RFC 2908
The Internet Multicast Address Allocation Architecture
RFC 2908 💖 The Internet Multicast Address Allocation Architecture
okay bestie, so this one is RFC 2908 🎀 and it's called The Internet Multicast Address Allocation Architecture. It's from the Network Working Group and it's dated September 2000. The authors are D. Thaler (Microsoft), M. Handley (ACIRI), and D. Estrin (ISI) 👑. It's an Informational doc, which means it gives info to the Internet community and doesn't specify any kind of Internet standard. The header doesn't list anything it obsoletes or updates ✨.
Status of this Memo 🧚
This memo gives info to the Internet community. It doesn't specify an Internet standard of any kind, and distribution of it is unlimited 💅.
Copyright Notice 🌸
Copyright (C) The Internet Society (2000). All Rights Reserved.
Abstract 🦋
This doc proposes a multicast address allocation architecture (MALLOC) for the Internet, and it's giving modular. It has three layers: a host-server mechanism, an intra-domain server-server coordination mechanism, and an inter-domain mechanism 🫶.
Table of Contents 🍓
1: Introduction ................................................ 2
2: Requirements ................................................ 2
3.1: Address Dynamics .......................................... 4
3: Overview of the Architecture ................................ 5
4: Scoping ..................................................... 7
4.1: Allocation Scope .......................................... 8
4.1.1: The IPv4 Allocation Scope -- 239.251.0.0/16 ............. 9
4.1.2: The IPv6 Allocation Scope -- SCOP 6 ..................... 9
5: Overview of the Allocation Process .......................... 9
6: Security Considerations ..................................... 10
7: Acknowledgments ............................................. 11
8: References .................................................. 11
9: Authors' Addresses .......................................... 12
10: Full Copyright Statement ................................... 131. Introduction ✨
This doc proposes the MALLOC architecture for the Internet, and it's meant to be generic enough to work for both IPv4 and IPv6 environments 🌈.
Just like unicast addresses, the way any multicast address gets used is limited in two dimensions, and they're both kind of a big deal:
- Lifetime: An address has a start time and a (possibly infinite) end time, and it's only valid in between 🦋.
- Scope: An address is valid over a specific area of the network. For example, it might be globally valid and unique, or it might be a private address that's only valid within a local area 🌍.
This architecture assumes that the main scoping mechanism in use is administrative scoping, as described in RFC 2365 [1]. Solutions that work for TTL scoping are possible, but they add a lot of extra complication for address allocation [2]. Honestly, TTL scoping is a poor solution for multicast scope control 💀, and the assumption here is that people will stop using TTL scoping before this architecture is widely used 🙈.
2. Requirements 💜
From a design point of view, the important properties of multicast allocation mechanisms are robustness, timeliness, a low probability of clashing allocations, and good address space utilization in situations where space is scarce. When this interacts with multicast routing, it's desirable for multicast addresses to be allocated in a way that helps aggregate routing state 🔥.
- Robustness/Availability
The robustness requirement is that an application that needs an address should always be able to get one, even when there are other network failures going on 🫶.
- Timeliness
From a timeliness point of view, a short delay of up to a few seconds is probably fine before the client gets an address it can have reasonable confidence in being unique ⏱️. If the session is defined in advance, the address should be allocated as soon as possible, and it should not wait until just before the session starts. It's sometimes okay to change the multicast addresses the session uses right up until it actually starts, but this should only be done when it avoids a significant problem, like an address clash that was discovered after the initial session definition 😭.
- Low Probability of Clashes
A multicast address allocation scheme should always be able to allocate an address that can be guaranteed not to clash with another session's address ✨. A top-down partitioning of the address space would be required to completely guarantee that no clashes would ever occur 🧩.
- Address Space Packing in Scarcity Situations
When address space is scarce, simply partitioning it would cause a lot of fragmentation 🍰. That's because each partition would need enough spare space to give a reasonable degree of assurance that addresses could still be allocated for a significant time if the network got partitioned. On top of that, providing backup allocation servers in a hierarchy like this, so that fail-over (including partitioning a server and its backup away from each other) doesn't cause collisions, would add even more fragmentation 😩.
Since guaranteeing no clashes in a robust way requires partitioning the address space, a hard guarantee leads to inefficient address space usage. So when address space is scarce, it's hard to get constant availability and timeliness, guarantee no clashes, and get good address space usage all at once. That means we must prioritize these properties 👑. When address space is scarce, we believe that good address space packing and constant availability matter more than guaranteeing that address clashes never happen 💅. What we aim for in these situations is a very high probability that an address clash doesn't occur, but we accept that there's a finite probability of it happening anyway.
Should a clash happen (or should an application start using an address it didn't allocate, which may also lead to a clash), either the clash can be detected and the addresses changed, or hosts that receive the extra traffic can prune it using source-specific prunes available in IGMP version 3. So we don't think this is a disastrous situation 🙌. Tolerating the possibility of clashes is likely to allow allocation of a very high proportion of the address space, even with network conditions like those observed in [3].
We believe we can get good packing and good availability with good collision avoidance, whereas avoiding all collisions would mean compromising packing and availability significantly 🌸.
Finally, in situations where address space is not scarce, such as with IPv6, good address space usage is less important. So partitioning may potentially be used to guarantee no collisions among hosts that use this architecture 🧚.
2.1. Address Dynamics 🦋
Multicast addresses may be allocated in any of three ways 🌈:
Static: Statically allocated addresses are allocated by IANA for specific protocols that need well-known addresses to work. Examples are 224.0.1.1, which is used for the Network Time Protocol [13], and 224.2.127.255, which is used for global scope multicast session announcements 📡. Applications that use multicast for bootstrap purposes should not normally be given their own static multicast address. Instead, they should bootstrap themselves using a well-known service location address that can be used to announce the binding between local services and multicast addresses.
Static addresses typically have a permanent lifetime, and a scope defined by the scope range they live in. So a static address is valid everywhere (although the set of receivers may be different depending on location), and it may be hard-coded into applications, devices, embedded systems, and so on 🤖. Static addresses are also useful for devices that support sending but not receiving multicast IP datagrams (Level 1 conformance as specified in RFC 1112 [7]), or even for devices that can't receive any data at all, like a wireless broadcasting device 📻.
Scope-relative: RFC 2365 [1] reserves the highest 256 addresses in every administrative scope range for relative assignments. Relative assignments are made by IANA and consist of an offset that's valid in every scope. Relative addresses are reserved for infrastructure protocols that need an address in every scope, and this offset may be hard-coded into applications, devices, embedded systems, and so on 🧸. Such devices must have a way (e.g. via MZAP [9] or via MADCAP [4]) to obtain the list of scopes they reside in 🗺️.
The offsets that get assigned typically have a permanent lifetime, and they're valid in every scope and location. So the scope-relative address in a given scope range has a lifetime equal to that of the scope range it falls in 💫.
Dynamic: For most purposes, the correct way to use multicast is to get a dynamic multicast address 💖. These addresses are provided on demand and have a specific lifetime. An application should only request an address for as long as it expects to need it ⏳. In some circumstances, an address will be granted for a shorter period than was requested. This will happen rarely if the request is for a reasonable amount of time, but applications should be prepared to cope with it when it does 😅.
At any time during the lifetime of an existing address, applications may also request an extension of that lifetime, and such extensions will be granted when possible 🎀. When the extension isn't granted, the application is expected to request a new address to take over from the old one when it expires, and to be able to cope with this situation gracefully 🫶. As with unicast addresses, the network doesn't guarantee reachability of an address once its lifetime expires 🚫.
These restrictions on address lifetime are necessary so the address allocation architecture can be organized around address usage patterns, which keeps addresses aggregatable and makes multicast routing reasonably close to optimal 🏆. Statically allocated addresses, in contrast, may be given sub-optimal routing 🐢.
3. Overview of the Architecture ✨
okay so the architecture is super modular, bestie 🧩 Each layer can be used, upgraded, or swapped out totally on its own. Layering also gives you isolation, so different mechanisms at the same layer can be used by different orgs without messing up the other layers. it's giving clean separation 💅
There are three layers (see Figure 1 🦋). Quick heads up: these layer numbers are NOT the same as the layer numbers in the TCP/IP stack, which describe the path of data packets. don't mix them up 🙈
+--------------------------+ +------------------------+
| | | |
| to other peers | | to other peers |
| || // | | || // || |
| Prefix | | Prefix Prefix |
| Coordinator | |Coordinator Coordinator|
+------------||------------+ +-------||----//---------+
||Layer 3 || //
+------------||------------------------------||--//-----------+
| Prefix Prefix |
| Coordinator=======================Coordinator |
| ^ ^ |
| +----------------+-------------+ |
| | Layer 2 | | |
| MAAS<---/ | +---> MAAS |
| ^ ^ v ^ |
| . . MAAS . |
| . .Layer 1 ^ .Layer 1 |
| v v .Layer 1 v |
| Client Client v Client |
| Client |
+-------------------------------------------------------------+Figure 1: An Overview of the Multicast Address Allocation Architecture
Layer 1 🎀 This is the protocol or mechanism a multicast client uses to request a multicast address from a multicast address allocation server (MAAS). When the server grants an address, it becomes the server's responsibility to make sure that address isn't reused elsewhere within the address's scope during the lifetime that was granted. main character server energy 👑
Examples of possible protocols or mechanisms at this layer include MADCAP [4], HTTP to access a web page for allocation, and IANA static address assignments.
There's also an abstract API for applications to use for dynamic allocation, and it's independent of whichever Layer 1 protocol or mechanism is in use. It's given in [11] 🧚
Layer 2 🌸 This is an intra-domain protocol or mechanism that MAAS's use to coordinate allocations so they don't hand out duplicate addresses. A MAAS must have stable storage, or some equivalent robustness mechanism, so that uniqueness is preserved across MAAS failures and reboots. no cap, this part is serious 💀
MAASs also use the Layer 2 protocol or mechanism to acquire (from "Prefix Coordinators") the ranges of multicast addresses they're allowed to allocate from.
In this document, "allocation domain" means an administratively scoped, multicast-capable region of the network, where addresses in a specific range may be allocated by a Layer 2 protocol or mechanism. 🦋
Examples of protocols or mechanisms at this layer include AAP [5], and manual configuration of MAAS's.
Layer 3 🌈 This is an inter-domain protocol or mechanism that allocates multicast address ranges (with lifetimes) to Prefix Coordinators. Individual addresses then get allocated out of these ranges by MAAS's inside the allocation domains, as described above.
Examples of protocols or mechanisms at this layer include MASC [6] (where Prefix Coordinators are typically routers without any stable storage requirement), and static allocations by AS number as described in [10] (where Prefix Coordinators are typically human administrators).
Each of the three layers serves a slightly different purpose, so protocols or mechanisms at each layer may need different design tradeoffs. it's all about the tradeoffs, bestie ⭐
4. Scoping 🔍
okay so, to allocate dynamic addresses within administrative scopes, a MAAS has to be able to learn which scopes are in effect, what their address ranges and names are, and which addresses or subranges within each scope are valid for dynamic allocation by that MAAS. it's a whole info-gathering era 📚
The first two tasks, learning which scopes are in effect and the address range and name(s) of each scope, can come from static configuration or be dynamically learned. For example, a MAAS can just passively listen to MZAP [9] messages to pick up this info. lowkey the easiest option 🫶
To work out the subrange for dynamic allocation, there are two cases for each scope, depending on whether it's a small "indivisible" scope or a big "divisible" scope. Note that MZAP is the one that tells you which scopes are divisible and which aren't 💖
- Small scopes. The allocation domain is the entire topology within the administrative scope. So all MAASs inside the scope may use the entire address range (minus the last 256 addresses reserved as scope-relative addresses), and they use the Layer 2 mechanism or protocol to coordinate allocations. Prefix Coordinators aren't involved at all for small scopes 🙅♀️
Because of that, the effective "allocation domain" area can be different for different scopes. Also, a small, indivisible scope could be larger or smaller than the Allocation Scope used for big scopes (more on that below) 🌸
- Big scopes (including the global scope). The area inside the scope may be large enough that simply using a Layer 2 mechanism or protocol would be inefficient or otherwise undesirable. In this case the scope must span multiple allocation domains, and the Layer 3 mechanism or protocol must be used to divvy up the scoped address space among those allocation domains. So a MAAS can learn about the scope via MZAP, but it has to acquire a subrange to allocate from from a Prefix Coordinator 🎀
For simplicity, the effective "allocation domain" area is the same for all big scopes. It's the granularity at which all big scopes get divided up, and that administrative scope at this granularity is called the "Allocation Scope" ✨
4.1. Allocation Scope 🦋
The Allocation Scope is a brand-new administrative scope, defined in this document and to be reserved by IANA with the values noted below. It's the scope a Layer 2 protocol or mechanism uses to coordinate address allocation for addresses in larger, divisible scopes. the main character of the scoping drama 🎬
We expect the Allocation Scope will often line up with a unicast Autonomous System (AS) boundary.
If an AS is too large, or the network administrator wants to run different intra-domain multicast routing in different parts of an AS, that AS can be split by manually setting up an allocation scope boundary that isn't an AS boundary. This is done by setting up a multicast boundary that divides the unicast AS into two or more multicast allocation domains. splitting the squad 🍓
If an AS is too small and address space is scarce, address space fragmentation may happen if the AS is its own allocation domain. In that case the AS can instead be treated as part of its provider's allocation domain, and it can use a Layer 2 protocol or mechanism to coordinate allocation between its MAAS's (if any) and those of its provider. An AS should probably take this course of action if:
- it is connected to a single provider,
- it does not provide transit for another AS, and
- it needs fewer than (say) 256 multicast addresses of larger than AS scope allocated on average.
4.1.1. The IPv4 Allocation Scope -- 239.251.0.0/16 🌈
The address space 239.251.0.0/16 is to be reserved for the Allocation Scope. The ranges 239.248.0.0/16, 239.249.0.0/16 and 239.250.0.0/16 are to be left unassigned and available for expansion of this space. Those ranges should be left unassigned until the 239.251.0.0/16 space is no longer sufficient. save the extra space for later, bestie 🧸
4.1.2. The IPv6 Allocation Scope -- SCOP 6 💜
The IPv6 "scop" value 6 is to be used for the Allocation Scope. one value, one job 💅
5. Overview of the Allocation Process 🌸
okay so once Layer 3 allocation has been done for large, divisible scopes, and each Prefix Coordinator has acquired one or more ranges, those ranges get passed to all MAAS's within that Prefix Coordinator's domain via a Layer 2 mechanism or protocol. it's a whole relay race 🏃♀️
MAAS's in the domain receive these ranges and store them as the currently allowable addresses for that domain. Each range is valid for a given lifetime (also acquired via the Layer 3 mechanism or protocol) and isn't revoked before that lifetime has expired. MAAS's also learn about small scopes (e.g., via MZAP) and store the ranges associated with them. 📚
Using the Layer 2 mechanism or protocol, each MAAS makes sure it excludes any addresses that have been, or will be, allocated by other MAAS's within its domain. no double-booking, periodt 🙅♀️
When a client needs a multicast address, it first has to decide what the scope of the intended session should be, and then locate a MAAS that can allocate addresses within that scope. 🧚
To pick a scope, the client will either just choose a well-known scope, such as the global scope, or it will enumerate the available scopes (e.g., by sending a MADCAP query, or by listening to MZAP messages over time) and let a user select one. 🦋
Locating a MAAS can be done in a bunch of ways, including manual configuration, using a service location protocol such as SLP [12], or via a mechanism provided by a Layer 1 protocol itself. MADCAP, for instance, includes such a facility. ✨
Once the client has chosen a scope and located a MAAS, it requests an address in that scope from that MAAS. Along with the request, it also passes the acceptable range for the lifetimes of the allocation it wants. For example, if the Layer 1 protocol in use is MADCAP, the client sends a MADCAP REQUEST message to the MAAS and waits for either a NAK message or an ACK message containing the allocated information. 💌
When the MAAS gets a request from a client, it chooses an unused address in a range for the specified scope, with a lifetime that both satisfies the acceptable range the client specified and falls within the lifetime of the actual range. precision is everything 🎯
The MAAS uses the Layer 2 mechanism or protocol to make sure that address doesn't clash with any addresses allocated by other MAASs. For example, if Layer 2 uses manual configuration of non-overlapping ranges, this just means sticking to the range configured in the local MAAS. If, on the other hand, AAP is used at Layer 2 to get less address space fragmentation, the MAAS advertises the proposed allocation domain-wide using AAP. If no clashing AAP claim is received within a short time interval, the address is returned to the client via the Layer 1 protocol or mechanism. If a clashing claim is received by the MAAS, it chooses a different address and tries again. AAP also lets each MAAS pre-reserve a small "pool" of addresses that it doesn't have to wait on to detect clashes for. pool party 🏊♀️
If a domain ever starts running out of available multicast addresses, a Prefix Coordinator in that domain uses the Layer 3 protocol or mechanism to acquire more space. 🫶
6. Security Considerations 🔐💅
okay bestie, this section is lowkey the most important one 😭 so let's get into it 🧚
The architecture described here does not stop an application from just sending to or joining a multicast address without allocating it first 🙈 (same deal as unicast addresses today, no cap). BUT there is no guarantee that data for unallocated addresses will actually get delivered by the network 💀 Routers may drop data for unallocated addresses if they have some way of checking whether a destination address has been allocated. For example, if the border routers of a domain take part in the Layer 2 protocol/mechanism and cache the set of allocated addresses, then data for unallocated addresses in a range allocated by that domain can be dropped by creating multicast forwarding state with an empty outgoing interface list and/or pruning back the tree branches for those groups 🌸
A malicious application might try a denial-of-service attack 😈 by trying to allocate a huge number of addresses, so it exhausts the supply of available ones. Other attacks include releasing or modifying the allocation of another party 🙅♀️ These attacks can be combatted through authentication with policy restrictions (like a maximum number of addresses that a single party can allocate) 🫶
So, hence 💖 protocols and mechanisms that implement layers of this architecture should be deployable in a secure fashion ✨ For example, one should support authentication with policy restrictions, and should not allow someone unauthorized to release or modify the allocation of another party 👑 Periodt.
7. Acknowledgments 🎀
Steve Hanna gave valuable feedback on this document 🫶 The members of the MALLOC WG and the MBone community provided the motivation for this work 🌈
8. References 📚
[1] Meyer, D., "Administratively Scoped IP Multicast", BCP 23, RFC 2365, July 1998.
[2] Mark Handley, "Multicast Session Directories and Address Allocation", Chapter 6 of PhD Thesis entitled "On Scalable Multimedia Conferencing Systems", University of London, 1997.
[3] Mark Handley, "An Analysis of Mbone Performance", Chapter 4 of PhD Thesis entitled "On Scalable Multimedia Conferencing Systems", University of London, 1997.
[4] Hanna, S., Patel, B. and M. Shah, "Multicast Address Dynamic Client Allocation Protocol (MADCAP)", RFC 2730, December 1999.
[5] Handley, M. and S. Hanna, "Multicast Address Allocation Protocol (AAP)", Work in Progress.
[6] Estrin, D., Govindan, R., Handley, M., Kumar, S., Radoslavov, P. and D. Thaler, "The Multicast Address-Set Claim (MASC) Protocol", RFC 2909, September 2000.
[7] Deering, S., "Host Extensions for IP Multicasting", STD 5, RFC 1112, August 1989.
[8] Rekhter, Y. and T. Li, "A Border Gateway Protocol 4 (BGP-4)", RFC 1771, March 1995.
[9] Handley, M., Thaler, D. and R. Kermode, "Multicast-Scope Zone Announcement Protocol (MZAP)", RFC 2776, February 2000.
[10] Meyer, D. and P. Lothberg, "GLOP Addressing in 233/8", RFC 2770, February 2000.
[11] Finlayson, R., "Abstract API for Multicast Address Allocation", RFC 2771, February 2000.
[12] Guttman, E., Perkins, C., Veizades, J. and M. Day, "Service Location Protocol, Version 2", RFC 2608, June 1999.
[13] Mills, D., "Network Time Protocol (Version 3) Specification, Implementation and Analysis", RFC 1305, March 1992.
9. Authors' Addresses 🏡
Dave Thaler Microsoft Corporation One Microsoft Way Redmond, WA 98052-6399
EMail: dthaler@microsoft.com
Mark Handley AT&T Center for Internet Research at ICSI 1947 Center St, Suite 600 Berkeley, CA 94704
EMail: mjh@aciri.org
Deborah Estrin Computer Science Dept/ISI University of Southern California Los Angeles, CA 90089
EMail: estrin@usc.edu
10. Full Copyright Statement ⚖️
Copyright (C) The Internet Society (2000). All Rights Reserved.
This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English.
The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Acknowledgement 🌷
Funding for the RFC Editor function is currently provided by the Internet Society 💕
📦 where this page comes from
- ietf: rfc2908 2000-09
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms