 - d) obtain OPAD by replicating the octet
0x5C64 times 🔁 - e) obtain K2 by XORing
extendedAuthKeywith OPAD 💫
4) Prepend K1 to the wholeMsg, and calculate the MD5 digest over it according to [MD5] 🧮
5) Prepend K2 to the result of step 4, and calculate the MD5 digest over it according to [MD5]. Take the first 12 octets of the final digest, which is the Message Authentication Code (MAC) 🎯
6) Replace the msgAuthenticationParameters field with the MAC obtained in step 5 🔄
7) The authenticatedWholeMsg is then returned to the caller, together with statusInformation indicating success 🎉
6.3.2. Processing an Incoming Message ✨
okay so this is the procedure an SNMP engine follows whenever it has to authenticate an incoming message with the usmHMACMD5AuthProtocol 💅
- if the digest in the msgAuthenticationParameters field isn't exactly 12 octets long, that's a failure 🙅♀️ and the engine returns an errorIndication (authenticationError) to the calling module
- save the MAC that came in the msgAuthenticationParameters field 🔐
- replace the digest in the msgAuthenticationParameters field with 12 zero octets 🫧
- now we derive two keys, K1 and K2, from the secret authKey 🔑
- a) extend the authKey to 64 octets by appending 48 zero octets, and save the result as extendedAuthKey
- b) get IPAD by replicating the octet 0x36 64 times
- c) get K1 by XORing extendedAuthKey with IPAD
- d) get OPAD by replicating the octet 0x5C 64 times
- e) get K2 by XORing extendedAuthKey with OPAD
- time to calculate the MAC over the wholeMsg 🧮
- a) prepend K1 to the wholeMsg and compute the MD5 digest over that
- b) prepend K2 to the result of step 5.a and compute the MD5 digest over that
- c) the first 12 octets of the result from step 5.b is the MAC
Then the msgAuthenticationParameters field gets the MAC value we saved back in step 2 🎀
- compare the newly calculated MAC with the MAC saved in step 2. if they don't match, that's a failure and the engine returns an errorIndication (authenticationFailure) to the calling module 💀
- if everything checks out, the authenticatedWholeMsg and a statusInformation showing success go back to the caller 🎉
7. HMAC-SHA-96 Authentication Protocol 🦋
this section is all about the HMAC-SHA-96 authentication protocol 💖 it uses the SHA hash-function described in [SHA-NIST], in the HMAC mode from [RFC2104], and then truncates the output down to 96 bits ✂️
this protocol is identified by usmHMACSHAAuthProtocol 🏷️
Over time, other authentication protocols may be defined, either as a replacement for this one or alongside it. 🌸
7.1. Mechanisms 🧚
- For data integrity, a message digest algorithm is required. A digest gets calculated over an appropriate chunk of an SNMP message, and that digest is included in the message sent to the recipient 📦
- For data origin authentication and data integrity, a secret value gets prepended to the SNMP message before the digest is computed. The calculated digest is then partly inserted into the message before it's transmitted. The prepended secret is never transmitted 🤫 The secret value is shared by all SNMP engines that are authorized to originate messages on behalf of the appropriate user 👑
7.1.1. Digest Authentication Mechanism 💅
The Digest Authentication Mechanism defined in this memo gives you two things ✨
- Integrity verification: the message you receive is the message that was sent 🔍 The digest is computed over an appropriate portion of the message, computed by the originator, transmitted along with the message, and verified by the recipient 🫶
- User verification: we check the user the message was generated for 🪪 A secret value, known only to SNMP engines authorized to generate messages on behalf of a user, is used in HMAC mode (see [RFC2104]). It also recommends truncating the hash-function output that's used as the Message Authentication Code 🎀
This mechanism uses the SHA [SHA-NIST] message digest algorithm. A 160-bit SHA digest is calculated in a special (HMAC) way over the designated portion of an SNMP message, and the first 96 bits of that digest are included in the message sent to the recipient 🔥 The digest carried in a message is 12 octets, and the private authentication key (the secret) is 20 octets. For the details, see section 7.3 📖
7.2. Elements of the HMAC-SHA-96 Authentication Protocol 💜
this section has the definitions you need to build the authentication module described here 🧩
7.2.1. Users 🙋♀️
Authentication with this protocol uses a defined set of userNames 🏷️ For any user on whose behalf a message must be authenticated at a particular SNMP engine, that engine has to know that user. An SNMP engine that wants to talk to another SNMP engine also has to know a user known to that engine, including that user's applicable attributes 🫧
a user and its attributes are defined like this:
<userName>: a string representing the name of the user<authKey>: the user's secret key, used when calculating a digest. It MUST be 20 octets long for SHA. 🔑
7.2.2. msgAuthoritativeEngineID 🌈
the msgAuthoritativeEngineID value in an authenticated message says which authoritative SNMP engine is responsible for that message (see the definition of SnmpEngineID in the SNMP Architecture document [RFC2271]) 🧠
a user's (private) authentication key is usually different at each authoritative SNMP engine, so the snmpEngineID is what picks the right key for the authentication process 🗝️
7.2.3. SNMP Messages Using this Authentication Protocol 💌
messages using this protocol carry a msgAuthenticationParameters field as part of the msgSecurityParameters 🎁 For this protocol, the msgAuthenticationParameters field is the serialized OCTET STRING holding the first 12 octets of the HMAC-SHA-96 output computed over the wholeMsg 💖
the digest is calculated over the wholeMsg, so if a message gets authenticated, that also means all of its fields are intact and haven't been tampered with 🛡️ no cap, that's iconic 😭
7.2.4. Services provided by the HMAC-SHA-96 Authentication Module 🛎️
this section describes the inputs and outputs the HMAC-SHA-96 Authentication module expects and produces when the User-based Security module calls it for services 📞
7.2.4.1. Services for Generating an Outgoing SNMP Message 📤
HMAC-SHA-96 authentication assumes the caller picks the authKey and passes in the secret key to use 🔑
when it's done, the authentication module returns statusInformation and, if the message digest was calculated correctly, the wholeMsg with the digest inserted in the right spot ✨ The abstract service primitive looks like this:
statusInformation = -- success or failure
authenticateOutgoingMsg(
IN authKey -- secret key for authentication
IN wholeMsg -- unauthenticated complete message
OUT authenticatedWholeMsg -- complete authenticated message
)the abstract data elements are:
statusInformation: says whether the authentication process worked. If it didn't, it says what the problem was 🙈authKey: the secret key the authentication algorithm uses. The length of this key MUST be 20 octets. 🔐wholeMsg: the message to be authenticated 📨authenticatedWholeMsg: the authenticated message, including the inserted digest, on output 🎀
the authParameters field gets filled in by the authentication module, but that field should already be present in the wholeMsg before the Message Authentication Code (MAC) is generated. don't skip this bit bestie 🚨
7.2.4.2. Services for Processing an Incoming SNMP Message 📥
same deal here: HMAC-SHA-96 authentication assumes the caller picks the authKey and passes in the secret key to use 🗝️
when it's done, the authentication module returns statusInformation and, if the message digest was calculated correctly, the wholeMsg as it was processed 💅 The abstract service primitive looks like this:
statusInformation = -- success or failure
authenticateIncomingMsg(
IN authKey -- secret key for authentication
IN authParameters -- as received on the wire
IN wholeMsg -- as received on the wire
OUT authenticatedWholeMsg -- complete authenticated message
)the abstract data elements are:
statusInformation: says whether the authentication process worked. If it didn't, it says what the problem was 🙈authKey: the secret key the authentication algorithm uses. The length of this key MUST be 20 octets. 🔐authParameters: the authParameters from the incoming message 📬wholeMsg: the message to be authenticated on input, and the authenticated message on output 🎀authenticatedWholeMsg: the whole message once the authentication check is complete ✨
7.3. Elements of Procedure 🎬
this section describes the procedures for the HMAC-SHA-96 authentication protocol 💖
7.3.1. Processing an Outgoing Message 📤
this section describes the procedure an SNMP engine follows whenever it has to authenticate an outgoing message with the usmHMACSHAAuthProtocol 🔥
- set the msgAuthenticationParameters field to the serialization, following the rules in [RFC1906], of an OCTET STRING containing 12 zero octets 🫧
- from the secret authKey, derive two keys, K1 and K2 🔑
- a) extend the authKey to 64 octets by appending 44 zero octets, and save the result as extendedAuthKey
- b) get IPAD by replicating the octet 0x36 64 times
- c) get K1 by XORing extendedAuthKey with IPAD
- d) get OPAD by replicating the octet 0x5C 64 times
- e) get K2 by XORing extendedAuthKey with OPAD
- prepend K1 to the wholeMsg and calculate the SHA digest over it, according to [SHA-NIST] 🧮
- prepend K2 to the result of step 4, and calculate the SHA digest over it, according to [SHA-NIST]. Take the first 12 octets of the final digest, which is the Message Authentication Code (MAC) 🎀
- replace the msgAuthenticationParameters field with the MAC obtained in step 5 💌
- the authenticatedWholeMsg goes back to the caller, together with a statusInformation showing success 🎉
7.3.2. Processing an Incoming Message ✨
Okay bestie, this is the part where an SNMP engine checks the authentication on an incoming message whenever it has to use usmHMACSHAAuthProtocol 🔐. Here's the order it goes in:
- If the digest in the
msgAuthenticationParametersfield is not 12 octets long, it's a failure. An errorIndication (authenticationError) gets sent back to the calling module 🚫 - The MAC that came in the
msgAuthenticationParametersfield gets saved 💾 - The digest in
msgAuthenticationParametersgets replaced with 12 zero octets 🫧 - Two keys, K1 and K2, get derived from the secret
authKey🗝️- a) Extend
authKeyto 64 octets by appending 44 zero octets, and save the result asextendedAuthKey - b) Get IPAD by repeating the octet 0x36 64 times
- c) Get K1 by XORing
extendedAuthKeywith IPAD - d) Get OPAD by repeating the octet 0x5C 64 times
- e) Get K2 by XORing
extendedAuthKeywith OPAD
- a) Extend
- Calculate the MAC over the
wholeMsg🧮- a) Prepend K1 to
wholeMsgand calculate the SHA digest over that - b) Prepend K2 to the result of step 5.a and calculate the SHA digest over that
- c) The first 12 octets of the result of step 5.b is the MAC
- a) Prepend K1 to
Then the msgAuthenticationParameters field gets replaced with the MAC value that was saved in step 2.
- Compare the newly calculated MAC with the MAC saved in step 2. If they don't match, it's a failure and an errorIndication (
authenticationFailure) goes back to the calling module 💔 - If everything checks out, the
authenticatedWholeMsgand a statusInformation showing success get returned to the caller 🎉
8. CBC-DES Symmetric Encryption Protocol 🔒
This section is all about the CBC-DES Symmetric Encryption Protocol 💅. It's the first privacy protocol ever defined for the User-based Security Model, so it's kind of a big deal 👑
This protocol is identified by usmDESPrivProtocol 🪪
Over time, other privacy protocols may be defined, either to replace this one or to sit alongside it 🌈
8.1. Mechanisms 🧩
- For data confidentiality, an encryption algorithm is required 🤫 An appropriate portion of the message gets encrypted before it's sent. The User-based Security Model says the scopedPDU is the part of the message that needs encrypting.
- A secret value, combined with a timeliness value, is used to create the en/decryption key and the initialization vector 🔑 That secret value is shared by all SNMP engines that are allowed to send messages on behalf of the right user.
8.1.1. Symmetric Encryption Protocol 🔐
The Symmetric Encryption Protocol in this memo gives you data confidentiality 🫶 The part of an SNMP message you designate gets encrypted and then included in the message that's sent to the recipient.
Two organizations have published specs for DES: the National Institute of Standards and Technology (NIST) [DES-NIST] and the American National Standards Institute [DES-ANSI]. Each one has a companion Modes of Operation spec ([DESO-NIST] and [DESO-ANSI], respectively) 📚
NIST has also put out three extra docs that implementers might find useful 🤓
- There's a guidelines document for implementing and using DES, including functional specs for DES and its modes of operation [DESG-NIST].
- There's a validation test suite for DES [DEST-NIST]. It's built to test every aspect of DES and is super useful for pinpointing specific problems 🎯
- There's a maintenance test for DES [DESM-NIST]. It uses a minimal amount of data and processing to test all the components of DES. It gives you a simple yes-or-no answer on whether things are working, and it's handy to run as part of an initialization step, like when a computer reboots 🔄
8.1.1.1. DES key and Initialization Vector. 🗝️
The first 8 octets of the 16-octet secret (private privacy key) are used as the DES key. Since DES only uses 56 bits, the Least Significant Bit in each octet gets ignored 🙈
The Initialization Vector for encryption is obtained with the following procedure 🪄
The last 8 octets of the 16-octet secret (private privacy key) are used as the pre-IV.
Now here's the thing 👀 If two different packets are encrypted with the same key, their IVs must not be the same (in other words, the IV must never repeat). So we need to "salt" the pre-IV with something unique for each packet. The "salt" is an 8-octet string. Its input is the concatenation of the generating SNMP engine's 32-bit snmpEngineBoots and a local 32-bit integer that the encryption engine maintains. That 32-bit integer gets set to an arbitrary value at boot time 🥹
The 32-bit snmpEngineBoots becomes the first 4 octets of the "salt" (Most Significant Byte first). The 32-bit integer then becomes the last 4 octets of the "salt" (also Most Significant Byte first). The resulting "salt" is XOR-ed with the pre-IV. The 8-octet "salt" then goes into the privParameters field, encoded as an OCTET STRING. After that, the "salt" integer gets modified. We recommend incrementing it by one, and wrapping around when it hits the maximum value ♻️
How exactly the value of the "salt" (and therefore the IV) changes is an implementation issue, as long as you take steps to avoid producing a duplicate IV 💖
The "salt" must be placed in the privParameters field so the receiving entity can compute the correct IV and decrypt the message 🔓
8.1.1.2. Data Encryption. 🍓
The data to be encrypted is treated as a sequence of octets. Its length should be an integral multiple of 8. If it isn't, the data gets padded at the end as needed. The actual pad value doesn't matter 🙌
The data is encrypted in Cipher Block Chaining mode 🔗
The plaintext is split into 64-bit blocks 🧱
The plaintext for each block is XOR-ed with the ciphertext of the previous block. That result gets encrypted, and the output of the encryption is the ciphertext for that block. This keeps going until there are no more plaintext blocks 🔁
For the very first block, the Initialization Vector is used instead of the ciphertext of a previous block ✨
8.1.1.3. Data Decryption 🫧
Before decryption, the encrypted data length gets verified. If the length of the OCTET STRING to be decrypted isn't an integral multiple of 8 octets, the decryption process halts and an appropriate exception gets noted 🚨 When decrypting, the padding is ignored.
The first ciphertext block is decrypted, the decryption output is XOR-ed with the Initialization Vector, and the result is the first plaintext block 🪞
For each block after that, the ciphertext block is decrypted, the decryption output is XOR-ed with the previous ciphertext block, and the result is the plaintext block 💜
8.2. Elements of the DES Privacy Protocol 🧚
This section has the definitions needed to make the privacy module in this memo work 🛠️
8.2.1. Users 👯
Data en/decryption with this Symmetric Encryption Protocol uses a defined set of userNames. Any user on whose behalf a message has to be en/decrypted at a particular SNMP engine must be known to that engine. An SNMP engine that wants to talk to another SNMP engine also has to know a user known to that other engine, including the applicable attributes of that user 🫶
A user and its attributes are defined like this 📋
<userName>: An octet string with the name of the user.<privKey>: The user's secret key, used as input for the DES key and IV. The length of this key MUST be 16 octets. 🔑
8.2.2. msgAuthoritativeEngineID 🪪
The msgAuthoritativeEngineID value in an authenticated message says which authoritative SNMP engine handles that message. (Check the definition of SnmpEngineID in the SNMP Architecture document [RFC2271] for more.)
The user's (private) privacy key is usually different at each authoritative SNMP engine, so the snmpEngineID is used to pick the right key for the en/decryption process 🗝️
8.2.3. SNMP Messages Using this Privacy Protocol 📨
Messages using this privacy protocol carry a msgPrivacyParameters field as part of msgSecurityParameters. For this protocol, msgPrivacyParameters is the serialized OCTET STRING that represents the "salt" used to create the IV 🧂
8.2.4. Services provided by the DES Privacy Module 🎀
This section describes the inputs and outputs the DES Privacy module expects and produces when the User-based Security module calls it for services 📞
8.2.4.1. Services for Encrypting Outgoing Data 🔐
This DES privacy protocol assumes the caller picks the privKey and passes in the secret key to be used 🙋♀️
When it's done, the privacy module returns statusInformation. If the encryption worked, it also returns the encryptedPDU and the msgPrivacyParameters encoded as an OCTET STRING. The abstract service primitive looks like this 👇
statusInformation = -- success of failure
encryptData(
IN encryptKey -- secret key for encryption
IN dataToEncrypt -- data to encrypt (scopedPDU)
OUT encryptedData -- encrypted data (encryptedPDU)
OUT privParameters -- filled in by service provider
)The abstract data elements are:
statusInformation: Says whether the encryption worked or failed. If it failed, it says what the error was.encryptKey: The secret key the encryption algorithm uses. The length of this key MUST be 16 octets. 🗝️dataToEncrypt: The data that must be encrypted.encryptedData: The encrypted data, once it's successfully done.privParameters: The privParameters, encoded as an OCTET STRING.
8.2.4.2. Services for Decrypting Incoming Data 🔓
This DES privacy protocol also assumes the caller picks the privKey and passes in the secret key to use 💅
When it's done, the privacy module returns statusInformation. If the decryption worked, it also returns the scopedPDU in plain text. The abstract service primitive looks like this 👇
statusInformation =
decryptData(
IN decryptKey -- secret key for decryption
IN privParameters -- as received on the wire
IN encryptedData -- encrypted data (encryptedPDU)
OUT decryptedData -- decrypted data (scopedPDU)
)The abstract data elements are:
statusInformation: Says whether the data was decrypted successfully and, if not, what the error was.decryptKey: The secret key the decryption algorithm uses. The length of this key MUST be 16 octets. 🔑privParameters: The "salt" used to calculate the IV.encryptedData: The data to be decrypted.decryptedData: The decrypted data.
8.3. Elements of Procedure. 📝
This section describes the procedures for the DES privacy protocol 🌸
8.3.1. Processing an Outgoing Message ✨
okay so this is the procedure an SNMP engine follows any time it needs to encrypt part of an outgoing message using the usmDESPrivProtocol 🔐 it's giving very ordered recipe energy
- The secret
cryptKeygets used to build the DES encryption key, the"salt", and the DES pre-IV (see section 8.1.1.1 for the deets) 🔑 - The
privParametersfield gets set to the serialization of an OCTET STRING representing the"salt"string, serialized by the rules in [RFC1906] 🧂 - The scopedPDU gets encrypted (see section 8.1.1.2), and then the encrypted data gets serialized as an OCTET STRING following the rules in [RFC1906] 🔒
- The serialized OCTET STRING of the encrypted scopedPDU, together with
privParametersand astatusInformationsaying success, gets handed back to the calling module 🎀
8.3.2. Processing an Incoming Message 📥
this part is the same vibe but in reverse. it's the procedure an SNMP engine follows whenever it has to decrypt part of an incoming message using the usmDESPrivProtocol 🪄
- If the
privParametersfield isn't an 8-octet OCTET STRING, an error indication (decryptionError) gets returned to the calling module 🙅 - The
"salt"gets pulled out of theprivParametersfield 🫧 - The secret
cryptKeyand the"salt"then get used to build the DES decryption key and pre-IV (see section 8.1.1.1) 🗝️ - The
encryptedPDUgets decrypted (see section 8.1.1.3) 🔓 - If the
encryptedPDUcan't be decrypted, an error indication (decryptionError) gets returned to the calling module 💀 - The decrypted scopedPDU and a
statusInformationsaying success get returned to the calling module 💅
9. Intellectual Property 🧠
okay so the IETF doesn't take a side on whether any intellectual property or other rights might be claimed over the tech in this doc, or on whether you can actually get a license for those rights. it's also not saying it went looking for any of them. if you want info on how the IETF handles rights for standards-track and standards-related docs, check out BCP-11 📚 claims of rights that were made available for publication, assurances that licenses will be available, or the result of trying to get a general license or permission for implementers or users of this spec can all be obtained from the IETF Secretariat 📬
the IETF also invites anyone to flag any copyrights, patents, patent applications, or other proprietary rights that might cover tech needed to practice this standard. send that info to the IETF Executive Director 🫶
10. Acknowledgements 🙌
this doc is the result of the SNMPv3 Working Group's hard work, and they deserve a huge shoutout 📣 special thanks go to these SNMPv3 WG members:
- Dave Battle (SNMP Research, Inc.)
- Uri Blumenthal (IBM T.J. Watson Research Center)
- Jeff Case (SNMP Research, Inc.)
- John Curran (BBN)
- T. Max Devlin (Hi-TECH Connections)
- John Flick (Hewlett Packard)
- David Harrington (Cabletron Systems Inc.)
- N.C. Hien (IBM T.J. Watson Research Center)
- Dave Levi (SNMP Research, Inc.)
- Louis A Mamakos (UUNET Technologies Inc.)
- Paul Meyer (Secure Computing Corporation)
- Keith McCloghrie (Cisco Systems)
- Russ Mundy (Trusted Information Systems, Inc.)
- Bob Natale (ACE*COMM Corporation)
- Mike O'Dell (UUNET Technologies Inc.)
- Dave Perkins (DeskTalk)
- Peter Polkinghorne (Brunel University)
- Randy Presuhn (BMC Software, Inc.)
- David Reid (SNMP Research, Inc.)
- Shawn Routhier (Epilogue)
- Juergen Schoenwaelder (TU Braunschweig)
- Bob Stewart (Cisco Systems)
- Bert Wijnen (IBM T.J. Watson Research Center)
the document is based on recommendations from the IETF Security and Administrative Framework Evolution for SNMP Advisory Team 🧩 the members of that Advisory Team were:
- David Harrington (Cabletron Systems Inc.)
- Jeff Johnson (Cisco Systems)
- David Levi (SNMP Research Inc.)
- John Linn (Openvision)
- Russ Mundy (Trusted Information Systems) chair
- Shawn Routhier (Epilogue)
- Glenn Waters (Nortel)
- Bert Wijnen (IBM T. J. Watson Research Center)
as the Advisory Team and the SNMPv3 Working Group Charter recommended, the design pulls in as much as is practical from previous RFCs and drafts 🍓 which means big thanks to the authors of the earlier designs known as SNMPv2u and SNMPv2*:
- Jeff Case (SNMP Research, Inc.)
- David Harrington (Cabletron Systems Inc.)
- David Levi (SNMP Research, Inc.)
- Keith McCloghrie (Cisco Systems)
- Brian O'Keefe (Hewlett Packard)
- Marshall T. Rose (Dover Beach Consulting)
- Jon Saperia (BGS Systems Inc.)
- Steve Waldbusser (International Network Services)
- Glenn W. Waters (Bell-Northern Research Ltd.)
11. Security Considerations 🚨
okay this part is SERIOUS bestie. security is literally the whole point 😭 don't skip these, they keep the whole thing from falling apart 🛡️
11.1. Recommended Practices 🛡️
this section covers practices that help the mechanisms defined in this memo run securely and effectively 💖
- An SNMP engine must discard SNMP Response messages that do not correspond to any currently outstanding Request message. It is the responsibility of the Message Processing module to take care of this. For example, it can use a msgID for that.
An SNMP Command Generator Application must discard any Response PDU for which there is no currently outstanding Request PDU. For example, for SNMPv2 [RFC1905] PDUs, the request-id component in the PDU can be used to correlate Responses to outstanding Requests.
Even though it would be typical for an SNMP engine and an SNMP Command Generator Application to do this as a matter of course, it's a big deal when you're using these security protocols because of the possibility of message duplication, whether malicious or otherwise 🙈
- If an SNMP engine uses a msgID for correlating Response messages to outstanding Request messages, then it MUST use different msgIDs in all such Request messages that it sends out during a Time Window (150 seconds) period.
A Command Generator or Notification Originator Application MUST use different request-ids in all Request PDUs that it sends out during a TimeWindow (150 seconds) period.
This has to be done to protect against the possibility of message duplication, whether malicious or otherwise 🚫
For example, starting operations with a msgID and/or request-id value of zero is not a good idea 🙅 starting with an unpredictable number (so they don't start out the same after each reboot) and then incrementing by one would be acceptable 🧚
- An SNMP engine should perform time synchronization using authenticated messages to protect against the possibility of message duplication, whether malicious or otherwise 🕰️
- When sending state altering messages to a managed authoritative SNMP engine, a Command Generator Application should delay sending successive messages to that managed SNMP engine until a positive acknowledgement is received for the previous message or until the previous message expires.
No message ordering is imposed by SNMP. Messages may be received in any order relative to their time of generation, and each will be processed in the order received. Note that when an authenticated message is sent to a managed SNMP engine, it will be valid for approximately 150 seconds under normal circumstances, and it's subject to replay during that window 😬 an SNMP engine and SNMP Command Generator Applications have to cope with the loss and re-ordering of messages caused by network anomalies as a matter of course 🌪️
however, a managed object, snmpSetSerialNo [RFC1907], is specifically defined for use with SNMP Set operations. it gives you a mechanism to make sure SNMP messages get processed in a specific order 🎯
- The frequency with which the secrets of a User-based Security Model user should be changed is indirectly related to how often those secrets get used ⏳
protecting the secrets from disclosure is critical to the overall security of the protocols 🔐 frequent use of a secret gives a cryptanalyst a steady stream of data they could use to exploit known or perceived weaknesses in an algorithm. frequently changing the secret shuts that down 🔄
changing a secret after every single use is generally seen as the most secure practice, but it can come with a lot of overhead 😮💨
also, in a local environment the threat of disclosure may be less significant, so changing secrets can happen less often. however, when public data networks are used as the communication paths, more caution is wise 🌐✨
11.2. Defining Users 🎀
okay so the whole User-based Security Model is built around the idea of users, the people or things a message gets sent on behalf of. 🫶 How you define a user is totally up to your network's security policy. A user could be a person like joe or jane, a role like operator or administrator, or a combo like joe-operator, jane-operator or joe-admin. 💅 A user can even be a logical entity, like an SNMP Application or a set of them, acting for an individual, a role, a set of individuals, a set of roles, or any combination. It's giving flexible 😭
Appendix A has an algorithm that turns a user "password" into a 16 octet value, which can be used as an authentication key, a privacy key, or both. ngl, using the same password (and therefore the same key) for both authentication and privacy is very poor security practice and should be strongly discouraged 💀. Human-generated passwords can be shorter than the 16 octets the auth and privacy protocols need, and brute force attacks are pretty easy on a relatively short ASCII character set 🙈. So the Appendix A algorithm performs a transformation on the password. If you use the Appendix A algorithm, SNMP implementations (and SNMP configuration applications) must ensure that passwords are at least 8 characters in length.
Because the Appendix A algorithm uses these passwords (nearly) directly, it is very important that they not be easily guessed 🔥. The suggestion is to mix uppercase and lowercase letters, numbers, and punctuation, and to avoid words or phrases that you'd find in a dictionary. Longer passwords make the system more secure. Users may also want to type in multiword phrases, which makes the password longer while keeping it memorable ✨.
Here's the tea 🍵: it's impossible for human users to keep a different password for every SNMP engine, but security requirements strongly discourage using the same key for more than one SNMP engine. So the User-based Security Model uses a compromise proposed in [Localized-key]. It derives the user keys for each SNMP engine from the user's password, in a way that makes it practically impossible to figure out either the user's password or the user's key for another SNMP engine, even from any combination of the user's keys on SNMP engines. big brain energy 🧠
Note however, that if the user's password is disclosed, then key localization won't help and network security may be compromised in this case. Therefore a user's password or non-localized key MUST NOT be stored on a managed device/node. Instead the localized key SHALL be stored (if at all), so that, in case a device does get compromised, no other managed or managing devices get compromised 🛡️.
11.3. Conformance ✨
To be termed a "Secure SNMP implementation" based on the User-based Security Model, an SNMP implementation MUST:
- implement one or more Authentication Protocol(s). The HMAC-MD5-96 and HMAC-SHA-96 Authentication Protocols defined in this memo are examples of such protocols.
- to the maximum extent possible, prohibit access to the secret(s) of each user about which it maintains information in a Local Configuration Datastore (LCD) under all circumstances except as required to generate and/or validate SNMP messages with respect to that user.
- implement the key-localization mechanism.
- implement the SNMP-USER-BASED-SM-MIB.
In addition, an authoritative SNMP engine SHOULD provide initial configuration in accordance with Appendix A.1 🌸.
Implementation of a Privacy Protocol (the DES Symmetric Encryption Protocol defined in this memo is one such protocol) is optional 🦋.
12. References 📚
[RFC1321] Rivest, R., "Message Digest Algorithm MD5", RFC 1321, April 1992.
[RFC1903] Case, J., McCloghrie, K., Rose, M. and S. Waldbusser, "Textual Conventions for Version 2 of the Simple Network Management Protocol (SNMPv2)", RFC 1903, January 1996.
[RFC1905] Case, J., McCloghrie, K., Rose, M. and S. Waldbusser, "Protocol Operations for Version 2 of the Simple Network Management Protocol (SNMPv2)", RFC 1905, January 1996.
[RFC1906] Case, J., McCloghrie, K., Rose, M. and S. Waldbusser, "Transport Mappings for Version 2 of the Simple Network Management Protocol (SNMPv2)", RFC 1906, January 1996.
[RFC1907] Case, J., McCloghrie, K., Rose, M. and S. Waldbusser, "Management Information Base for Version 2 of the Simple Network Management Protocol (SNMPv2)", RFC 1907 January 1996.
[RFC2104] Krawczyk, H., Bellare, M., and R. Canetti, "HMAC: Keyed-Hashing for Message Authentication", RFC 2104, February 1997.
[RFC2028] Hovey, R., and S. Bradner, "The Organizations Involved in the IETF Standards Process", BCP 11, RFC 2028, October 1996.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC2271] Harrington, D., Presuhn, R., and B. Wijnen, "An Architecture for describing SNMP Management Frameworks", RFC 2271, January 1998.
[RFC2272] Case, J., Harrington, D., Presuhn, R., and B. Wijnen, "Message Processing and Dispatching for the Simple Network Management Protocol (SNMP)", RFC 2272, January 1998.
[Localized-Key] U. Blumenthal, N. C. Hien, B. Wijnen "Key Derivation for Network Management Applications" IEEE Network Magazine, April/May issue, 1997.
[DES-NIST] Data Encryption Standard, National Institute of Standards and Technology. Federal Information Processing Standard (FIPS) Publication 46-1. Supersedes FIPS Publication 46, (January, 1977; reaffirmed January, 1988).
[DES-ANSI] Data Encryption Algorithm, American National Standards Institute. ANSI X3.92-1981, (December, 1980).
[DESO-NIST] DES Modes of Operation, National Institute of Standards and Technology. Federal Information Processing Standard (FIPS) Publication 81, (December, 1980).
[DESO-ANSI] Data Encryption Algorithm - Modes of Operation, American National Standards Institute. ANSI X3.106-1983, (May 1983).
[DESG-NIST] Guidelines for Implementing and Using the NBS Data Encryption Standard, National Institute of Standards and Technology. Federal Information Processing Standard (FIPS) Publication 74, (April, 1981).
[DEST-NIST] Validating the Correctness of Hardware Implementations of the NBS Data Encryption Standard, National Institute of Standards and Technology. Special Publication 500-20.
[DESM-NIST] Maintenance Testing for the Data Encryption Standard, National Institute of Standards and Technology. Special Publication 500-61, (August, 1980).
[SHA-NIST] Secure Hash Algorithm. NIST FIPS 180-1, (April, 1995) http://csrc.nist.gov/fips/fip180-1.txt (ASCII) http://csrc.nist.gov/fips/fip180-1.ps (Postscript)
13. Editors' Addresses 🌷
Uri Blumenthal
IBM T. J. Watson Research
30 Saw Mill River Pkwy,
Hawthorne, NY 10532
USA
EMail: uri@watson.ibm.com
Phone: +1-914-784-7064Bert Wijnen
IBM T. J. Watson Research
Schagen 33
3461 GL Linschoten
Netherlands
EMail: wijnen@vnet.ibm.com
Phone: +31-348-432-794APPENDIX A - Installation 🧚
A.1. SNMP engine Installation Parameters
During installation, an authoritative SNMP engine SHOULD (in the meaning as defined in [RFC2119]) be configured with several initial parameters. These include:
- A security posture
The choice of security posture determines if initial configuration is implemented and if so how. One of three possible choices is selected:
minimum-secure,
semi-secure,
very-secure (i.e., no-initial-configuration)In the case of a very-secure posture, there is no initial configuration, and so the following steps are irrelevant.
- one or more secrets
These are the authentication/privacy secrets for the first user to be configured.
One way to accomplish this is to have the installer enter a "password" for each required secret. The password is then algorithmically converted into the required secret by:
- forming a string of length 1,048,576 octets by repeating the value of the password as often as necessary, truncating accordingly, and using the resulting string as the input to the MD5 algorithm [MD5]. The resulting digest, termed "
digest1", is used in the next step. - a second string is formed by concatenating
digest1, the SNMP engine'ssnmpEngineIDvalue, anddigest1. This string is used as input to the MD5 algorithm [MD5].
The resulting digest is the required secret (see Appendix A.2).
With these configured parameters, the SNMP engine instantiates the following usmUserEntry in the usmUserTable:
no privacy support privacy support
------------------ ---------------
usmUserEngineID localEngineID localEngineID
usmUserName "initial" "initial"
usmUserSecurityName "initial" "initial"
usmUserCloneFrom ZeroDotZero ZeroDotZero
usmUserAuthProtocol usmHMACMD5AuthProtocol usmHMACMD5AuthProtocol
usmUserAuthKeyChange "" ""
usmUserOwnAuthKeyChange "" ""
usmUserPrivProtocol none usmDESPrivProtocol
usmUserPrivKeyChange "" ""
usmUserOwnPrivKeyChange "" ""
usmUserPublic "" ""
usmUserStorageType anyValidStorageType anyValidStorageType
usmUserStatus active activeA.2. Password to Key Algorithm
A sample code fragment (section A.2.1) demonstrates the password to key algorithm which can be used when mapping a password to an authentication or privacy key using MD5. The reference source code of MD5 is available in [RFC1321].
Another sample code fragment (section A.2.2) demonstrates the password to key algorithm which can be used when mapping a password to an authentication or privacy key using SHA (documented in SHA-NIST).
An example of the results of a correct implementation is provided (section A.3) which an implementor can use to check if his implementation produces the same result 🍓.
A.2.1. Password to Key Sample Code for MD5
void password_to_key_md5(
u_char *password, /* IN */
u_int passwordlen, /* IN */
u_char *engineID, /* IN - pointer to snmpEngineID */
u_int engineLength /* IN - length of snmpEngineID */
u_char *key) /* OUT - pointer to caller 16-octet buffer */{
MD5_CTX MD;
u_char *cp, password_buf[64];
u_long password_index = 0;
u_long count = 0, i;
MD5Init (&MD); /* initialize MD5 */
/**********************************************/
/* Use while loop until we've done 1 Megabyte */
/**********************************************/
while (count < 1048576) {
cp = password_buf;
for (i = 0; i < 64; i++) {
/*************************************************/
/* Take the next octet of the password, wrapping */
/* to the beginning of the password as necessary.*/
/*************************************************/
*cp++ = password[password_index++ % passwordlen];
}
MD5Update (&MD, password_buf, 64);
count += 64;
}
MD5Final (key, &MD); /* tell MD5 we're done */
/*****************************************************/
/* Now localize the key with the engineID and pass */
/* through MD5 to produce final key */
/* May want to ensure that engineLength <= 32, */
/* otherwise need to use a buffer larger than 64 */
/*****************************************************/
memcpy(password_buf, key, 16);
memcpy(password_buf+16, engineID, engineLength);
memcpy(password_buf+engineLength, key, 16);
MD5Init(&MD);
MD5Update(&MD, password_buf, 32+engineLength);
MD5Final(key, &MD);
return;
}A.2.2. Password to Key Sample Code for SHA
void password_to_key_sha(
u_char *password, /* IN */
u_int passwordlen, /* IN */
u_char *engineID, /* IN - pointer to snmpEngineID */
u_int engineLength /* IN - length of snmpEngineID */
u_char *key) /* OUT - pointer to caller 20-octet buffer */
{
SHA_CTX SH;
u_char *cp, password_buf[72];
u_long password_index = 0;
u_long count = 0, i;
SHAInit (&SH); /* initialize SHA */
/**********************************************/
/* Use while loop until we've done 1 Megabyte */
/**********************************************/
while (count < 1048576) {
cp = password_buf;
for (i = 0; i < 64; i++) {
/*************************************************/
/* Take the next octet of the password, wrapping */
/* to the beginning of the password as necessary.*/
/*************************************************/
*cp++ = password[password_index++ % passwordlen];
}
SHAUpdate (&SH, password_buf, 64);
count += 64;
}
SHAFinal (key, &SH); /* tell SHA we're done */
/*****************************************************/
/* Now localize the key with the engineID and pass */
/* through SHA to produce final key */
/* May want to ensure that engineLength <= 32, */
/* otherwise need to use a buffer larger than 72 */
/*****************************************************/
memcpy(password_buf, key, 20);
memcpy(password_buf+20, engineID, engineLength);
memcpy(password_buf+engineLength, key, 20);
SHAInit(&SH);
SHAUpdate(&SH, password_buf, 40+engineLength);
SHAFinal(key, &SH);
return;
}A.3. Password to Key Sample Results
A.3.1. Password to Key Sample Results using MD5
Here's a sample of what the password to key algorithm spits out for a 16-octet key using MD5. 💖
With a password of "maplesyrup", the output of the password to key algorithm before the key gets localized with the SNMP engine's snmpEngineID looks like this:
'9f af 32 83 88 4e 92 83 4e bc 98 47 d8 ed d9 63'HAfter that intermediate key (shown above) is localized with the snmpEngineID value of:
'00 00 00 00 00 00 00 00 00 00 00 02'Hthe final output of the password to key algorithm is:
'52 6f 5e ed 9f cc e2 6f 89 64 c2 93 07 87 d8 2b'HA.3.2. Password to Key Sample Results using SHA
Same vibe, but now for a 20-octet key using SHA. 🌸
With a password of "maplesyrup", the output before localization with the SNMP engine's snmpEngineID is:
'f1 be a9 ae 66 7f 4f b6 34 1e 51 af 06 80 7e 91 e4 3b 01 ac'HAfter the intermediate key (shown above) is localized with the snmpEngineID value of:
'00 00 00 00 00 00 00 00 00 00 00 02'Hthe final output of the password to key algorithm is:
'8a a3 d9 9e 3e 30 56 f2 bf e3 a9 ee f3 45 d5 39 54 91 12 be'HA.4. Sample encoding of msgSecurityParameters
Okay so msgSecurityParameters in an SNMP message is represented as an OCTET STRING, and that OCTET STRING should be treated as opaque outside a specific Security Model. Hands off, everyone else 🙈
The User-based Security Model is the one that defines what's inside that OCTET STRING: it's a SEQUENCE (see section 2.4). 🧚
Put those two things together and you get an example of msgSecurityParameters for the User-based Security Model, encoded as an OCTET STRING:
04 <length>
30 <length>
04 <length> <msgAuthoritativeEngineID>
02 <length> <msgAuthoritativeEngineBoots>
02 <length> <msgAuthoritativeEngineTime>
04 <length> <msgUserName>
04 0c <HMAC-MD5-96-digest>
04 08 <salt>And here's that same example again, but now with real values. The only things left as placeholders are the digest in msgAuthenticationParameters and the salt in msgPrivacyParameters, since those depend on variable data we haven't defined here. 🎀
Hex Data Description
-------------- -----------------------------------------------
04 39 OCTET STRING, length 57
30 37 SEQUENCE, length 55
04 0c 80000002 msgAuthoritativeEngineID: IBM
01 IPv4 address
09840301 9.132.3.1
02 01 01 msgAuthoritativeEngineBoots: 1
02 02 0101 msgAuthoritativeEngineTime: 257
04 04 62657274 msgUserName: bert
04 0c 01234567 msgAuthenticationParameters: sample value
89abcdef
fedcba98
04 08 01234567 msgPrivacyParameters: sample value
89abcdefB. Full Copyright Statement ✨
Copyright (C) The Internet Society (1998). All Rights Reserved. 🌈
Okay bestie, time for the legal bit 💅 You're allowed to copy this document and share it with anyone. You can also write derivative works that comment on it, explain it, or help people implement it, and publish and distribute those in whole or in part with no restrictions, as long as the copyright notice and this paragraph stay on every copy and derivative work.
BUT the document itself can't be modified in any way. That means no removing the copyright notice or the references to the Internet Society or other Internet organizations. The only exceptions are developing Internet standards (in which case the Internet Standards process copyright procedures must be followed) or translating it into languages other than English. 🫶
These permissions are perpetual, and the Internet Society or its successors or assigns won't revoke them. 🦋
And the fine print, which is very serious and very not fun 😭
This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
📦 where this page comes from
- ietf: rfc2274 1998-01
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms