RFC 9063
Host Identity Protocol Architecture
RFC 9063: Host Identity Protocol Architecture 🦋
Okay bestie, this one is a whole vibe. It's RFC 9063, titled Host Identity Protocol Architecture 💖
- Authors: R. Moskowitz, Ed. (HTT Consulting) and M. Komu (Ericsson) 🫶
- Category: Informational. It's published for information, not as a Standards Track spec 🎀
- Date: July 2021
- ISSN: 2070-1721
- Obsoletes: RFC 4423 💅
- Updates: the header doesn't list any updates 🌸
The Abstract is next, and it's the TL;DR. ✨
Abstract
Okay so this memo describes the Host Identity (HI) namespace 🧚. It gives applications a cryptographic namespace, and it sits in a protocol layer called the Host Identity Protocol. That layer lives between the internetworking and transport layers, and it supports end-host mobility, multihoming, and NAT traversal. It's giving main character energy 👑. The memo also covers the basics of the current namespaces, their strengths and weaknesses, and how a HI namespace will add completeness to them. Then it defines the roles of the HI namespace in the protocols 💜.
This document obsoletes RFC 4423 and addresses the concerns the IESG raised, especially crypto agility 🔥. The Security Considerations section is a big deal too. It covers measures against flooding attacks, using identities in access control lists, weaker types of identifiers, and trust on first use 😭. It also builds on lessons learned from implementing RFC 7401, and it goes further to explain how HIP works as a secure signaling channel 🌈.
Status of This Memo
This document is not an Internet Standards Track specification. It's published for informational purposes 🍓.
This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community, and it went through public review. The Internet Engineering Steering Group (IESG) approved it for publication. Not every document the IESG approves is a candidate for any level of Internet Standard. See Section 2 of RFC 7841 for more on that 🦋.
For the current status of this document, any errata, and how to give feedback on it, check https://www.rfc-editor.org/info/rfc9063 💌.
Copyright Notice
Copyright (c) 2021 IETF Trust and the persons identified as the document authors. All rights reserved 🎀.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication. Please read these carefully, because they spell out your rights and restrictions for this document 📜. Code Components extracted from this document must include Simplified BSD License text, as described in Section 4.e of the Trust Legal Provisions. They come without warranty, as described in the Simplified BSD License 💅.
This document may contain material from IETF Documents or IETF Contributions published or made publicly available before November 10, 2008. The person or people controlling the copyright in some of that material may not have granted the IETF Trust the right to allow modifications outside the IETF Standards Process. Without an adequate license from those copyright holders, this document may not be modified outside the IETF Standards Process, and derivative works of it may not be created outside the IETF Standards Process. The exception is formatting it for publication as an RFC or translating it into languages other than English 🌸.
Table of Contents 🗺️
1. Introduction
2. Terminology
2.1. Terms Common to Other Documents
2.2. Terms Specific to This and Other HIP Documents
3. Background
3.1. A Desire for a Namespace for Computing Platforms
4. Host Identity Namespace
4.1. Host Identifiers
4.2. Host Identity Hash (HIH)
4.3. Host Identity Tag (HIT)
4.4. Local Scope Identifier (LSI)
4.5. Storing Host Identifiers in Directories
5. New Stack Architecture
5.1. On the Multiplicity of Identities
6. Control Plane
6.1. Base Exchange
6.2. End-Host Mobility and Multihoming
6.3. Rendezvous Mechanism
6.4. Relay Mechanism
6.5. Termination of the Control Plane
7. Data Plane
8. HIP and NATs
8.1. HIP and Upper-Layer Checksums
9. Multicast
10. HIP Policies
11. Security Considerations
11.1. MitM Attacks
11.2. Protection against Flooding Attacks
11.3. HITs Used in ACLs
11.4. Alternative HI Considerations
11.5. Trust on First Use
12. IANA Considerations
13. Changes from RFC 4423
14. References
14.1. Normative References
14.2. Informative References
Appendix A. Design Considerations
A.1. Benefits of HIP
A.2. Drawbacks of HIP
A.3. Deployment and Adoption Considerations
A.3.1. Deployment Analysis
A.3.2. HIP in 802.15.4 Networks
A.3.3. HIP and Internet of Things
A.3.4. Infrastructure Applications
A.3.5. Management of Identities in a Commercial Product
A.4. Answers to NSRG Questions
Acknowledgments
Authors' Addresses1. Introduction ✨
Okay so the Internet has two big global namespaces: IP addresses and DNS names 🌐. They've powered the Internet into what it is today, but they have weaknesses too. Basically, since they're all we've got, we keep trying to do too much with them 😭. Semantic overloading and functionality extensions have made both of them really complicated 💀.
The proposed Host Identity namespace is also a global namespace, and it fills an important gap between the IP and DNS namespaces 🫶. A Host Identity conceptually refers to a computing platform. A platform can have more than one Host Identity, because it might want to show a different identity to different communicating peers 🦋. The Host Identity namespace is made up of Host Identifiers (HI). There's exactly one Host Identifier for each Host Identity. The one exception is short transient periods, like key replacement, when more than one identifier might be active 🔄.
Quick heads up, bestie: this text later talks about non-cryptographic Host Identifiers, but the architecture is really focused on cryptographic ones 🔐. Specifically, the Host Identifier is the public key of an asymmetric key pair 💜. Each Host Identity uniquely identifies a single host, so no two hosts share the same Host Identity 🎀. If two or more computing platforms share the same Host Identifier, then together they're instantiating a distributed host 🌈. A Host Identifier can be public, like published in the DNS, or unpublished. Client systems tend to have both public and unpublished Host Identifiers 📱.
There's a subtle but important difference between Host Identities and Host Identifiers 👀. An Identity is the abstract entity being identified. An Identifier is the concrete bit pattern used in the identification process 🧠.
Now, the Host Identifiers could be used in lots of authentication systems, like IKEv2 [RFC7296] ✨. But this architecture introduces a brand-new protocol, the Host Identity Protocol (HIP), along with a cryptographic exchange called the HIP base exchange. See also Section 6 💅. HIP gives limited forms of trust between systems. It also enhances mobility, multihoming, and dynamic IP renumbering. It helps with protocol translation and transition, and it reduces certain types of denial-of-service (DoS) attacks 🛡️.
When HIP is used, the actual payload traffic between two HIP hosts is typically, but not necessarily, protected with Encapsulating Security Payload (ESP) [RFC7402] 🔒. The Host Identities are used to create the ESP Security Associations (SAs) the hosts need, and to authenticate those hosts 🤝. When ESP is used, the actual payload IP packets look exactly like standard ESP-protected IP packets. Nothing about them is different 🎀.
Oh, and tons has been learned about HIP [RFC6538] since [RFC4423] was published 📚. This document stretches Host Identities beyond their original use. They now enable IP connectivity and security, plus general interhost secure signaling at any protocol layer 🚀. The signal might establish a security association between the hosts, or it might just pass information along the channel 💫.
2. Terminology 📖
2.1. Terms Common to Other Documents
This table is the glossary for the basics, so let's get the vocab down 📝.
+==========+===================================================+
| Term | Explanation |
+==========+===================================================+
| Public | The public key of an asymmetric cryptographic key |
| key | pair. Used as a publicly known identifier for |
| | cryptographic identity authentication. Public is |
| | a relative term here, ranging from "known to |
| | peers only" to "known to the world". |
+----------+---------------------------------------------------+
| Private | The private or secret key of an asymmetric |
| key | cryptographic key pair. Assumed to be known only |
| | to the party identified by the corresponding |
| | public key. Used by the identified party to |
| | authenticate its identity to other parties. |
+----------+---------------------------------------------------+
| Public | An asymmetric cryptographic key pair consisting |
| key pair | of public and private keys. For example, Rivest- |
| | Shamir-Adleman (RSA), Digital Signature Algorithm |
| | (DSA) and Elliptic Curve DSA (ECDSA) key pairs |
| | are such key pairs. |
+----------+---------------------------------------------------+
| Endpoint | A communicating entity. For historical reasons, |
| | the term 'computing platform' is used in this |
| | document as a (rough) synonym for endpoint. |
+----------+---------------------------------------------------+Table 1 🌸
2.2. Terms Specific to This and Other HIP Documents ✨
okay so heads up bestie, a lot of these terms are kinda tautologous, self-referential, or defined through circular reference to other terms 💀 that's just how short definitions work. for the deep dive, check the text elsewhere in this doc and the base spec [RFC7401] 📚
+==============+=============================================+
| Term | Explanation |
+==============+=============================================+
| Computing | An entity capable of communicating and |
| platform | computing, for example, a computer. See |
| | the definition of 'Endpoint', above. |
+--------------+---------------------------------------------+
| HIP base | A cryptographic protocol; see also |
| exchange | Section 6. |
+--------------+---------------------------------------------+
| HIP packet | An IP packet that carries a 'Host Identity |
| | Protocol' message. |
+--------------+---------------------------------------------+
| Host | An abstract concept assigned to a |
| Identity | 'computing platform'. See 'Host |
| | Identifier', below. |
+--------------+---------------------------------------------+
| Host | A public key used as a name for a Host |
| Identifier | Identity. |
+--------------+---------------------------------------------+
| Host | A name space formed by all possible Host |
| Identity | Identifiers. |
| namespace | |
+--------------+---------------------------------------------+
| Host | A protocol used to carry and authenticate |
| Identity | Host Identifiers and other information. |
| Protocol | |
+--------------+---------------------------------------------+
| Host | The cryptographic hash used in creating the |
| Identity | Host Identity Tag from the Host Identifier. |
| Hash | |
+--------------+---------------------------------------------+
| Host | A 128-bit datum created by taking a |
| Identity Tag | cryptographic hash over a Host Identifier |
| | plus bits to identify which hash was used. |
+--------------+---------------------------------------------+
| Local Scope | A 32-bit datum denoting a Host Identity. |
| Identifier | |
+--------------+---------------------------------------------+
| Public Host | A published or publicly known Host |
| Identifier | Identifier used as a public name for a Host |
| and Identity | Identity, and the corresponding Identity. |
+--------------+---------------------------------------------+
| Unpublished | A Host Identifier that is not placed in any |
| Host | public directory, and the corresponding |
| Identifier | Host Identity. Unpublished Host Identities |
| and Identity | are typically short lived in nature, being |
| | often replaced and possibly used just once. |
+--------------+---------------------------------------------+
| Rendezvous | A mechanism used to locate mobile hosts |
| Mechanism | based on their HIT. |
+--------------+---------------------------------------------+
Table 23. Background 🌸
okay the internet is built from three main pieces: computing platforms (endpoints), packet transport (aka internetworking) infrastructure, and services (applications) 🧚 and the internet exists to serve two main groups: people and robotic services (silicon-based people, if you will 🤖). everything needs a name so it can interact at scale, and this section is mostly about naming computing platforms and packet transport elements 💅
there are two main namespaces in use for these components: IP addresses and Domain Names 🌐. Domain Names give hierarchical names to some computing platforms and some services. each hierarchy is delegated from the level above, so there's no anonymity in Domain Names 🙈. email addresses, HTTP addresses, and SIP addresses all point at Domain Names.
the IP addressing namespace got overloaded, so it names both interfaces (at Layer 3) and endpoints (for the endpoint-specific part of Layer 3 and for Layer 4) 😭. when IP addresses act as interface names, they're sometimes called "locators", and they serve as an endpoint within a routing topology 🦋
IP addresses are numbers that name networking interfaces, and usually only while the interface is connected to the network ✨. originally, IP addresses had long-term significance. today, the vast number of interfaces use ephemeral and/or non-unique IP addresses. that means every time an interface connects to the network, it gets assigned an IP address 💜
in the current internet, the transport layers are coupled to the IP addresses. neither one can evolve separately from the other 🔥 IPng deliberations were strongly shaped by the decision that a corresponding TCPng would not be created 👑
there are three critical deficiencies with the current namespaces, and they're kind of a big deal 💀 First, setting up initial contact and keeping data flows going between two hosts can be challenging because of private address realms and the ephemeral nature of addresses. Second, confidentiality is not provided in a consistent, trustable manner. Finally, authentication for systems and datagrams is not provided. all of these problems happen because computing platforms aren't well named with the current namespaces 😩
3.1. A Desire for a Namespace for Computing Platforms 💖
okay so an independent namespace for computing platforms could be used in end-to-end operations, no matter how the internetworking layer evolves and across many internetworking layers 🧚 this could support rapid readdressing of the internetworking layer because of mobility, rehoming, or renumbering 🚀
if the namespace for computing platforms is based on public-key cryptography, it can also provide authentication services 🔐. if this namespace is created locally without requiring registration, it can provide anonymity 🫶
so a namespace like this, and the names in it, should have the following characteristics:
- the namespace should be applied to the IP 'kernel' or stack. the IP stack is the 'component' between applications and the packet transport infrastructure 🎀
- the namespace should fully decouple the internetworking layer from the higher layers. the names should replace all occurrences of IP addresses within applications (like in the Transport Control Block, TCB). this replacement can be handled transparently for legacy applications, since the Local Scope Identifiers (LSIs) and HITs are compatible with IPv4 and IPv6 addresses [RFC5338]. however, HIP-aware applications require some modifications from the developers, who may employ networking API extensions for HIP [RFC6317] ✨
- the introduction of the namespace should not mandate any administrative infrastructure. deployment must come from the bottom up, in a pairwise deployment 🌈
- the names should have a fixed-length representation, for easy inclusion in datagram headers and existing programming interfaces (e.g., the TCB) 🍓
- using the namespace should be affordable when used in protocols. this is primarily a packet size issue, and there's also a computational concern in affordability 💅
- name collisions should be avoided as much as possible. the mathematics of the birthday paradox can be used to estimate the chance of a collision in a given population and hash space 🧮 in general, for a random hash space of size n bits, we would expect to obtain a collision after approximately 1.2*sqrt(2^n) hashes were obtained. for 64 bits, this number is roughly 4 billion. a hash size of 64 bits may be too small to avoid collisions in a large population; for example, there is a 1% chance of collision in a population of 640M 😱 for 100 bits (or more), we would not expect a collision until approximately 2^50 (1 quadrillion) hashes were generated. with the currently used hash size of 96 bits [RFC7343], the figure is 2^48 (281 trillions) 🌟
- the names should have a localized abstraction so that they can be used in existing protocols and APIs 🧩
- it must be possible to create names locally. when such names are not published, this can provide anonymity at the cost of making resolvability very difficult 🕵️♀️
- the namespace should provide authentication services 🔑
- the names should be long-lived, but replaceable at any time. this impacts access control lists; short lifetimes will tend to result in tedious list maintenance or require a namespace infrastructure for central control of access lists 📝
in this document, the namespace that approaches these ideas is called the Host Identity namespace 🦋 using Host Identities requires its own protocol layer, the Host Identity Protocol, between the internetworking and transport layers. the names are based on public-key cryptography to supply authentication services 🔐 properly designed, it can deliver all of the above-stated requirements 👑
4. Host Identity Namespace 🦋
a name in the Host Identity namespace, a Host Identifier (HI), represents a statistically globally unique name for naming any system with an IP stack ✨ this identity is normally associated with, but not limited to, an IP stack. a system can have multiple identities, some 'well known', some unpublished or 'anonymous' 🎭 a system may self-assert its own identity, or it may use a third-party authenticator like DNSSEC [RFC4033], Pretty Good Privacy (PGP), or X.509 to 'notarize' the identity assertion to another namespace 📜
in theory, any name that can claim to be 'statistically globally unique' may serve as a Host Identifier 💫 in the HIP architecture, the public key of a private-public key pair was chosen as the Host Identifier, because it can be self-managed and it is computationally difficult to forge 🔒 as specified in the Host Identity Protocol specification [RFC7401], a public-key-based HI can authenticate the HIP packets and protect them from man-in-the-middle (MitM) attacks 🛡️ since authenticated datagrams are mandatory to provide much of HIP's denial-of-service protection, the Diffie-Hellman exchange in the HIP base exchange has to be authenticated 💪 thus, only public-key HI and authenticated HIP messages are supported in practice 💖
in this document, some non-cryptographic forms of HI and HIP are referenced, but cryptographic forms should be preferred, because they are more secure than their non-cryptographic counterparts 🫶 there has been past research in challenge puzzles using non-cryptographic HI for Radio Frequency IDentification (RFID), in a HIP exchange tailored to the workings of such challenges (as described further in [urien-rfid] and [urien-rfid-draft]) 🏷️
4.1. Host Identifiers 🎀
Okay bestie, so Host Identity gives Internet protocols two main glow-ups 💖 First, it decouples the internetworking and transport layers (see Section 5 🦋). That means the two layers can evolve independently, which is so iconic ✨ It can also give you end-to-end services across multiple internetworking realms. Second glow-up: host authentication 🔥 Since the Host Identifier is a public key, that same key can be used for authentication in security protocols like ESP 🔐
Okay, so an identity in HIP is built on public-private key cryptography 🗝️ The Host Identity is referred to by its public component, the public key, so the name for a Host Identity in the Host Identity namespace (the Host Identifier) is the public key itself. Kind of wild, right? The possession of the private key is basically what defines the Identity 👑 And if more than one node holds that private key, the Identity can be considered a distributed one 🌈
Architecturally, any other Internet naming convention could be a usable base for Host Identifiers 🧚 BUT, non-cryptographic names should only be used in situations of high trust and/or low risk. That's any place where host authentication isn't needed (no risk of host spoofing) and there's no use of ESP. Honestly, at least for interconnected networks spanning several operational domains, the set of environments where host spoofing risk from non-cryptographic Host Identifiers is acceptable is the null set 💀 So the current HIP documents don't specify how to use any other types of Host Identifiers except public keys. For example, Apple's Back to My Mac service [RFC6281] comes pretty close to what HIP does, but unlike HIP, it's based on non-cryptographic identifiers 🙈
The actual Host Identifiers are never directly used at the transport or network layers 🙅♀️ The corresponding Host Identifiers (public keys) may be stored in various DNS or other directories, as identified elsewhere in this document, and they're passed in the HIP base exchange. A Host Identity Tag (HIT) is what other protocols use to represent the Host Identity 🦋 Another representation of Host Identities, the Local Scope Identifier (LSI), can also be used in protocols and APIs 💜
4.2. Host Identity Hash (HIH) 🍓
The Host Identity Hash (HIH) is the cryptographic hash algorithm used to produce the HIT from the HI 🧁 It's also the hash used throughout HIP for consistency and simplicity. The two hosts in a HIP exchange are allowed to use different hash algorithms, which is a fun little detail ✨
Multiple HIHs within HIP are needed to deal with the moving target of cryptographic hashes being created and eventually compromised 😭 Yes, it makes HIP more complicated, and it gives an attacker an extra downgrade attack, but that attack is mitigated in HIP [RFC7401] 🛡️
4.3. Host Identity Tag (HIT) 🌸
A Host Identity Tag (HIT) is a 128-bit representation for a Host Identity 💅 Because it's that size, it fits in the existing sockets API in place of IPv6 addresses (for example, in the sockaddr_in6 structure, sin6_addr member) without modifying applications. Lowkey genius 🫶 It's created from an HIH, an IPv6 prefix [RFC7343], and a hash identifier. Using the HIT instead of the Host Identifier in protocols has two perks. First, its fixed length makes protocol coding easier and helps manage the packet size cost of this technology. Second, it presents the identity in a consistent format to the protocol, no matter which cryptographic algorithms are being used 👑
Essentially, the HIT is a hash over the public key 🔐 Two algorithms affect how a HIT is generated: the public-key algorithm of the HI and the HIH being used. Both algorithms are encoded in the bit presentation of the HIT. The two communicating parties might support different algorithms, so [RFC7401] defines the minimum set for interoperability 🧩 For further interoperability, the Responder may store its keys in DNS records, and then the Initiator may have to pair destination HITs with the appropriate source HITs according to matching HIH 💖
In the HIP packets, HITs identify the sender and recipient of a packet 📦 So a HIT should be unique across the whole IP universe while it's in use. In the extremely rare case that a single HIT maps to more than one Host Identity, the Host Identifiers (public keys) make the final difference 🕵️♀️ And if there's more than one public key for a given node, the HIT acts as a hint for which public key to use.
Although it's rare for an accidental collision to make one HIT map to more than one Host Identity, an attacker might manage to find, by brute force or through an algorithmic weakness, a second Host Identity that hashes to the same HIT 😱 This type of attack is known as a preimage attack. Resistance to finding a second Host Identifier (public key) that hashes to the same HIT is called second preimage resistance. In HIP, second preimage resistance depends on the strength of the hash algorithm and the length of the hash output used. Through HIPv2 [RFC7401], this resistance is 96 bits (less than the 128-bit width of an IPv6 address field, because of the presence of the Overlay Routable Cryptographic Hash Identifiers (ORCHID) prefix [RFC7343]) 🌈 96 bits of resistance was considered acceptable strength during the design of HIP, but it may eventually be considered insufficient for the threat model of an envisioned deployment. One possible mitigation is to pair the use of HITs in the deployment with the HIs themselves (plus mechanisms to securely bind the HIs to the HITs), so that the HI becomes the final authority 🧚. It also may be possible to make a brute force attack harder by making the generation of the HI more computationally difficult, like the hash extension approach of Secure Neighbor Discovery Cryptographically Generated Addresses (CGAs) [RFC3972]. Just note that the HIP specifications through HIPv2 don't provide such a mechanism. Finally, deployments that don't use ORCHIDs (such as certain types of overlay networks) might also use the full 128-bit width of an IPv6 address field for the HIT ✨
4.4. Local Scope Identifier (LSI) 🦋
An LSI is a 32-bit localized representation for a Host Identity 🎀 Because of its size, it fits in the existing sockets API in place of IPv4 addresses (for example, in the sockaddr_in structure, sin_addr member) without modifying applications. The whole point of an LSI is to make Host Identities usable in existing APIs for IPv4-based applications 🧃 LSIs are never sent on the wire. When an application sends data using a pair of LSIs, the HIP layer (or sockets handler) translates the LSIs to the matching HITs, and does the reverse when receiving data 🔄 Besides helping legacy IPv4 applications get HIP-based connectivity, LSIs are useful in two other scenarios [RFC6538] 💅
In the first scenario, two IPv4-only applications live on two separate hosts connected by an IPv6-only network 🌐 With HIP-based connectivity, the two applications can talk to each other even though their application protocol families don't match the underlying network. The reason is that the HIP layer translates the LSIs coming from the upper layers into routable IPv6 locators before the packets hit the wire 🚀
The second scenario is the same as the first, except one of the applications only supports IPv6 😭 Now there are two obstacles in the way: the addressing families of the two applications differ, and the IPv4-only application still can't communicate because of the mismatch between its addressing family (IPv4) and the network's (IPv6). With HIP-based connectivity for applications, this scenario works, because the HIP layer can choose whether to translate the locator of an incoming packet into an LSI or a HIT 🧚
Effectively, LSIs improve IPv6 interoperability at the network layer (as in the first scenario) and at the application layer (as in the second example) 🌸 The interoperability mechanism should not be used to avoid transitioning to IPv6. The authors firmly believe in IPv6 adoption and encourage developers to port existing IPv4-only applications to IPv6 💖 That said, some proprietary, closed-source, IPv4-only applications may never see the daylight of IPv6, and the LSI mechanism is suitable for extending the lifetime of such applications, even in IPv6-only networks 🫶
The main disadvantage of an LSI is its local scope 🙈 Applications may violate layering principles and pass LSIs to each other in application-layer protocols. Since LSIs are only valid in the context of the local host, they may represent an entirely different host when passed to another host 😬 However, it's worth emphasizing that the LSI concept is effectively a host-based NAT, and it doesn't introduce any more issues than the prevalent middlebox-based NATs for IPv4 do. In other words, applications that violate layering principles are already broken by the NAT boxes that are everywhere in deployments 💀
4.5. Storing Host Identifiers in Directories 📚
The public Host Identifiers should be stored in DNS, while the unpublished Host Identifiers should not be stored anywhere (besides the communicating hosts themselves) 🔒 The (public) HI, along with the supported HIHs, is stored in a new Resource Record (RR) type. This RR type is defined in the HIP DNS extension [RFC8005] 🧾
Alternatively, or in addition to storing Host Identifiers in DNS, they may be stored in other directories too 🗂️ For example, a directory based on the Lightweight Directory Access Protocol (LDAP) or a Public Key Infrastructure (PKI) [RFC8002] may be used. Distributed Hash Tables (DHTs) [RFC6537] have also been successfully used [RFC6538] 🌈 This kind of practice may allow them to be used for purposes other than pure host identification ✨
Some types of applications may cache and use Host Identifiers directly, while others may discover them indirectly by looking up a symbolic host name (such as a Fully Qualified Domain Name (FQDN)) in a directory 🧭 Even though Host Identities can have a substantially longer lifetime than routable IP addresses, directories may be a better approach for managing the lifespan of Host Identities 🦋 For example, an LDAP-based directory or DHT can be used for locally published identities, whereas DNS can be more suitable for public advertisement 💜
5. New Stack Architecture ✨
okay so bestie, the easiest way to get what Host Identity is about is to compare the new HI-based setup with the current one 💅. Using the terminology from the IRTF Name Space Research Group Report [nsrg-report] and, e.g., the document on "Endpoints and Endpoint Names" [chiappa-endpoints], IP addresses right now are doing a double shift 🙈. Each IP address names a topological location on the Internet, so it works like a routing direction vector, or locator. At the same time, it names the physical network interface sitting at the point-of-attachment, so it's also acting as an endpoint name.
In HIP, the endpoint names and locators get split apart 🦋. IP addresses keep the locator job. Host Identifiers take over as endpoint identifiers. Important detail: endpoint names based on Host Identities are slightly different from interface names, because a Host Identity can be reachable through several interfaces at once 🌈.
Figure 1 shows how the logical entities are bound differently. The left side is the current TCP/IP architecture and the right side is the HIP-based one.
Transport ---- Socket Transport ------ Socket
association | association |
| |
| |
| |
Endpoint | Endpoint --- Host Identity
\ | |
\ | |
\ | |
\ | |
Location --- IP address Location --- IP address
Figure 1Architecturally, HIP binds transport-layer protocols differently 🎀. The transport-layer associations, meaning TCP connections and UDP associations, are no longer tied to IP addresses. They're tied to Host Identities. In practice, the Host Identities show up as LSIs and HITs for legacy applications and the transport layer, so existing networking APIs and stacks keep working 🫶.
The HIP layer sits logically at Layer 3.5, between the transport and network layers 🧚. It's a shim layer: it handles transport data using LSIs or HITs and leaves other data alone. It translates the two forms of HIP identifiers coming from the transport layer into routable IPv4/IPv6 addresses for the network layer, and does the reverse in the other direction.
5.1. On the Multiplicity of Identities
A host may have multiple identities, on both the client and server side 👑. This raises some extra concerns, which this section addresses.
For security reasons, it may be a bad idea to duplicate the same Host Identity on multiple hosts, because compromising one host taints the identities of the others 💀. Managing machines with identical Host Identities is also a headache in other ways, so it's advisable to give each host its own unique identity 💜.
On the server side, DNS is a better alternative than a shared Host Identity for load balancing 🍓. A single FQDN entry can be configured to refer to multiple Host Identities. Each FQDN entry can be associated with its related locators, or with a single shared locator when the servers use the same HIP rendezvous server (Section 6.3) or HIP relay server (Section 6.4).
Instead of duplicating identities, you can use HIP opportunistic mode ✨. The Initiator leaves out the Responder's identifier when starting the key exchange, then learns it once the exchange completes. There are trade-offs with lowered security guarantees, but the perk is that Host Identifiers don't get published in any directories [komu-leap]. Since many public servers already use DNS as their directory, opportunistic mode may be more suitable for, e.g., peer-to-peer connectivity 🌸. Also worth knowing: opportunistic mode is required in practice when anycast IP addresses are used as locators.
Opportunistic mode can also work with HIP rendezvous or relay servers [komu-diss] 🧚. The Initiator sends an I1 message with a wildcard destination HIT to the locator of a rendezvous/relay server. If that server is serving multiple registered Responders, it can choose the ultimate destination HIT, acting as a HIP-based load balancer. This approach is still experimental though, and it needs further investigation 😭.
On the client side, a host may have multiple Host Identities, for instance for privacy purposes 🙈. Another reason is that the person using the host uses different identities for different administrative domains, as an extra security measure. If a HIP-aware middlebox, like a HIP-based firewall, sits on the path between client and server, the user or the underlying system should carefully pick the correct identity so the firewall doesn't unnecessarily drop HIP-based connectivity [komu-diss].
A server may also have multiple Host Identities 🌈. For instance, one web server might serve several different administrative domains. Usually that's split by DNS name, but a Host Identity could be used for this too. A more compelling reason, though, is a HIP-aware firewall that can't see the HTTP traffic inside the encrypted IPsec tunnel. In that case, each service could get its own identity, letting the firewall segregate the different services on the single web server [lindqvist-enterprise].
6. Control Plane 🎀
HIP decouples the control and data planes from each other ✨. Two end-hosts initialize the control plane with a key exchange called the base exchange. That procedure can be assisted by HIP-specific infrastructure intermediaries called rendezvous or relay servers. If IP addresses change, the end-hosts sustain control plane connectivity using the mobility and multihoming extensions. Eventually, the end-hosts terminate the control plane and remove the associated state.
6.1. Base Exchange 💖
The base exchange is a key exchange that authenticates the Initiator and Responder to each other using their public keys 🔐. Typically the Initiator is the client-side host and the Responder is the server-side host. The roles are used by the HIP implementation's state machine, then discarded upon successful completion.
The exchange is four messages long, and during it the hosts also create symmetric keys to protect the control plane with Hash-based Message Authentication Codes (HMACs) 💞. The keys can also protect the data plane. IPsec ESP [RFC7402] is typically the data plane protocol, although HIP can accommodate others too. Both the control and data planes are terminated with a closing procedure of two messages.
The base exchange also includes a computational puzzle [RFC7401] that the Initiator must solve 🧩. The Responder picks the difficulty of the puzzle, which lets it delay new Initiators according to local policies, for instance when it's under heavy load. The puzzle offers some resiliency against DoS attacks, because its design lets the Responder stay stateless until the very end of the base exchange [aura-dos]. HIP puzzles have also been studied under steady-state DDoS attacks [beal-dos], on multiple adversary models with varying puzzle difficulties [tritilanunt-dos], and with ephemeral Host Identities [komu-mitigation].
6.2. End-Host Mobility and Multihoming 🦋
HIP decouples the transport from the internetworking layer and binds the transport associations to Host Identities (actually through either the HIT or the LSI) 🔗. After the initial key exchange, the HIP layer keeps transport-layer connectivity and data flows going using its extensions for mobility [RFC8046] and multihoming [RFC8047]. So HIP can provide a degree of internetworking mobility and multihoming at a low infrastructure cost 🌸. HIP mobility covers IP address changes (via any method) to either party.
A system is considered mobile if its IP address can change dynamically for any reason, like PPP, DHCP, IPv6 prefix reassignments, or a NAT device remapping its translation 🏃. A system is considered multihomed if it has more than one globally routable IP address at the same time. HIP links IP addresses together when multiple IP addresses correspond to the same Host Identity. If one address becomes unusable, or a more preferred address becomes available, existing transport associations can easily be moved to another address ✨.
When a mobile node moves while communication is ongoing, the address changes are pretty straightforward 🚗. The mobile node sends a HIP UPDATE packet to inform the peer of the new address(es), and the peer then verifies that the mobile node is reachable through these addresses. This way, the peer can avoid flooding attacks, as further discussed in Section 11.2 🛡️.
6.3. Rendezvous Mechanism 🧚
Establishing contact with a mobile, moving node is a bit more involved 🙃. To start the HIP exchange, the Initiator node has to know how to reach the mobile node. For instance, the mobile node can use Dynamic DNS [RFC2136] to update its reachability information in the DNS. To avoid depending on DNS, HIP provides its own alternative: the HIP rendezvous mechanism, as defined in the HIP rendezvous specification [RFC8004].
Using the HIP rendezvous extensions, the mobile node keeps the rendezvous infrastructure continuously updated with its current IP address(es) 📡. The mobile node trusts the rendezvous mechanism to properly maintain its HIT and IP address mappings.
The rendezvous mechanism is especially useful when both nodes are expected to change their address at the same time 💫. In that case, the HIP UPDATE packets will cross each other in the network and never reach the peer node 💔.
6.4. Relay Mechanism 💜
The HIP relay mechanism [RFC9028] is an alternative to the HIP rendezvous mechanism 🌈. It's more suitable for IPv4 networks with NATs, because a HIP relay can forward all control and data plane communications, which guarantees successful NAT traversal 🎯.
6.5. Termination of the Control Plane 👑
The control plane between two hosts is terminated using a secure two-message exchange, as specified in the base exchange specification [RFC7401] 🫶. The related state (host associations) should be removed upon successful termination ✨.
7. Data Plane 🦋
Okay so the data plane is where the actual app traffic goes 📦, and its encapsulation format can be dynamically negotiated during the key exchange ✨. For instance, HICCUPS extensions [RFC6078] define one way to transport application-layer datagrams directly over the HIP control plane, protected by asymmetric key cryptography 🔐. Secure Real-time Transport Protocol (SRTP) has also been considered as the data encapsulation protocol [hip-srtp]. But the most widely implemented method is the Encapsulated Security Payload (ESP) [RFC7402] 💅, which is protected by symmetric keys derived during the key exchange. ESP Security Associations (SAs) offer both confidentiality and integrity protection, and the confidentiality part can be disabled during the key exchange. Other ways of transporting application-layer data may be defined in the future 🔮.
The ESP SAs are established and terminated between the Initiator and the Responder hosts. Usually, the hosts create at least two SAs, one in each direction (Initiator-to-Responder SA and Responder-to-Initiator SA). If the IP addresses of either host changes, the HIP mobility extensions can be used to renegotiate the corresponding SAs 🔄.
On the wire, the big difference in how identifiers are used between the HIP control and data planes is this: the HITs are included in all control packets, but not in the data plane when ESP is employed. Instead, ESP uses Security Parameter Index (SPI) numbers that act as compressed HITs 🗜️. Any HIP-aware middlebox (for instance, a HIP-aware firewall) that's interested in the ESP-based data plane should keep track of the relationship between the control and data plane identifiers so it can associate them with each other 👀.
Since HIP does not negotiate any SA lifetimes, all lifetimes are subject to local policy ⏳. The only lifetimes a HIP implementation must support are sequence number rollover (for replay protection) and SA timeout. An SA times out if no packets are received using that SA. Implementations may support lifetimes for the various ESP transforms and other data plane protocols 🌸.
8. HIP and NATs 🌐
Passing packets between different IP addressing realms means changing the IP addresses in the packet header 🔁. This may happen, for example, when a packet travels between the public Internet and a private address space, or between IPv4 and IPv6 networks. The address translation is usually implemented as Network Address Translation (NAT) [RFC3022] or the historic NAT Protocol Translation (NAT-PT) [RFC2766].
In a network environment where identification is based on IP addresses, identifying the communicating nodes is hard when NATs are employed, because private address spaces overlap 😵💫. In other words, two hosts cannot be told apart solely based on their IP addresses. With HIP, the transport-layer endpoints (i.e., applications) are bound to unique Host Identities instead of overlapping private addresses 🎀. This lets two endpoints tell each other apart even when they're sitting in different private address realms. So the IP addresses are only used for routing purposes and can be changed freely by NATs when a packet between two HIP-capable hosts goes through multiple private address realms 🦋.
NAT traversal extensions for HIP [RFC9028] can be used to set up the actual end-to-end connectivity through NAT devices ✨. For backward compatibility with legacy NATs, the extensions encapsulate both the HIP control and data planes in UDP. They define mechanisms for forwarding the two planes through an intermediary host called a HIP relay, plus procedures for establishing direct end-to-end connectivity by penetrating NATs 🚪. Besides this "native" NAT traversal mode for HIP, other NAT traversal mechanisms have also been successfully used, such as Teredo [RFC4380] (described in further detail in [varjonen-split]) 🧩.
Besides legacy NATs, a HIP-aware NAT has been designed and implemented [ylitalo-spinat] 🤖. For a HIP-based flow, a HIP-aware NAT or HIP-aware historic NAT-PT system tracks the mapping of HITs, and the corresponding ESP SPIs, to an IP address. The NAT system has to learn mappings both from HITs and from SPIs to IP addresses 📚. Many HITs (and SPIs) can map to a single IP address on a NAT, which makes connections on address-poor NAT interfaces much simpler 💜. The NAT can gain much of its knowledge from the HIP packets themselves, though some NAT configuration may be necessary ⚙️.
8.1. HIP and Upper-Layer Checksums
Here's the tea ☕: there's no way for a host to know whether any of the IP addresses in an IP header are the addresses used to calculate the TCP checksum 🙃. It isn't feasible to calculate the TCP checksum using the actual IP addresses in the pseudo header, because the addresses received in an incoming packet aren't necessarily the same as the ones on the sending host. It's also not possible to recompute the upper-layer checksums in the NAT/NAT-PT system, since the traffic is ESP protected 🔒. So the TCP and UDP checksums are calculated using the HITs in place of the IP addresses in the pseudo header. Furthermore, only the IPv6 pseudo header format is used, which provides for IPv4 / IPv6 protocol translation 🌈.
9. Multicast 📣
A bunch of studies looking into HIP-based multicast have been published (including [shields-hip], [zhu-hip], [amir-hip], [kovacshazi-host], and [zhu-secure]) 📚. In particular, so-called Bloom filters, which allow multiple labels to be compressed into small data structures, may be a promising way forward [sarela-bloom] 🌟. However, the different schemes have not been adopted by the HIP working group (nor by the HIP research group in the IRTF), so the details are not elaborated further here 🤷♀️.
10. HIP Policies 📋
There are a number of variables that influence the HIP exchange, and each host has to deal with them 🧠. All HIP implementations should support at least two HIs: one to publish in DNS or a similar directory service, and an unpublished one for anonymous usage (which should be expected to be rotated frequently to disrupt linkability and/or trackability) 🎭. Unpublished HIs will rarely be used as Responder HIs, but they're likely to be common for Initiators 🙈. As stated in [RFC7401]:
all HIP implementations MUST support more than one simultaneous HI, at least one of which SHOULD be reserved for anonymous usage
support for more than two HIs is RECOMMENDED
This creates new challenges for systems and users, who have to decide which type of HI to expose when they start a new session 😬.
Opportunistic mode (where the Initiator starts a HIP exchange without prior knowledge of the Responder's HI) comes with a security trade-off ⚖️. At the expense of being subject to MitM attacks, opportunistic mode lets the Initiator learn the identity of the Responder during communication, rather than from an external directory 🕵️♀️. Opportunistic mode can be used for registration to HIP-based services [RFC8003] (i.e., utilized by HIP for its own internal purposes), or by the application layer [komu-leap]. For security reasons, the application-layer version especially needs some involvement from the user to accept the identity of the Responder, similar to how the Secure Shell (SSH) protocol prompts the user when connecting to a server for the first time [pham-leap] 🔑. In practice, this can be realized in end-host-based firewalls for legacy applications [karvonen-usable], or with native HIP APIs [RFC6317] for HIP-aware applications 🛡️.
As stated in [RFC7401]:
Initiators MAY use a different HI for different Responders to provide basic privacy. Whether such private HIs are used repeatedly with the same Responder, and how long these HIs are used, are decided by local policy and depend on the privacy requirements of the Initiator.
According to [RFC7401]:
Responders that only respond to selected Initiators require an Access Control List (ACL), representing for which hosts they accept HIP base exchanges, and the preferred transport format and local lifetimes. Wildcarding SHOULD be supported for such ACLs, and also for Responders that offer public or anonymous services.
11. Security Considerations 🛡️💀
This section includes discussion on some issues and solutions related to security in the HIP architecture 🔐. Stay tuned, bestie, because this is where things get serious ✨.
11.1. MitM Attacks ✨
okay so HIP takes the Host Identity paradigm and uses it for secure authentication of hosts AND a fast key exchange for ESP 🔐. It also tries to keep hosts from being exposed to various denial-of-service (DoS) and man-in-the-middle (MitM) attacks. BUT here's the plot twist: HIP itself can get hit by its own DoS and MitM attacks, and those could be even more damaging to a host's ability to conduct business as usual 💀.
Resource-exhausting DoS attacks are all about how expensive it is to set up state for a protocol on the Responder compared to the "cheapness" on the Initiator 🙈. HIP lets a Responder make starting state more expensive for the Initiator, and it also tries to make it cheaper for the Responder. How? The Responder starts the authenticated Diffie-Hellman exchange instead of the Initiator, which makes the HIP base exchange four packets long 🦋. The first packet the Responder sends can be prebuilt to cut costs even more. That packet also includes a computational puzzle that can optionally be used to delay the Initiator further, for instance when the Responder is overloaded 🧩. The details are explained in the base exchange specification [RFC7401].
MitM attacks are really hard to defend against without third-party authentication 😭. A skillful MitM could easily handle all parts of the HIP base exchange, but HIP still indirectly gives you this protection: if the Responder's HI is retrieved from a signed DNS zone or securely obtained by some other means, the Initiator can use it to authenticate the signed HIP packets 🔏. Likewise, if the Initiator's HI is in a secure DNS zone, the Responder can retrieve it and validate the signed HIP packets. BUT since an Initiator may choose to use an unpublished HI, it knowingly risks a MitM attack. The Responder may choose not to accept a HIP exchange with an Initiator using an unknown HI 🙅.
Other types of MitM attacks against HIP can be mounted using ICMP messages that signal about problems 📡. As an overall guideline, ICMP messages should be considered as unreliable "hints" and should be acted upon only after timeouts ⏳. The exact attack scenarios and countermeasures are described in full detail in the base exchange specification [RFC7401].
A MitM attacker could try to replay older I1 or R1 messages using weaker cryptographic algorithms, as described in Section 4.1.4 of [RFC7401] 🕵️. The base exchange has been augmented to deal with this attack by restarting when the attack is detected. At worst, the base exchange would never finish (or would be aborted after some retries). The catch is that this leads to a six-way base exchange, which may seem bad at first 😬. But since this only happens in an attack scenario, and the attack can be handled (so it's not interesting to mount anymore), we assume the subsequent messages do not represent a security threat. Since the MitM cannot succeed with a downgrade attack, these sorts of attacks will only occur as 'nuisance' attacks. So the base exchange would still usually be just four packets, even though implementations must be prepared to protect themselves against the downgrade attack 🧁.
In HIP, the Security Association for ESP is indexed by the SPI. The source address is always ignored, and the destination address may be ignored as well. Therefore, HIP-enabled ESP is IP address independent 🧚. This might seem to make attacking easier, but ESP with replay protection is already as well protected as possible, and removing the IP address as a check should not increase the exposure of ESP to DoS attacks 💜.
11.2. Protection against Flooding Attacks 🌊
Although the idea of informing about address changes by simply sending packets with a new source address sounds appealing, it is not secure enough 🙅. Even if HIP does not rely on the source address for anything (once the base exchange has been completed), it appears to be necessary to check a mobile node's reachability at the new address before actually sending any larger amounts of traffic to that address.
Blindly accepting new addresses would potentially lead to flooding DoS attacks against third parties [RFC4225] 💥. In a distributed flooding attack, an attacker opens high-volume HIP connections with a large number of hosts (using unpublished HIs) and then claims to all of them that it has moved to a target node's IP address. If the peer hosts were to simply accept the move, the result would be a packet flood to the target node's address 🌪️. To prevent this type of attack, HIP mobility extensions include a return routability check procedure, where the reachability of a node is separately checked at each address before using that address for larger amounts of traffic ✅.
A credit-based authorization approach for "Host Mobility with the Host Identity Protocol" [RFC8046] can be used between hosts for sending data prior to completing the address tests 💸. Otherwise, if HIP is used between two hosts that fully trust each other, the hosts may optionally decide to skip the address tests. However, such performance optimization must be restricted to peers that are known to be trustworthy and capable of protecting themselves from malicious software 🔐.
11.3. HITs Used in ACLs 📋
At end-hosts, HITs can be used in IP-based access control lists at the application and network layers. At middleboxes, HIP-aware firewalls [lindqvist-enterprise] can use HITs or public keys to control both ingress and egress access to networks or individual hosts, even in the presence of mobile devices, because the HITs and public keys are topology independent 🗺️. As discussed earlier in Section 7, once a HIP session has been established, the SPI value in an ESP packet may be used as an index, indicating the HITs. In practice, firewalls can inspect HIP packets to learn of the bindings between HITs, SPI values, and IP addresses. They can even explicitly control ESP usage, dynamically opening ESP only for specific SPI values and IP addresses 🚪. The signatures in HIP packets allow a capable firewall to ensure that the HIP exchange is indeed occurring between two known hosts. This may increase firewall security 🛡️.
A potential drawback of HITs in ACLs is their 'flatness', which means they cannot be aggregated, and this could potentially result in larger table searches in HIP-aware firewalls 📚. A way to optimize this could be to utilize Bloom filters for grouping HITs [sarela-bloom]. However, it should be noted that it is also easier to exclude individual, misbehaving hosts when the firewall rules concern individual HITs rather than groups 🚫.
There has been considerable bad experience with distributed ACLs that contain material related to public keys, for example, with SSH 😵. If the owner of a key needs to revoke it for any reason, the task of finding all locations where the key is held in an ACL may be impossible. If the reason for the revocation is due to private key theft, this could be a serious issue 🚨.
A host can keep track of all of its partners that might use its HIT in an ACL by logging all remote HITs. It should only be necessary to log Responder hosts. With this information, the host can notify the various hosts about the change to the HIT. There have been attempts to develop a secure method to issue the HIT revocation notice [zhang-revocation] 📝.
Some of the HIP-aware middleboxes, such as firewalls [lindqvist-enterprise] or NATs [ylitalo-spinat], may observe the on-path traffic passively. Such middleboxes are transparent by their nature and may not get a notification when a host moves to a different network. Thus, such middleboxes should maintain soft state and time out when the control and data planes between two HIP end-hosts have been idle too long. Correspondingly, the two end-hosts may send periodically keepalives, such as UPDATE packets or ICMP messages inside the ESP tunnel, to sustain state at the on-path middleboxes 💓.
One general limitation related to end-to-end encryption is that middleboxes may not be able to participate in the protection of data flows 🔒. While the issue may also affect other protocols, Heer et al. [heer-end-host] have analyzed the problem in the context of HIP. More specifically, when ESP is used as the data plane protocol for HIP, the association between the control and data planes is weak and can be exploited under certain assumptions. In the scenario, the attacker has already gained access to the target network protected by a HIP-aware firewall, but wants to circumvent the HIP-based firewall. To achieve this, the attacker passively observes a base exchange between two HIP hosts and later replays it. This way, the attacker manages to get through the firewall and can use a fake ESP tunnel to transport its own data 😈. This is possible because the firewall cannot distinguish when the ESP tunnel is valid. As a solution, HIP-aware middleboxes may participate in the control plane interaction by adding random nonce parameters to the control traffic, which the end-hosts have to sign to guarantee the freshness of the control traffic [heer-midauth]. As an alternative, extensions for transporting the data plane directly over the control plane can be used [RFC6078] 🌸.
11.4. Alternative HI Considerations 🔑
The definition of the Host Identifier states that the HI need not be a public key. It implies that the HI could be any value, for example, a FQDN 🤔. This document does not describe how to support such a non-cryptographic HI, but examples of such protocol variants do exist ([urien-rfid], [urien-rfid-draft]). A non-cryptographic HI would still offer the services of the HIT or LSI for NAT traversal. It would be possible to carry HITs in HIP packets that had neither privacy nor authentication. Such schemes may be employed for resource-constrained devices, such as small sensors operating on battery power, but are not further analyzed here 🔋.
If it is desirable to use HIP in a low-security situation where public key computations are considered expensive, HIP can be used with very short Diffie-Hellman and Host Identity keys. Such use makes the participating hosts vulnerable to MitM and connection hijacking attacks. However, it does not cause flooding dangers, since the address check mechanism relies on the routing system and not on cryptographic strength 🌈.
11.5. Trust on First Use ✨
Okay bestie, let's talk about Trust On First Use (TOFU) 🫶, also known as "Leap of Faith" protocols. [RFC7435] lays out four design principles for these, and they apply to opportunistic HIP too. They're giving main character energy 💅
- Coexist with explicit policy
- Prioritize communication
- Maximize security peer by peer
- No misrepresentation of security
Let's start with the first one. In the RFC's words, "Opportunistic security never displaces or preempts explicit policy". 👑 So if some application data is super sensitive, the policy can require authentication (the public key or certificate) instead of the unauthenticated opportunistic mode. In HIP implementations, this played out like so [RFC6538].
The OpenHIP implementation let an Initiator use opportunistic mode only with an explicitly configured Responder IP address, and only when the Responder's HIT is unknown. On the Responder side, OpenHIP had an option to allow opportunistic mode with any Initiator. Basically, "trust any Initiator" 🙈
HIP for Linux (HIPL) developers went even more fine-grained, with policies at the application level 🧚 They used LD_PRELOAD hooking at the application layer, which let a dynamically linked library intercept socket-related calls without rebuilding the application binaries. The library sat as a shim between the application and transport layers, and it translated the non-HIP-based socket calls from the application into HIP-based socket calls. The shim got a bit complicated, as described in more detail in [komu-leap], but it hit the goal of applying opportunistic mode per individual application. Iconic 🌈
The second TOFU principle basically says communication should be prioritized over security 🍓 So opportunistic mode should generally be allowed even with no authentication present. Falling back to unencrypted communication could even be allowed (if policy permits) instead of blocking the connection. In practice, this works in three steps:
- The HIP Initiator looks up the Responder's HI from a directory such as DNS. If it discovers an HI, it can use that for authentication and skip the rest of the steps.
- If it fails to find an HI, the Initiator can try opportunistic mode with the Responder.
- If opportunistic mode fails, the Initiator can fall back to non-HIP-based communications, if the policy allows it.
This three-step model has been implemented successfully and is described in more detail in [komu-leap]. ✨
The third TOFU principle says security should be maximized, so at least opportunistic security gets used. The three-step model prefers authentication whenever it's available, for example through DNS records (and possibly even DNSSEC when that's available). When no out-of-band credentials exist, it falls back to opportunistic mode. As the very last resort, the model can fall back to non-HIP-based communications if the policy allows it. Also, since perfect forward secrecy (PFS) is explicitly mentioned in the third design principle, it's worth noting that HIP supports it 🔥
The fourth TOFU principle says users and noninteractive applications should be properly informed about the level of security being applied. Non-HIP-aware applications would assume no extra security is being applied, so misleading at least a noninteractive application should not be possible 🙈 For interactive desktop applications, earlier HIP experiments used system-level prompts [karvonen-usable] [RFC6538] to tell the user about the HIP-based security underneath. Overall, users in those experiments could tell when HIP-based security was in use and when it wasn't. 💀 But they failed to notice the difference between opportunistic, non-authenticated HIP and non-opportunistic, authenticated HIP. The reason? Opportunistic HIP (the lower level of security) wasn't clearly indicated in the prompt. That was a valuable lesson for improving the user interface 📚
For HIP-aware applications, the native sockets APIs for HIP specified in [RFC6317] can be used to build application-specific logic instead of relying on generic system-level prompts. The application can then prompt the user directly or handle the situation in other ways. Noninteractive applications can also properly log the level of security being used, since the developer can explicitly program the use of authenticated HIP, opportunistic HIP, and plaintext communication. Developer main character energy 👩💻✨
There are a few more things [RFC7435] discusses that are worth mentioning. For active attacks, HIP has built-in protection against ciphersuite downgrade attacks, as described in detail in [RFC7401] 🛡️ Also, pre-deployed certificates could be used to mitigate active attacks in the opportunistic mode case, as mentioned in [RFC6538].
Detecting peer capabilities comes up in the TOFU context too 🦋 As discussed in this section, the three-step model can be used to detect them. A host can do the first step, discovering a public key, via DNS for instance. If the host finds no keys, it can then try opportunistic mode as the second step. If that times out, the host can proceed to the third step and fall back to non-HIP-based communications if the policy permits. This last step relies on an implicit timeout rather than an explicit (negative) acknowledgment like the one DNS gives, so the user may conclude too early that the connection has failed 😭
To speed up detection by explicitly checking whether the peer supports opportunistic HIP, researchers have proposed TCP-specific extensions [RFC6538] [komu-leap]. The idea is that an Initiator sends an opportunistic I1 packet and a related TCP SYN datagram, equipped with a special TCP option, to a peer at the same time. If the peer supports HIP, it drops the SYN packet and responds with an R1. If the peer is HIP incapable, it drops the HIP packet (and the unknown TCP option) and responds with a TCP SYN-ACK. The benefit of this scheme is a faster fallback to non-HIP-based communications, taking only one round-trip. The drawback is that the approach is tied to TCP. IP options were also considered, but they don't work well with firewalls and NATs. The approach also doesn't work against an active attacker, but opportunistic mode isn't supposed to protect against that kind of adversary anyway 🙅♀️
It's also worth noting that while opportunistic mode has some benefits for incremental deployment, it doesn't deliver all the benefits of authenticated HIP [komu-diss]. Authenticated HIP supports persistent identifiers, which means hosts are identified with the same HI no matter how they move around 🌸 Opportunistic HIP only partially meets this goal. After the first contact between two hosts, HIP can keep connectivity going with its mobility management extensions, but problems pop up when the hosts close the HIP association and try to reconnect. Since hosts can change location, the same IP address isn't guaranteed to belong to the same host anymore. It can be temporarily assigned to different hosts, for example because of IP address reuse (such as by a DHCP service), overlapping private address realms (see also the discussion on Internet transparency in Appendix A.1), or an attempted attack 😬
12. IANA Considerations
This document has no IANA actions. 🌸
13. Changes from RFC 4423
In a nutshell, the changes from RFC 4423 [RFC4423] are mostly editorial 📝 They include clarifications on topics that were described in a difficult way, and omitting some non-architectural (implementation) details that are already described in other documents. A number of missing references to the literature were also added. New topics include the drawbacks of HIP, a discussion on 802.15.4 and MAC security, HIP for IoT scenarios, deployment considerations, and a description of the base exchange. Glow-up era 🦋✨
14. References
14.1. Normative References
[RFC5482] Eggert, L. and F. Gont, "TCP User Timeout Option",
RFC 5482, DOI 10.17487/RFC5482, March 2009,
<https://www.rfc-editor.org/info/rfc5482>.
[RFC6079] Camarillo, G., Nikander, P., Hautakorpi, J., Keranen, A.,
and A. Johnston, "HIP BONE: Host Identity Protocol (HIP)
Based Overlay Networking Environment (BONE)", RFC 6079,
DOI 10.17487/RFC6079, January 2011,
<https://www.rfc-editor.org/info/rfc6079>.
[RFC7086] Keranen, A., Camarillo, G., and J. Maenpaa, "Host Identity
Protocol-Based Overlay Networking Environment (HIP BONE)
Instance Specification for REsource LOcation And Discovery
(RELOAD)", RFC 7086, DOI 10.17487/RFC7086, January 2014,
<https://www.rfc-editor.org/info/rfc7086>.
[RFC7343] Laganier, J. and F. Dupont, "An IPv6 Prefix for Overlay
Routable Cryptographic Hash Identifiers Version 2
(ORCHIDv2)", RFC 7343, DOI 10.17487/RFC7343, September
2014, <https://www.rfc-editor.org/info/rfc7343>.
[RFC7401] Moskowitz, R., Ed., Heer, T., Jokela, P., and T.
Henderson, "Host Identity Protocol Version 2 (HIPv2)",
RFC 7401, DOI 10.17487/RFC7401, April 2015,
<https://www.rfc-editor.org/info/rfc7401>.
[RFC7402] Jokela, P., Moskowitz, R., and J. Melen, "Using the
Encapsulating Security Payload (ESP) Transport Format with
the Host Identity Protocol (HIP)", RFC 7402,
DOI 10.17487/RFC7402, April 2015,
<https://www.rfc-editor.org/info/rfc7402>.
[RFC8002] Heer, T. and S. Varjonen, "Host Identity Protocol
Certificates", RFC 8002, DOI 10.17487/RFC8002, October
2016, <https://www.rfc-editor.org/info/rfc8002>.
[RFC8003] Laganier, J. and L. Eggert, "Host Identity Protocol (HIP)
Registration Extension", RFC 8003, DOI 10.17487/RFC8003,
October 2016, <https://www.rfc-editor.org/info/rfc8003>.
[RFC8004] Laganier, J. and L. Eggert, "Host Identity Protocol (HIP)
Rendezvous Extension", RFC 8004, DOI 10.17487/RFC8004,
October 2016, <https://www.rfc-editor.org/info/rfc8004>.
[RFC8005] Laganier, J., "Host Identity Protocol (HIP) Domain Name
System (DNS) Extension", RFC 8005, DOI 10.17487/RFC8005,
October 2016, <https://www.rfc-editor.org/info/rfc8005>.
[RFC8046] Henderson, T., Ed., Vogt, C., and J. Arkko, "Host Mobility
with the Host Identity Protocol", RFC 8046,
DOI 10.17487/RFC8046, February 2017,
<https://www.rfc-editor.org/info/rfc8046>.
[RFC8047] Henderson, T., Ed., Vogt, C., and J. Arkko, "Host
Multihoming with the Host Identity Protocol", RFC 8047,
DOI 10.17487/RFC8047, February 2017,
<https://www.rfc-editor.org/info/rfc8047>.
[RFC9028] Keränen, A., Melén, J., and M. Komu, Ed., "Native NAT
Traversal Mode for the Host Identity Protocol", RFC 9028,
DOI 10.17487/RFC9028, July 2021,
<https://www.rfc-editor.org/info/rfc9028>.14.2. Informative References
[amir-hip] Amir, K., Forsgren, H., Grahn, K., Karvi, T., and G. Pulkkis, "Security and Trust of Public Key Cryptography for HIP and HIP Multicast", International Journal of Dependable and Trustworthy Information Systems (IJDTIS), Vol. 2, Issue 3, pp. 17-35, DOI 10.4018/jdtis.2011070102, 2013, https://doi.org/10.4018/jdtis.2011070102. 🔐✨
[aura-dos] Aura, T., Nikander, P., and J. Leiwo, "DOS-Resistant Authentication with Client Puzzles", 8th International Workshop on Security Protocols, Security Protocols 2000, Lecture Notes in Computer Science, Vol. 2133, pp. 170-177, Springer, DOI 10.1007/3-540-44810-1_22, September 2001, https://doi.org/10.1007/3-540-44810-1_22. 🧩
[beal-dos] Beal, J. and T. Shepard, "Deamplification of DoS Attacks via Puzzles", October 2004. 🌸
[camarillo-p2psip] Camarillo, G., Mäenpää, J., Keränen, A., and V. Anderson, "Reducing delays related to NAT traversal in P2PSIP session establishments", IEEE Consumer Communications and Networking Conference (CCNC), pp. 549-553, DOI 10.1109/CCNC.2011.5766540, 2011, https://doi.org/10.1109/CCNC.2011.5766540. 🦋
[chiappa-endpoints] Chiappa, J., "Endpoints and Endpoint Names: A Proposed Enhancement to the Internet Architecture", 1999, http://mercury.lcs.mit.edu/~jnc/tech/endpoints.txt. 🧚
[heer-end-host] Heer, T., Hummen, R., Komu, M., Gotz, S., and K. Wehrle, "End-Host Authentication and Authorization for Middleboxes Based on a Cryptographic Namespace", 2009 IEEE International Conference on Communications, DOI 10.1109/ICC.2009.5198984, 2009, https://doi.org/10.1109/ICC.2009.5198984. 💜
[heer-midauth] Heer, T., Ed., Hummen, R., Wehrle, K., and M. Komu, "End-Host Authentication for HIP Middleboxes", Work in Progress, Internet-Draft, draft-heer-hip-middle-auth-04, 31 October 2011, https://datatracker.ietf.org/doc/html/draft-heer-hip-middle-auth-04. 🍓
[henderson-vpls] Henderson, T. R., Venema, S. C., and D. Mattes, "HIP-based Virtual Private LAN Service (HIPLS)", Work in Progress, Internet-Draft, draft-henderson-hip-vpls-11, 3 August 2016, https://datatracker.ietf.org/doc/html/draft-henderson-hip-vpls-11. 🎀
[hip-dex] Moskowitz, R., Ed., Hummen, R., and M. Komu, "HIP Diet EXchange (DEX)", Work in Progress, Internet-Draft, draft-ietf-hip-dex-24, 19 January 2021, https://datatracker.ietf.org/doc/html/draft-ietf-hip-dex-24. 🥗
[hip-lte] Liyanage, M., Kumar, P., Ylianttila, M., and A. Gurtov, "Novel secure VPN architectures for LTE backhaul networks", Security and Communication Networks, Vol. 9, pp. 1198-1215, DOI 10.1002/sec.1411, January 2016, https://doi.org/10.1002/sec.1411. 📡
[hip-srtp] Tschofenig, H., Shanmugam, M., and F. Muenz, "Using SRTP transport format with HIP", Work in Progress, Internet-Draft, draft-tschofenig-hiprg-hip-srtp-02, 25 October 2006, https://datatracker.ietf.org/doc/html/draft-tschofenig-hiprg-hip-srtp-02. 🎧
[hummen] Hummen, R., Hiller, J., Henze, M., and K. Wehrle, "Slimfit - A HIP DEX compression layer for the IP-based Internet of Things", 2013 IEEE 9th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob), pp. 259-266, DOI 10.1109/WiMOB.2013.6673370, October 2013, https://doi.org/10.1109/WiMOB.2013.6673370. 🧃
[IEEE.802.15.4] IEEE, "IEEE Standard for Low-Rate Wireless Networks", IEEE Standard 802.15.4, DOI 10.1109/IEEESTD.2020.9144691, July 2020, https://ieeexplore.ieee.org/document/9144691. 📶
[IEEE.802.15.9] IEEE, "IEEE Draft Recommended Practice for Transport of Key Management Protocol (KMP) Datagrams", IEEE P802.15.9/D04, May 2015. 🗝️
[karvonen-usable] Karvonen, K., Komu, M., and A. Gurtov, "Usable security management with host identity protocol", 2009 IEEE/ACS International Conference on Computer Systems and Applications, pp. 279-286, DOI 10.1109/AICCSA.2009.5069337, 2009, https://doi.org/10.1109/AICCSA.2009.5069337. 🙌
[komu-cloud] Komu, M., Sethi, M., Mallavarapu, R., Oirola, H., Khan, R., and S. Tarkoma, "Secure Networking for Virtual Machines in the Cloud", 2012 IEEE International Conference on Cluster Computing Workshops, pp. 88-96, DOI 10.1109/ClusterW.2012.29, 2012, https://doi.org/10.1109/ClusterW.2012.29. ☁️
[komu-diss] Komu, M., "A Consolidated Namespace for Network Applications, Developers, Administrators and Users", Dissertation, Aalto University, Espoo, Finland, ISBN 978-952-60-4904-5 (printed), ISBN 978-952-60-4905-2 (electronic), December 2012. 🎓
[komu-leap] Komu, M. and J. Lindqvist, "Leap-of-Faith Security is Enough for IP Mobility", 2009 6th IEEE Consumer Communications and Networking Conference, Las Vegas, NV, USA, pp. 1-5, DOI 10.1109/CCNC.2009.4784729, January 2009, https://doi.org/10.1109/CCNC.2009.4784729. 🏃♀️
[komu-mitigation] Komu, M., Tarkoma, S., and A. Lukyanenko, "Mitigation of Unsolicited Traffic Across Domains with Host Identities and Puzzles", 15th Nordic Conference on Secure IT Systems, NordSec 2010, Lecture Notes in Computer Science, Vol. 7127, pp. 33-48, Springer, ISBN 978-3-642-27936-2, DOI 10.1007/978-3-642-27937-9_3, October 2010, https://doi.org/10.1007/978-3-642-27937-9_3. 🚫
[kovacshazi-host] Kovacshazi, Z. and R. Vida, "Host Identity Specific Multicast", International Conference on Networking and Services (ICNS '07), Athens, Greece, pp. 1-1, DOI 10.1109/ICNS.2007.66, 2007, https://doi.org/10.1109/ICNS.2007.66. 📣
[levae-barriers] Levä, T., Komu, M., and S. Luukkainen, "Adoption barriers of network layer protocols: the case of host identity protocol", Computer Networks, Vol. 57, Issue 10, pp. 2218-2232, ISSN 1389-1286, DOI 10.1016/j.comnet.2012.11.024, March 2013, https://doi.org/10.1016/j.comnet.2012.11.024. 🚧
[lindqvist-enterprise] Lindqvist, J., Vehmersalo, E., Komu, M., and J. Manner, "Enterprise Network Packet Filtering for Mobile Cryptographic Identities", International Journal of Handheld Computing Research (IJHCR), Vol. 1, Issue 1, pp. 79-94, DOI 10.4018/jhcr.2010090905, 2010, https://doi.org/10.4018/jhcr.2010090905. 🛡️
[Nik2001] Nikander, P., "Denial-of-Service, Address Ownership, and Early Authentication in the IPv6 World", 9th International Workshop on Security Protocols, Security Protocols 2001, Lecture Notes in Computer Science, Vol. 2467, pp. 12-21, Springer, DOI 10.1007/3-540-45807-7_3, 2002, https://doi.org/10.1007/3-540-45807-7_3. 👑
[nsrg-report] Lear, E. and R. Droms, "What's In A Name: Thoughts from the NSRG", Work in Progress, Internet-Draft, draft-irtf-nsrg-report-10, 22 September 2003, https://datatracker.ietf.org/doc/html/draft-irtf-nsrg-report-10. 🏷️
[paine-hip] Paine, R. H., "Beyond HIP: The End to Hacking As We Know It", BookSurge Publishing, ISBN-10 1439256047, ISBN-13 978-1439256046, 2009. 📚
[pham-leap] Pham, V. and T. Aura, "Security Analysis of Leap-of-Faith Protocols", 7th International ICST Conference, Security and Privacy for Communication Networks, SecureComm 2011, Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, Vol. 96, DOI 10.1007/978-3-642-31909-9_19, 2012, https://doi.org/10.1007/978-3-642-31909-9_19. 🔍
[ranjbar-synaptic] Ranjbar, A., Komu, M., Salmela, P., and T. Aura, "SynAPTIC: Secure and Persistent Connectivity for Containers", 2017 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID), Madrid, 2017, pp. 262-267, DOI 10.1109/CCGRID.2017.62, 2017, https://doi.org/10.1109/CCGRID.2017.62. 📦
[RFC2136] Vixie, P., Ed., Thomson, S., Rekhter, Y., and J. Bound, "Dynamic Updates in the Domain Name System (DNS UPDATE)", RFC 2136, DOI 10.17487/RFC2136, April 1997, https://www.rfc-editor.org/info/rfc2136. 🔄
[RFC2766] Tsirtsis, G. and P. Srisuresh, "Network Address Translation - Protocol Translation (NAT-PT)", RFC 2766, DOI 10.17487/RFC2766, February 2000, https://www.rfc-editor.org/info/rfc2766. 🔀
[RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network Address Translator (Traditional NAT)", RFC 3022, DOI 10.17487/RFC3022, January 2001, https://www.rfc-editor.org/info/rfc3022. 🏠
[RFC3102] Borella, M., Lo, J., Grabelsky, D., and G. Montenegro, "Realm Specific IP: Framework", RFC 3102, DOI 10.17487/RFC3102, October 2001, https://www.rfc-editor.org/info/rfc3102. 🗺️
[RFC3748] Aboba, B., Blunk, L., Vollbrecht, J., Carlson, J., and H. Levkowetz, Ed., "Extensible Authentication Protocol (EAP)", RFC 3748, DOI 10.17487/RFC3748, June 2004, https://www.rfc-editor.org/info/rfc3748. 🪪
[RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", RFC 3972, DOI 10.17487/RFC3972, March 2005, https://www.rfc-editor.org/info/rfc3972. 🧾
[RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S. Rose, "DNS Security Introduction and Requirements", RFC 4033, DOI 10.17487/RFC4033, March 2005, https://www.rfc-editor.org/info/rfc4033. 🛠️
[RFC4225] Nikander, P., Arkko, J., Aura, T., Montenegro, G., and E. Nordmark, "Mobile IP Version 6 Route Optimization Security Design Background", RFC 4225, DOI 10.17487/RFC4225, December 2005, https://www.rfc-editor.org/info/rfc4225. 🧭
[RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through Network Address Translations (NATs)", RFC 4380, DOI 10.17487/RFC4380, February 2006, https://www.rfc-editor.org/info/rfc4380. 🚇
[RFC4423] Moskowitz, R. and P. Nikander, "Host Identity Protocol (HIP) Architecture", RFC 4423, DOI 10.17487/RFC4423, May 2006, https://www.rfc-editor.org/info/rfc4423. 💕
[RFC5218] Thaler, D. and B. Aboba, "What Makes for a Successful Protocol?", RFC 5218, DOI 10.17487/RFC5218, July 2008, https://www.rfc-editor.org/info/rfc5218. 🌈
- [RFC5338] Henderson, T., Nikander, P., and M. Komu, "Using the Host Identity Protocol with Legacy Applications", RFC 5338, DOI 10.17487/RFC5338, September 2008,
https://www.rfc-editor.org/info/rfc5338. 🦋 - [RFC5887] Carpenter, B., Atkinson, R., and H. Flinck, "Renumbering Still Needs Work", RFC 5887, DOI 10.17487/RFC5887, May 2010,
https://www.rfc-editor.org/info/rfc5887. - [RFC6078] Camarillo, G. and J. Melen, "Host Identity Protocol (HIP) Immediate Carriage and Conveyance of Upper-Layer Protocol Signaling (HICCUPS)", RFC 6078, DOI 10.17487/RFC6078, January 2011,
https://www.rfc-editor.org/info/rfc6078. - [RFC6250] Thaler, D., "Evolution of the IP Model", RFC 6250, DOI 10.17487/RFC6250, May 2011,
https://www.rfc-editor.org/info/rfc6250. - [RFC6281] Cheshire, S., Zhu, Z., Wakikawa, R., and L. Zhang, "Understanding Apple's Back to My Mac (BTMM) Service", RFC 6281, DOI 10.17487/RFC6281, June 2011,
https://www.rfc-editor.org/info/rfc6281. 🍓 - [RFC6317] Komu, M. and T. Henderson, "Basic Socket Interface Extensions for the Host Identity Protocol (HIP)", RFC 6317, DOI 10.17487/RFC6317, July 2011,
https://www.rfc-editor.org/info/rfc6317. - [RFC6537] Ahrenholz, J., "Host Identity Protocol Distributed Hash Table Interface", RFC 6537, DOI 10.17487/RFC6537, February 2012,
https://www.rfc-editor.org/info/rfc6537. - [RFC6538] Henderson, T. and A. Gurtov, "The Host Identity Protocol (HIP) Experiment Report", RFC 6538, DOI 10.17487/RFC6538, March 2012,
https://www.rfc-editor.org/info/rfc6538. 🌸 - [RFC7296] Kaufman, C., Hoffman, P., Nir, Y., Eronen, P., and T. Kivinen, "Internet Key Exchange Protocol Version 2 (IKEv2)", STD 79, RFC 7296, DOI 10.17487/RFC7296, October 2014,
https://www.rfc-editor.org/info/rfc7296. - [RFC7435] Dukhovni, V., "Opportunistic Security: Some Protection Most of the Time", RFC 7435, DOI 10.17487/RFC7435, December 2014,
https://www.rfc-editor.org/info/rfc7435. ✨ - [sarela-bloom] Särelä, M., Esteve Rothenberg, C., Zahemszky, A., Nikander, P., and J. Ott, "BloomCasting: Security in Bloom Filter Based Multicast", Information Security Technology for Applications, NordSec 2010, Lecture Notes in Computer Science, Vol. 7127, pages 1-16, Springer, DOI 10.1007/978-3-642-27937-9_1, 2012,
https://doi.org/10.1007/978-3-642-27937-9_1. - [schuetz-intermittent] Schütz, S., Eggert, L., Schmid, S., and M. Brunner, "Protocol enhancements for intermittently connected hosts", ACM SIGCOMM Computer Communication Review, Vol. 35, Issue 3, pp. 5-18, DOI 10.1145/1070873.1070875, July 2005,
https://doi.org/10.1145/1070873.1070875. 💅 - [shields-hip] Shields, C. and J. J. Garcia-Luna-Aceves, "The HIP protocol for hierarchical multicast routing", Proceedings of the seventeenth annual ACM symposium on Principles of distributed computing, pp. 257-266, ISBN 0-89791-977-7, DOI 10.1145/277697.277744, 1998,
https://doi.org/10.1145/277697.277744. - [tempered-networks] Tempered Networks, "Identity-Defined Network (IDN) Architecture: Unified, Secure Networking Made Simple", White Paper, 2016. 🧚
- [tritilanunt-dos] Tritilanunt, S., Boyd, C., Foo, E., and J.M.G. Nieto, "Examining the DoS Resistance of HIP", On the Move to Meaningful Internet Systems 2006: OTM 2006 Workshops, Lecture Notes in Computer Science, Vol. 4277, pp. 616-625, Springer, DOI 10.1007/11915034_85, 2006,
https://doi.org/10.1007/11915034_85. - [urien-rfid] Urien, P., Chabanne, H., Pepin, C., Orga, S., Bouet, M., de Cunha, D.O., Guyot, V., Pujolle, G., Paradinas, P., Gressier, E., and J.-F. Susini, "HIP-based RFID Networking Architecture", 2007 IFIP International Conference on Wireless and Optical Communications Networks, pp. 1-5, DOI 10.1109/WOCN.2007.4284140, 2007,
https://doi.org/10.1109/WOCN.2007.4284140. - [urien-rfid-draft] Urien, P., Lee, G. M., and G. Pujolle, "HIP support for RFIDs", Work in Progress, Internet-Draft, draft-irtf-hiprg-rfid-07, 23 April 2013,
https://datatracker.ietf.org/doc/html/draft-irtf-hiprg-rfid-07. - [varjonen-split] Varjonen, S., Komu, M., and A. Gurtov, "Secure and Efficient IPv4/IPv6 Handovers Using Host-Based Identifier-Location Split", Journal of Communications Software and Systems, Vol. 6, Issue 1, ISSN 18456421, DOI 10.24138/jcomss.v6i1.193, 2010,
https://doi.org/10.24138/jcomss.v6i1.193. 💜 - [xin-hip-lib] Xin, G., "Host Identity Protocol Version 2.5", Master's Thesis, Aalto University, Espoo, Finland, June 2012.
- [ylitalo-diss] Ylitalo, J., "Secure Mobility at Multiple Granularity Levels over Heterogeneous Datacom Networks", Dissertation, Helsinki University of Technology, Espoo, Finland, ISBN 978-951-22-9531-9, 2008.
- [ylitalo-spinat] Ylitalo, J., Salmela, P., and H. Tschofenig, "SPINAT: Integrating IPsec into Overlay Routing", First International Conference on Security and Privacy for Emerging Areas in Communication Networks, SECURECOMM'05, Athens, Greece, pp. 315-326, ISBN 0-7695-2369-2, DOI 10.1109/SECURECOMM.2005.53, 2005,
https://doi.org/10.1109/SECURECOMM.2005.53. 🔥 - [zhang-revocation] Zhang, D., Kuptsov, D., and S. Shen, "Host Identifier Revocation in HIP", Work in Progress, Internet-Draft, draft-irtf-hiprg-revocation-05, 9 March 2012,
https://datatracker.ietf.org/doc/html/draft-irtf-hiprg-revocation-05. - [zhu-hip] Zhu, X., Ding, Z., and X. Wang, "A Multicast Routing Algorithm Applied to HIP-Multicast Model", 2011 International Conference on Network Computing and Information Security, Guilin, China, pp. 169-174, DOI 10.1109/NCIS.2011.42, 2011,
https://doi.org/10.1109/NCIS.2011.42. - [zhu-secure] Zhu, X. and J. W. Atwood, "A Secure Multicast Model for Peer-to-Peer and Access Networks Using the Host Identity Protocol", 2007 4th IEEE Consumer Communications and Networking Conference, Las Vegas, NV, USA, pages 1098-1102, DOI 10.1109/CCNC.2007.221, 2007,
https://doi.org/10.1109/CCNC.2007.221. 👑
Appendix A. Design Considerations ✨
A.1. Benefits of HIP ✨
okay so let's start with the OG rules 🦋. In the beginning, the network layer protocol (aka IP) had four "classic" invariants that everything was built on:
- Non-mutable: The address sent is the address received. 💅
- Non-mobile: The address doesn't change during the course of an "association". 🙈
- Reversible: A return header can always be formed by reversing the source and destination addresses. 🔁
- Omniscient: Each host knows what address a partner host can use to send packets to it. 🧠
Actually, the fourth one can be inferred from 1 and 3, but it's worth mentioning explicitly, for reasons that will be obvious soon if they aren't already 👀.
In the current "post-classic" world, we're intentionally trying to get rid of the second invariant (for both mobility and multihoming 📱), and we've been forced to give up the first and the fourth 💀. Realm Specific IP [RFC3102] is an attempt to bring back the fourth invariant without the first one. IPv6 attempts to bring back the first invariant. Iconic multitasking, honestly 👑.
Few client-side systems on the Internet have DNS names that are actually meaningful 😭. If they do have a Fully Qualified Domain Name (FQDN), that name usually belongs to a NAT device or a dial-up server. It doesn't really identify the system itself, just its current connectivity. FQDNs (and their extensions as email names) are application-layer names, and they're more often naming services than particular systems. This is why so many systems on the Internet aren't registered in the DNS at all: they don't have services of interest to other Internet hosts 🌸.
DNS names are references to IP addresses. This only shows how the networking and application layers are tangled together 🧶. DNS, as the Internet's only deployed and distributed database, is also the home of other namespaces, partly thanks to DNSSEC and application-specific key records. Each namespace can be stretched (IP with v6, DNS with KEY records), but neither one can adequately provide for host authentication or act as a separation between internetworking and transport layers. Not it, bestie 🙅♀️.
The Host Identity (HI) namespace fills an important gap between the IP and DNS namespaces 🧚. The iconic part is that HI lets a host give up all but the 3rd network-layer invariant. As long as the source and destination addresses in the network-layer protocol are reversible, HIP takes care of host identification, and reversibility lets a local host receive a packet back from a remote host. The address changes that happen during NAT transit (non-mutable) or host movement (non-omniscient or non-mobile) can be managed by the HIP layer 🫶.
With the exception of high-performance computing applications, the sockets API is the most common way to build network applications. Apps use it either directly or indirectly, through libraries or frameworks. BUT the sockets API is based on the assumption of static IP addresses, and DNS with its lifetime values was invented at later stages of the Internet's evolution. So the sockets API doesn't deal with the lifetime of addresses [RFC6250]. Most end-user equipment is mobile now, so its addresses are effectively ephemeral, yet the sockets API still gives an unwary developer a fallacious illusion of persistent IP addresses 🙃. HIP can solidify that illusion, because it provides persistent, surrogate addresses to the application layer in the form of LSIs and HITs 💖.
The persistent identifiers HIP provides are useful in multiple scenarios (see, e.g., [ylitalo-diss] or [komu-diss] for a more elaborate discussion) 📚:
- When a mobile host moves physically between two different WLAN networks and obtains a new address, an application using the identifiers stays isolated from the topology changes, while the HIP layer underneath reestablishes connectivity (i.e., a horizontal handoff) 🏃♀️.
- Similarly, the application stays unaware of topological changes when a host with WLAN and cellular interfaces switches between the two access technologies (i.e., a vertical handoff) 📶.
- Even when hosts sit in private address realms, applications can uniquely tell different hosts apart based on their identifiers. In other words, HIP improves Internet transparency for the application layer [komu-diss] 🌈.
- Site renumbering can happen because of corporate mergers or acquisitions, or changes in Internet service provider. It can involve changing the entire network prefix of an organization, which is a problem because of hard-coded addresses in service configuration files or cached IP addresses on the client side [RFC5887]. Considering such human errors, a site using location-independent identifiers as promoted by HIP may have fewer problems while renumbering its network 🧹.
- More agile IPv6 interoperability can be achieved, as discussed in Section 4.4. IPv6-based applications can communicate using HITs with IPv4-based applications that use LSIs. Plus, the underlying network type (IPv4 or IPv6) becomes independent of the addressing family the application uses 🔀.
- HITs (or LSIs) can be used in IP-based access control lists as a more secure replacement for IPv6 addresses. Besides security, HIT-based access control has two other perks. First, using HITs can potentially halve the size of access control lists, because separate rules for IPv4 aren't needed [komu-diss]. Second, HIT-based configuration rules in HIP-aware middleboxes stay static and independent of topology changes, which simplifies administration, especially in mobile environments 🗝️. For instance, HIT-based access control has been harnessed in HIP-aware firewalls, but it can be used directly at the end-hosts as well [RFC6538] 🛡️.
Some of these benefits could be, and have been, redundantly implemented by individual applications. Still, providing this generic functionality at the lower layers is useful, because it reduces software development effort and networking software bugs (since the layer gets tested by multiple applications) 🧪. It also lets the developer focus on building the application itself instead of diving into the intricacies of mobile networking, which makes for nice separation of concerns 💜.
HIP could also be realized by combining a number of different protocols, but the complexity of the resulting software may become substantially larger, and the interaction between multiple, possibly layered, protocols may hurt latency and throughput 😬. It's also worth noting that virtually nothing prevents realizing the HIP architecture as, for instance, an application-layer library, which has actually been implemented in the past [xin-hip-lib] 🧩. The trade-off of moving the HIP layer up to the application layer is that legacy applications may not be supported 🙁.
A.2. Drawbacks of HIP 💅
In computer science, many problems can be solved with an extra layer of indirection. BUT indirection always has costs, because there's no such thing as a "free lunch" 🍓. For HIP, the main costs are:
- In general, an additional layer and a namespace always involve some initial effort in implementation, deployment, and maintenance. Some education of developers and administrators may also be needed. The good news: the HIP community at the IETF has spent years experimenting, exploring, testing, documenting, and implementing HIP to ease adoption costs 🎀.
- HIP introduces a need to manage HIs, and it requires a centralized approach to manage HIP-aware endpoints at scale 🗂️. What used to be IP address-based ACLs are now trusted HITs, and the HIT-to-IP address mappings as well as access policies must be managed. HIP-aware endpoints must also be able to operate autonomously to ensure mobility and availability (an endpoint must be able to run without having a persistent management connection). Users who want this better security and mobility of HIs instead of IP address-based ACLs have to manage this extra 'identity layer' in a nonpersistent fashion 😤. As exemplified in Appendix A.3.5, these challenges have already been solved in an infrastructure setting, where policy is distributed and the mappings and trust relationships between HIP-aware endpoints are managed 🙌.
- HIP decouples the identifier and locator roles of IP addresses. Consequently, a mapping mechanism is needed to associate them together. Failing to map a HIT to its corresponding locator may result in failed connectivity, because a HIT is "flat" by its nature and can't be looked up from the hierarchically organized DNS 🥲. HITs are flat by design, due to a security trade-off: the more bits allocated for the hash in the HIT, the less likely there will be (malicious) collisions 🔐.
- From a performance viewpoint, HIP control and data plane processing introduces some overhead in throughput and latency, as elaborated below ⏱️.
Related to deployment drawbacks: firewalls are commonly used to control access to various services and devices in the current Internet. Since HIP introduces an additional namespace, it's expected that the HIP namespace would also be filtered for unwanted connectivity 🚧. This can be done with existing tools directly on the end-hosts, but filtering at the middleboxes requires modifications to existing firewall software or additional middleboxes [RFC6538] 🫠.
The key exchange adds some extra latency (two round trips) to the initial transport-layer connection establishment between two hosts ⏳. With TCP, additional delay occurs if the underlying network stack implementation drops the triggering SYN packet during the key exchange. The same cost may also occur during HIP handoff procedures. However, subsequent TCP sessions using the same HIP association won't bear this cost (within the key lifetime) ✨. Both the key exchange and handoff penalties can be minimized by caching TCP packets. The latter case can be optimized further with TCP user timeout extensions [RFC5482], as described in more detail by Schütz et al. [schuetz-intermittent] 🧠.
The most CPU-intensive operations involve asymmetric keys and Diffie-Hellman key derivation at the control plane. This only happens during the key exchange, its maintenance (handoffs and refreshing of key material), and teardown procedures of HIP associations 🔥. The data plane is typically implemented with ESP, because it has smaller overhead thanks to symmetric key encryption. Even so, ESP involves some overhead in latency (processing costs) and throughput (tunneling) (see, e.g., [ylitalo-diss] for a performance evaluation) 📊.
A.3. Deployment and Adoption Considerations 🌷
This section describes some deployment and adoption considerations related to HIP from a technical perspective 🦋.
A.3.1. Deployment Analysis ✨
okay so this part is about where HIP has actually shown up in the real world, and bestie, it's giving "underrated but it's been quietly working" 🫶 HIP got adapted and deployed in an industrial control network in a production factory 🏭, where its strong network-layer identity lets the control network live safely alongside a bunch of untrusted devices from third-party vendors [paine-hip]. Lowkey iconic. HIP also got bundled into a security product for Layer 2 VPNs [henderson-vpls] so people could build security zones in a supervisory control and data acquisition (SCADA) network 🔐.
BUT, and this is the tea ☕, HIP has not been a "wild success" [RFC5218] on the Internet, as Levä et al. [levae-barriers] argued 😭. They interviewed 19 experts from industry and academia, so here are some of their findings 💅.
From a marketing perspective, demand for HIP has been low, and substitute technologies have been favored. Another reason is that some technical misconceptions from the early days of the HIP specifications still persist 🙈. Two of those misconceptions are that HIP does not support NAT traversal and that HIP must be implemented in the OS kernel. Both of these claims are untrue 💀. HIP does have NAT traversal extensions [RFC9028], and kernel modifications can be avoided with modern operating systems by diverting packets for userspace processing ✨.
The analysis by Levä et al. also clarifies the infrastructure requirements for HIP 🧚. In a minimal setup, a client and a server machine have to run HIP software. To avoid manual configuration, though, people usually set up DNS records for HIP 🎀. For example, the popular DNS server software Bind9 doesn't need any changes to hold DNS records for HIP, because they can be supported in binary format in its configuration files [RFC6538] 🌸. HIP rendezvous servers and firewalls are optional. No changes are required to network address points, NATs, edge routers, or core networks 🦋. HIP may require holes in legacy firewalls, though 👀.
The analysis also spells out the requirements for the host components, which have three parts 🍓. First, a HIP control plane component is required, and it's typically implemented as a userspace daemon. Second, a data plane component is needed. Most HIP implementations use the so-called Bound End-to-End Tunnel (BEET) mode of ESP, which has been available since Linux kernel 2.6.27, but the BEET mode is also included as a userspace component in a few of the implementations 💜. Third, HIP systems usually provide a DNS proxy for the local host. It translates HIP DNS records to LSIs and HITs, and it passes the corresponding locators to the HIP userspace daemon 🫧. The third component isn't mandatory, but it's super useful for avoiding manual configuration. The three components are described further in the HIP experiment report [RFC6538] 📖.
Based on those interviews, Levä et al. suggest more directions to help HIP get deployed 🚀. Moving a number of HIP specifications to the Standards Track in the IETF has already happened, but the authors suggest other measures too. As a more radical move, they suggest implementing HIP as a purely application-layer library [xin-hip-lib] or some other kind of middleware 🌈. On the other hand, more conservative measures include focusing on private deployments controlled by a single stakeholder. A concrete example of that scenario: HIP could be used by a single service provider to keep its own servers securely connected [komu-cloud] 🔒.
A.3.2. HIP in 802.15.4 Networks ✨
The IEEE 802 standards have been defining MAC-layer security 🛡️. Many of these standards use Extensible Authentication Protocol (EAP) [RFC3748] as a Key Management System (KMS) transport, but some, like IEEE 802.15.4 [IEEE.802.15.4], leave the KMS and its transport as "out of scope" 🤷♀️.
HIP is well suited as a KMS in these environments 💖:
- HIP is independent of IP addressing and can be directly transported over any network protocol 🦋.
- Master keys in 802 protocols are commonly pair-based, with group keys transported from the group controller using pairwise keys 🔑.
- Ad hoc 802 networks can be better served by a peer-to-peer KMS than by the EAP client/server model 🫶.
- Some devices are very memory constrained, and a common KMS for both MAC and IP security represents a considerable code savings 💅.
A.3.3. HIP and Internet of Things ✨
HIP requires a certain amount of computational resources from a device because of cryptographic processing 🧠. HIP scales down to phones and small system-on-chip devices (such as Raspberry Pis and Intel Edison) 📱, but small sensors running on small batteries have remained problematic 🔋. Different extensions to HIP have been developed to scale it down to smaller devices, usually with different security trade-offs ⚖️. For example, non-cryptographic identifiers have been proposed for RFID scenarios 🏷️. The Slimfit approach [hummen] proposes a compression layer for HIP to make it more suitable for constrained networks 🗜️. That approach is applied to a lightweight version of HIP (i.e., "Diet HIP") in order to scale down to small sensors 🌱.
The HIP Diet EXchange (DEX) [hip-dex] design aims to reduce the overhead of the cryptographic primitives it uses by omitting public-key signatures and hash functions 🥗. In doing so, the main goal is to still deliver security properties similar to the Base Exchange (BEX) 🎯.
DEX is primarily designed for computation- or memory-constrained sensor/actuator devices 🤖. Like BEX, it is expected to be used together with a suitable security protocol, such as ESP, for protecting upper-layer protocol data 🛡️. DEX can also be used as a keying mechanism for security primitives at the MAC layer, e.g., for IEEE 802.15.9 networks [IEEE.802.15.9] 🔐.
The main differences between HIP BEX and DEX are 👀:
- Minimum collection of cryptographic primitives to reduce protocol overhead 🧹.
- Static Elliptic Curve Diffie-Hellman (ECDH) key pairs for peer authentication and encryption of the session key 🗝️.
- AES-CTR for symmetric encryption and AES-CMAC for the MACing function 💎.
- A simple fold function for HIT generation 🧩.
- Forfeit of perfect forward secrecy with the dropping of an ephemeral Diffie-Hellman key agreement 😬.
- Forfeit of digital signatures with the removal of a hash function. Reliance on the ECDH-derived key used in HIP_MAC to prove ownership of the private key 🫣.
- Diffie-Hellman derived key ONLY used to protect the HIP packets. A separate secret exchange within the HIP packets creates the session key(s) 🎀.
- Optional retransmission strategy tailored to handle the potentially extensive processing time of the employed cryptographic operations on computationally constrained devices ⏳.
A.3.4. Infrastructure Applications ✨
The HIP experimentation report [RFC6538] lists a bunch of client and server applications that have been trialed with HIP 🧪. Based on that report, this section highlights and builds on some potential ways HIP could be used in existing infrastructure, like routers, gateways, and proxies 🏗️.
HIP has been used successfully with forward web proxies (i.e., client-side proxies) 🌐. HIP ran between a client host (a web browser) and a forward proxy (an Apache server) that terminated the HIP/ESP tunnel. The forward web proxy translated HIP-based traffic from the client into non-HIP traffic towards any web server on the Internet, so the HIP-capable client could talk to HIP-incapable web servers 🙌. This way, the client could use the mobility support HIP provides while using the fixed IP address of the web proxy, for instance to reach services that only allowed the proxy's IP address range 🧚♀️.
HIP with reverse web proxies (i.e., server-side proxies) has also been investigated, as described in more detail in [komu-cloud] 🔄. In this scenario, a HIP-incapable client accessed a HIP-capable web service through an intermediary load balancer (a web-based load balancer implementation called HAProxy). The load balancer translated non-HIP traffic from the client into HIP-based traffic for the web service, which consisted of front-end and back-end servers. Both the load balancer and the web service lived in a data center 🏢. One of the key benefits of encrypting the web traffic with HIP here was supporting a private-public cloud scenario (i.e., hybrid cloud), where the load balancer, front-end servers, and back-end servers were in different data centers. That meant the traffic needed protection as it crossed potentially insecure networks between the private and public clouds ☁️🔐.
HIP could be used to secure access to intermediary devices (e.g., access to switches with legacy telnet) 🔌, and it has also been used to secure intermittent connectivity between middlebox infrastructure 📦. For instance, earlier research [komu-mitigation] used HIP between Simple Mail Transport Protocol (SMTP) servers to exploit the computational puzzles of HIP as a spam mitigation mechanism 📧. A rather obvious practical challenge with this approach was the lack of HIP adoption on existing SMTP servers 😅.
To avoid deployment hurdles with existing infrastructure, HIP could be applied in the context of new protocols with little deployment 🌱. Specifically, HIP has been studied in the context of a new protocol, peer-to-peer SIP [camarillo-p2psip] 📞. The work has resulted in a number of related RFCs [RFC6078], [RFC6079], and [RFC7086] 📚. The key idea was to avoid redundant, time-consuming ICE procedures by grouping different connections (i.e., SIP and media streams) together using the low-layer HIP, which executes NAT traversal procedures only once per host 🪄. An interesting aspect of the approach was using P2P-SIP infrastructure as rendezvous servers for the HIP control plane, instead of the traditional HIP rendezvous services [RFC8004] 🌟.
Researchers have proposed using HIP in cellular networks as a mobility, multihoming, and security solution 📶. [hip-lte] provides a security analysis and simulation measurements of using HIP in Long Term Evolution (LTE) backhaul networks 📊.
HIP has been studied for securing cloud-internal connectivity ☁️. It was first tried with virtual machines [komu-cloud] and then between Linux containers [ranjbar-synaptic] 🐳. In both cases, HIP was suggested as a solution to NAT traversal that could be used both internally by a cloud network and between multi-cloud deployments 🌈. In the former case, HIP was beneficial in sustaining connectivity with a virtual machine while it migrated to a new location 🚚. In the latter case, a Software-Defined Networking (SDN) controller acted as a rendezvous server for HIP-capable containers 🎛️. The controller enforced strong replay protection by adding middlebox nonces [heer-end-host] to the passing HIP base exchange and UPDATE messages 🔒✨
A.3.5. Management of Identities in a Commercial Product 🎀
okay so Tempered Networks makes HIP-based products, and they call their platform Identity-Defined Networking (IDN) [tempered-networks] because HIP is identity-first networking. 🦋 Their goal is to make it simple and nondisruptive to deploy HIP-enabled services widely in production environments. They want transparent device authentication and authorization, cloaking, segmentation, and end-to-end networking. The vibe is getting rid of the circular dependencies, exploits, and layered complexity of traditional "address-defined networking" that blocks mobility and verifiable device access control. 💅 Their portfolio uses HIP like this:
- HIP Switches / Gateways These are physical or virtual appliances that act as the HIP gateway and policy enforcement point for non-HIP-aware applications and devices behind them. No IP or infrastructure changes are required to connect, cloak, and protect the non-HIP-aware devices. Currently known supported platforms for HIP gateways are x86 and ARM chipsets, ESXi, Hyper-V, KVM, AWS, Azure, and Google clouds. ✨
- HIP Relays / Rendezvous These are physical or virtual appliances that act as identity-based routers, authorizing and bridging HIP endpoints without decrypting the HIP session. 🧚 A HIP relay can be deployed as a standalone appliance or in a cluster for horizontal scaling. All HIP-aware endpoints, and the devices they're connecting and protecting, can stay privately addressed. The appliances eliminate IP conflicts, tunnel through NAT and carrier-grade NAT, and require no changes to the underlying infrastructure. The only requirement is that a HIP endpoint should have outbound access to the Internet and that a HIP Relay should have a public address. 🌈
- HIP-Aware Clients and Servers This is software installed in the host's network stack that enforces policy for that host. 🫶 HIP clients support split tunneling. Both the HIP client and HIP server can interface with the local host firewall, and the HIP server can be locked down to listen only on the port used for HIP, which makes the server invisible from unauthorized devices. Currently known supported platforms are Windows, OS X, iOS, Android, Ubuntu, CentOS, and other Linux derivatives. 👑
- Policy Orchestration Managers These physical or virtual appliances are the engine that defines and distributes network and security policy (HI and IP mappings, overlay networks, and whitelist policies, etc.) to HIP-aware endpoints. 🔥 Orchestration does not need to persist to the HIP endpoints, and vice versa, which allows for autonomous host networking and security. 🙈
A.4. Answers to NSRG Questions 🧠
okay bestie, the IRTF Name Space Research Group asked a bunch of evaluating questions in their report [nsrg-report]. Here are our answers, no cap. 💜
- How would a stack name improve the overall functionality of the Internet?
HIP decouples the internetworking layer from the transport layer, so each one can evolve separately. ✨ That decoupling makes end-host mobility and multihoming easier, including across IPv4 and IPv6 networks. HIs make network renumbering easier, and they also make process migration and clustered servers easier to implement. Being cryptographic in nature, they also provide the basis for solving the security problems related to end-host mobility and multihoming. 🔐
- What does a stack name look like?
A HI is a cryptographic public key. 🗝️ However, instead of using the keys directly, most protocols use a fixed-size hash of the public key. 🍓
- What is its lifetime?
HIP provides both stable and temporary Host Identifiers. Stable HIs are typically long-lived, with a lifetime of years or more. ⏳ The lifetime of temporary HIs depends on how long the upper-layer connections and applications need them, and it can range from a few seconds to years. 💀
- Where does it live in the stack?
The HIs live between the transport and internetworking layers. 🦋
- How is it used on the endpoints?
The Host Identifiers may be used directly or indirectly (in the form of HITs or LSIs) by applications when they access network services. 💖 Additionally, the Host Identifiers, as public keys, are used in the built-in key agreement protocol, called the HIP base exchange, to authenticate the hosts to each other. 🤝
- What administrative infrastructure is needed to support it?
In some environments, it is possible to use HIP opportunistically, without any infrastructure. 🌸 However, to gain full benefit from HIP, the HIs must be stored in the DNS or a PKI, and the rendezvous mechanism is needed [RFC8005]. 📚
- If we add an additional layer, would it make the address list in SCTP unnecessary?
Yes 💅
- What additional security benefits would a new naming scheme offer?
HIP reduces dependency on IP addresses, which makes the so-called address ownership [Nik2001] problems easier to solve. In practice, HIP provides security for end-host mobility and multihoming. 🛡️ Furthermore, since HIP Host Identifiers are public keys, standard public key certificate infrastructures can be applied on top of HIP. 🎀
- What would the resolution mechanisms be, or what characteristics of a resolution mechanism would be required?
For most purposes, an approach where DNS names are resolved simultaneously to HIs and IP addresses is sufficient. 🌈 However, if it becomes necessary to resolve HIs into IP addresses or back to DNS names, a flat resolution infrastructure is needed. Such an infrastructure could be based on the ideas of Distributed Hash Tables, but it would require significant new development and deployment. 😭
Acknowledgments 🫶
For the people historically involved in the early stages of HIP, see the Acknowledgments section in the Host Identity Protocol specification. ✨
During the later stages of this document, when the editing baton was transferred to Pekka Nikander, the comments from the early implementers and others were invaluable. 💜 Those people were Jari Arkko, Jeff Ahrenholz, Tom Henderson, Petri Jokela, Miika Komu, Mika Kousa, Andrew McGregor, Jan Melen, Tim Shepard, Jukka Ylitalo, Sasu Tarkoma, and Jorma Wall. Also, the comments from Lars Eggert, Spencer Dawkins, Dave Crocker, and Erik Giesa were also useful. 🙌
The authors want to express their special thanks to Tom Henderson, who took on the burden of editing the document in response to IESG comments at a time when both of the authors were busy doing other things. 👑 Without his perseverance, the original document might never have made it as RFC 4423. 🥹
This main effort to update and move HIP forward within the IETF process owes its impetus to a number of HIP development teams. The authors are grateful to Boeing, Helsinki Institute for Information Technology (HIIT), NomadicLab of Ericsson, and the three universities: RWTH Aachen, Aalto, and University of Helsinki for their efforts. 🎓 Without their collective efforts, HIP would have withered on the IETF vine as a nice concept. 🥀
Thanks also to Suvi Koskinen for her help with proofreading and with the reference jungle. 🌿
Authors' Addresses 💌
Robert Moskowitz (editor) HTT Consulting Oak Park, Michigan United States of America
Email: rgm@labs.htt-consult.com
Miika Komu Ericsson Hirsalantie 11 FI-02420 Jorvas Finland
Email: miika.komu@ericsson.com
📦 where this page comes from
- ietf: rfc9063 2021-07
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms