RFC 5902
IAB Thoughts on IPv6 Network Address Translation
RFC 5902: IAB Thoughts on IPv6 Network Address Translation 🦋
okay bestie, the vibe check first! ✨ This one is RFC 5902, titled IAB Thoughts on IPv6 Network Address Translation. It was written by D. Thaler, L. Zhang, and G. Lebovitz 💖 and published by the Internet Architecture Board (IAB) in July 2010 🌸. It's Informational, which means it's not a standards-track spec, and it has ISSN 2070-1721 🧚. The header doesn't list anything it obsoletes or updates, so it's just vibing on its own 💅.
Abstract
so basically people had been having a LOT of debates about whether the IETF should make standards for IPv6 Network Address Translators (NATs) 🔥. This doc breaks down the architectural issues IPv6 NATs bring up, the pros and cons of having them, and the IAB's take on the open questions and the possible solutions 🫶.
Status of This Memo
quick legal-ish moment, I promise it's short 🎀 This doc is not an Internet Standards Track spec. It's published just for information 💅.
This doc is a product of the Internet Architecture Board (IAB), and it holds info the IAB felt was worth keeping on the permanent record ✨. Docs the IAB approves for publication can't be upgraded to any level of Internet Standard, so see Section 2 of RFC 5741 🍓.
For the current status, any errata, and how to send feedback, check http://www.rfc-editor.org/info/rfc5902 🌈.
Copyright Notice
Copyright (c) 2010 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 (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document 🙈.
Table of Contents 🗂️
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. What is the problem? . . . . . . . . . . . . . . . . . . . . . 3
2.1. Avoiding Renumbering . . . . . . . . . . . . . . . . . . . 3
2.2. Site Multihoming . . . . . . . . . . . . . . . . . . . . . 4
2.3. Homogenous Edge Network Configurations . . . . . . . . . . 4
2.4. Network Obfuscation . . . . . . . . . . . . . . . . . . . 5
2.4.1. Hiding Hosts . . . . . . . . . . . . . . . . . . . . . 5
2.4.2. Topology Hiding . . . . . . . . . . . . . . . . . . . 8
2.4.3. Summary Regarding NAT as a Tool for Network
Obfuscation . . . . . . . . . . . . . . . . . . . . . 8
2.5. Simple Security . . . . . . . . . . . . . . . . . . . . . 9
2.6. Discussion . . . . . . . . . . . . . . . . . . . . . . . . 9
3. Architectural Considerations of IPv6 NAT . . . . . . . . . . . 9
4. Solution Space . . . . . . . . . . . . . . . . . . . . . . . . 11
4.1. Discussion . . . . . . . . . . . . . . . . . . . . . . . . 12
5. Security Considerations . . . . . . . . . . . . . . . . . . . 13
6. IAB Members at the Time of Approval . . . . . . . . . . . . . 13
7. Informative References . . . . . . . . . . . . . . . . . . . . 141. Introduction ✨
okay so the IAB has put out a bunch of docs about Internet transparency and the end-to-end principle 💖, and other IETF docs have touched on this stuff too 🌸. These are the big-picture principles the Internet architecture is built on, plus the core values the community wants to protect going forward 🧚. The most recent one, RFC 4924 [RFC4924], reaffirms all of that and reviews the other docs in this area ✨.
Okay, now the drama 🍿. We're running out of IPv4 address space, so there's been a big push for IPv6 deployment 🚀. But since late 2008, people have also been arguing about whether the IETF should standardize network address translation within IPv6 🔥. The people against it say there's no fundamental need for IPv6 NAT, and that as IPv6 keeps rolling out, the Internet should move back toward the end-to-end reachability that made it so successful 🦋. The people for it say NAT vendors would make IPv6 NAT implementations anyway, because NAT can solve a bunch of problems 💅. They also say the IETF should not repeat its IPv4 NAT mistake, where the lack of protocol standards led to lots of different IPv4 NAT implementations, which made NAT traversal really hard 💀.
an earlier doc, [RFC4864], talks about the real or perceived benefits of NAT and suggests alternatives for most of them 🌈. The goal was to show NAT isn't required to get those benefits. It also points out a few gaps that still need filling 🧩.
This doc is the IAB's current take on the debate 🫶. The IAB thinks the issue has to be looked at from an overall architectural standpoint, so we can really judge the pros and cons of IPv6 NAT for the global Internet and where it's headed 🔭.
2. What is the problem? 🤔
Here's the summary of why people want IPv6 NAT. NAT is seen as a way to get a bunch of things individual networks want: avoiding renumbering, making multihoming easier, keeping configurations homogenous, hiding internal network details, and providing simple security 🎀.
2.1. Avoiding Renumbering 🔁
as [RFC4864], Section 2.5 discusses, being able to switch service providers with minimal operational hassle matters a lot in many networks 💼. But renumbering is still super painful today, as [RFC5887] covers 😭. Right now it means reconfiguring every device that deals with IP addresses or prefixes, including DNS servers, DHCP servers, firewalls, IPsec policies, and possibly lots of other systems like intrusion detection systems, inventory management systems, patch management systems, and so on 🧰.
Honestly, renumbering doesn't seem like a big deal in consumer networks, like home networks 🏠. There, addresses or prefixes usually come through DHCP and almost never get configured by hand 💖. In managed networks, though, renumbering can be a seriously big problem 🚨.
Also, lots of large enterprise networks dodge the renumbering problem by using provider-independent (PI) IP address blocks 🏢. Using PI addresses is basically baked into how the Internet runs today ✨. But smaller managed networks that can't get PI address blocks still struggle with renumbering, and it's a serious issue for them 🥺. The Regional Internet Registries (RIRs) keep getting requests for PI address blocks 📬. One main reason they hesitate to hand PI blocks to everyone is worry about how PI addresses would hurt routing system scalability 🌐.
2.2. Site Multihoming 🌐
Another big requirement in many networks is site multihoming 👑. A multihomed site basically needs its IP prefixes to be in the global routing table so it gets the reliability it wants for Internet connectivity, plus load balancing 💅. Today, multihomed sites with PI addresses announce their PI prefixes to the global routing system 📢. Multihomed sites with provider-allocated (PA) addresses also announce the PA prefix they got from one service provider through another service provider 😬. That effectively turns off provider-based prefix aggregation 🙈. This practice makes the global routing table grow linearly with the number of multihomed user networks 📈.
This issue was identified in [RFC4864], Section 6.4 🎀. Sadly, no solution except NAT has been deployed so far that can shield the global routing system from the growing number of multihomed sites 😩. With NAT, a multihomed site just gives multiple IPv4 addresses (one from each of its service providers) to its exit router, which is an IPv4 NAT box 📦. Using address translation for multihoming has one super unique perk: it has no impact on routing system scalability ✨. The NAT box just takes one address from each service provider, and the multihomed site doesn't inject its own routes into the system 🧚. It also seems pretty natural to bring the same idea into IPv6 deployment 🤔. But keep in mind this approach brings all the drawbacks of putting a site behind a NAT box, including losing reachability to servers behind the NAT box 💀.
It's also important to point out that a multihomed site announcing its own prefixes gets two big benefits that NAT-based multihoming doesn't give you 🦋. First, end-to-end communications can be kept alive even when individual service providers have connectivity failures, as long as the site stays connected through at least one working service provider 🔥. Second, announcing your own prefixes lets a multihomed site do traffic engineering and load balancing 💅.
2.3. Homogenous Edge Network Configurations 🏡
Service providers that serve residential customers need to keep support costs down, like not getting flooded with help desk calls 📞. A huge factor in keeping support costs low is making sure customers have homogenous configurations, including the addressing architecture 💖. Today, when a service provider gives out IPv4 NATs, every customer gets the same address space on their home network, so the home gateway always has the same address ✨. From a customer-support point of view, that's probably the most important thing NAT gives us right now 🫶.
In IPv6, link-local addresses can be used to make sure all home gateways have the same address, and to give homogenous addresses to any other devices the service provider supports 🌸. Unlike IPv4, having a globally unique address doesn't stop you from using a homogenous address within the subnet 🧩. IPv6 NAT would only give some homogeneity that link-local addresses wouldn't, and that's in the case of multi-subnet customers 🤓. For multi-subnet customers (like someone using a wireless access point behind the service provider's router/modem), service providers today might only talk about problems, whether IPv4 or IPv6, from computers connected directly to the service provider's router 👀.
it's currently unknown whether IPv6 link-local addresses give enough homogeneity to cut down on help desk calls 🤷. If they don't, providers might still want IPv6 NATs in the residential gateways they provide 🎀.
2.4. Network Obfuscation 🕵️
Most network administrators want to hide the details of their computing resources, information infrastructure, and communications networks inside their borders 🔐. This comes from the basic security principle that an organization's assets are for its sole use, and that all information about those assets, how they operate, and how they get used are proprietary secrets 🤫. Some organizations use their information and communication technologies as a competitive advantage in their industries 👑. It's a generally held belief that measures must be taken to protect those secrets 🛡️. The first layer of protection is preventing access to the secrets, or knowledge about them, whenever possible 🧱. it's totally understandable why network admins would want to keep the details about the hosts on their network, and the network infrastructure itself, private 🙈. They believe NAT helps them reach that goal 💜.
2.4.1. Hiding Hosts ✨
Okay so network admins want to keep literally everything about the machines inside their network a secret 🤫. That includes workstations, laptops, servers, and those function-specific endpoints like printers, scanners, IP telephones, point-of-sale machines, building door access-control devices, and such 💅. They don't want outsiders to be able to count how many hosts are on the network. They also want to stop host fingerprinting, which is basically figuring out what a host is made of, what it contains, or what it does 🕵️. For example, they want to hide whether a host is a user workstation, a finance server, a source code build server, or a printer.
The second part of fingerprint prevention is hiding details that could help an attacker actually break into the host. That covers the type of operating system, its version number, whether it has patches (or doesn't), the make and model of the device hardware, any application software packages loaded, their version numbers and patches, and so on. Once an attacker has that info, they can run a way more focused, targeted attack 😬. So operators want to stop both host counting and host fingerprinting.
When host counting is a concern, it's worth knowing the challenges in stopping it 🧐. [Bellovin] showed how you can successfully count the number of hosts behind a certain type of simple NAT box. More complex NAT setups are better at stopping counting, like Network Address Port Translators (NAPTs) with a pool of public addresses that get randomly bound to internal hosts dynamically when any new connection arrives, without persistency across connections from the same host. BUT the more complex the NAT, the less likely that complex connection types like the Session Initiation Protocol (SIP) [RFC3261] and the Stream Control Transmission Protocol (SCTP) [RFC4960] will be able to get through it 💔. This is the age-old axiom for networked computer systems: for every unit of security you gain, you give up a unit of convenience, and for every unit of convenience you hope to gain, you must give up a unit of security 🔁.
If fields such as fragment ID, TCP initial sequence number, or ephemeral port number are chosen in a predictable way (like sequentially), an attacker can correlate packets or connections coming from the same host 🧩.
Now, you might think, "Just give every transport-layer connection its own IP address!" 🙃 Sadly, that creates other architectural problems. Switches, routers, and even the host's own hardware interface often have a limited amount of state available, and communication using a large number of addresses can run into significant performance problems 🐢. Also, if an attacker can figure out the average number of connections per host, they can still estimate the number of hosts from the connections they observe. So this approach can hurt legitimate communication at all times, just to raise the bar for an attacker. Not it 🙅♀️.
Where host fingerprinting is concerned, even a complex NAT can't prevent it completely 😭. The way different hosts respond to different requests and sequences of events consistently reveals the type of host, its OS, its version number, and sometimes the applications installed. There are even products that do this for network administrators as part of a vulnerability assessment.
These scanning tools open connections of various types across a range of possible IP addresses reachable through that network 🔍. They observe what comes back, then send follow-up messages accordingly until they've "fingerprinted" the host thoroughly. When it's part of a network assessment, these tools are normally run from inside the network, behind the NAT. But if the tool sits outside the network boundary along the path of packets (like in an external vulnerability assessment or penetration test) and passively observes and records connection exchanges, it can fingerprint hosts over time only if it can work out which externally viewed connections come from the same internal host 👀. If the NATing is simple and static, with each host's internal address always mapped to the same external address and vice versa, the tool has 100% success fingerprinting the host. Once the internal hosts are mapped to their external IP addresses and fingerprinted, the attacker can launch targeted attacks against them or reliably attempt to hijack their connections 💀. If the NAT uses a single external IP, or a pool of dynamically assigned IP addresses for each host, but does so in a deterministic and predictable way, fingerprinting is more complex but still quite achievable.
If the NAT uses dynamically assigned addresses with short-term persistency, but nothing about it can be learned from outside, the problem gets harder for the attacker 🙌. The observer may be able to fingerprint a host during the lifetime of a particular IP address mapping, and across connections. But once that mapping is terminated, the observer doesn't immediately know which new mapping belongs to the same host. After lots of observation and correlation, the attacker might sometimes figure out whether an observed new connection in flight comes from a familiar host. With that info, plus a good set of man-in-the-middle attack tools, the attacker could try to compromise the host by hijacking a new connection of adequately long duration. If temporal persistency isn't deployed on the NAT, this tactic becomes almost impossible 🙅. As the difficulty and cost of the attack go up, the number of attackers willing to try it goes down 📉. And the attacker can't initiate a connection toward a host whose current IP address binding they don't know. So they're limited to hijacking observed connections they think come from a familiar host, or blindly launching attacks on connections in flight. This is why network administrators appreciate complex NATs' ability to deter host counting and fingerprinting, but that deterrence comes at a cost to host reachability 🥲.
2.4.2. Topology Hiding 🗺️
It's perceived that a network operator may want to hide the details of the network topology, the size of the network, the identities of the internal routers, and how the routers are interconnected 🧠. This desire has been discussed in [RFC4864], Sections 4.4 and 6.2.
However, how well topology hiding works depends on the complexity, dynamism, and pervasiveness of the bindings the NAT employs (all of which were described above) 📊. The more complex the NAT, the more the topology gets hidden, but the less likely complex connection types will successfully traverse the NAT barrier. So the trade-off is reachability across applications ⚖️.
Even if you can hide the actual addresses of internal hosts through address translation, that doesn't necessarily hide the internal topology 🫣. Some topological info can be inferred from passively observing packets. For example, based on packet timing, delay measurements, the Hop Limit field, or other fields in the packet header, you could infer the relative distance between multiple hosts. Once an observed session is believed to match a previously fingerprinted host, that host's distance from the NAT device may be learned, but not its exact location or particular internal subnet 📏.
Host fingerprinting is required in order to do a thorough distance mapping 🔬. An attacker might then use message contents to lump certain types of devices into logical clusters and take educated guesses at attacks. That's not, however, a thorough mapping 🙄. Some NATs change the TTL hop counts, much like an application-layer proxy would, while others don't. This is an administrative setting on more advanced NATs. The simpler and more static the NAT, the more possible this is. The more complex, dynamic, and non-persistent the NAT bindings, the more difficult it gets 💅.
2.4.3. Summary Regarding NAT as a Tool for Network Obfuscation 📝
The degree of obfuscation a NAT can achieve is a function of its complexity, which is measured by:
- The use of one-to-many NAPT mappings;
- The randomness over time of the mappings from internal to external IP addresses, i.e., non-deterministic mappings from an outsider's perspective;
- The lack of persistence of mappings, i.e., the shortness of mapping lifetimes and not using the same mapping repeatedly;
- The use of re-writing in IP header fields such as TTL.
But deployers, be warned 🚨: as obfuscation increases, host reachability decreases. Mechanisms such as STUN [RFC5389] and Teredo [RFC4380] fail with the more complex NAT mechanisms 😩.
2.5. Simple Security 🔐
It's commonly perceived that a NAT box provides one level of protection, because external hosts can't directly initiate communication with hosts behind a NAT 🙈. However, you should not confuse NAT boxes with firewalls! As discussed in [RFC4864], Section 2.2, the act of translation does not provide security in itself. The stateful filtering function can provide the same level of protection without requiring a translation function. For further discussion, see [RFC4864], Section 4.2 💖.
2.6. Discussion 💬
At present, the primary benefits you may get from deploying NAT appear to be avoiding renumbering, facilitating multihoming without impacting routing scalability, and making edge consumer network configurations homogenous 🌸.
Network obfuscation (host hiding, both counting and fingerprinting prevention, and topology hiding) may well be achieved with more complex NATs, but at the cost of losing some reachability and application success 🥺. Again, when it comes to security, this is often the case: to gain security, one must give up some measure of convenience 🔒✨.
3. Architectural Considerations of IPv6 NAT ✨
okay bestie, first things first: a NAT box and a firewall are not the same thing 💅 A firewall's whole job is blocking unwanted traffic [RFC4948] without messing with the wanted stuff, but a NAT box also interferes with wanted traffic 🙈 So for the rest of this section, "reachability" means reachability of wanted traffic, fr fr.
The IPv6 NAT talk keeps pointing at how widely IPv4 NAT is deployed, where people have found real benefits and built up real operational experience 🎀 But so far the conversation is mostly about the maybe-benefits IPv6 NAT could or could not bring. Nobody's really looking at the bigger picture, and that's what we're going to dig into 🔍
When you think about what IPv6 NAT could do for a site that uses it, don't forget the huge question: if one site deploys IPv6 NAT, what happens to the rest of the Internet that doesn't? By "the rest of the Internet" we mean the community that builds, deploys, and uses end-to-end applications and protocols, which means they get hit by any loss of transparency (see [RFC2993] and [RFC4924] for more). This big question doesn't seem to have been addressed, or at least not properly 😭
We think the IPv6 NAT discussion should happen inside the context of the whole Internet architecture 🌸 The biggest question isn't how many perks you can get from IPv6 NAT. It's more basic than that: are enough parties on the Internet actually willing to deploy IPv6 NAT, and do we want IP address translation to become a permanent building block of the Internet architecture?
You could argue that the answers depend on whether we find good solutions for renumbering, site multihoming, and edge network configuration, and on whether those solutions keep transparency or not 🦋 If transparency isn't kept, making NAT a permanent building block would be a fundamental architectural change, and that's a whole thing.
It's really desirable that IPv6 users and apps can reach each other directly without worrying about address translation boxes sitting between the two ends 🫶 App developers should be able to build around end-to-end reachability (of wanted traffic) without having to figure out how to get through NAT boxes. Referrals and multi-party conversations are simple with end-to-end addressing, but get super messy once address translation is involved 💀 Network admins should also get to run their networks without the extra complexity of NATs, which can cost more than just the price of extra boxes. They can also make network monitoring and problem debugging way harder 🍓
Given how diverse Internet users are and how varied today's operational practices are, it's totally possible that some parties really want to deploy IPv6 NAT. The Internet should make room for different views that lead to different practices (some using IPv6 NAT, others not) 🌈
If we accept that some, but not all, parties want IPv6 NAT, then the real debate isn't about what benefits IPv6 NAT might give the people who deploy it. It's undeniable that network address translation can bring certain benefits to its users 💜 The real challenge is figuring out how to design IPv6 NAT so its impact stays hidden inside some localized scope ✨ If we can pull that off, the Internet as a whole can strive for (re-installing) the end-to-end reachability model 👑
4. Solution Space
From an end-to-end perspective, the solution space for renumbering and multihoming breaks down into three classes 🧚
- Endpoints get a stable, globally reachable address: In this class, end sites use provider-independent addressing, so endpoints aren't affected by changing service providers 💖 For this to be a complete solution, provider-independent addressing has to be available to all managed networks (every network that uses manual configuration of addresses or prefixes in any type of system). But today, giving provider-independent addresses to every network, even tiny ones, raises concerns about the scalability of the global routing system 😬 That's an area of ongoing research and experimentation. In practice, network admins have also been developing short-term approaches to close today's gap between continued routing table growth and the limits of existing router capacity [NANOG].
- Endpoints get a stable but non-globally-routable address on physical interfaces but a dynamic, globally routable address inside a tunnel: Hosts use locally-scoped (and so provider-independent) addresses for communication within the site over their physical interfaces 🎀 That means managed systems like routers and DHCP servers all see stable addresses. To communicate externally, the host tunnels to some infrastructure device. The tunnel gives the host globally routable addresses that may change, but address changes stay limited to systems that operate over or beyond the tunnel, including DNS servers and applications. These systems are often already able to handle changes today using things like DNS dynamic update ✨ However, if the endpoints and the tunnel infrastructure devices belong to different organizations, these solutions are harder to deploy incrementally because of the incentive and coordination issues involved.
- Endpoints get a stable address that gets translated in the network: End sites use non-globally-routable addresses inside the site and translate them to globally routable addresses somewhere in the network 🦋 In general this causes loss of end-to-end transparency, which is the subject of [RFC4924] and the documents it surveys. If the translation is reversible, and it actually gets reversed by the time it reaches the other end of the communication, then end-to-end transparency can be provided. But if the two translators involved belong to different organizations, these solutions are harder to deploy incrementally because of the incentive and coordination issues involved 😭
Routing scalability is next 🧠 Although there's no immediate danger, it's been a longtime concern in operational communities, and an effective, deployable solution has to be found. The real question isn't whether some parties can run NAT. It's whether the Internet as a whole would be willing to rely on NAT to curtail the routing scalability problem, and whether we've investigated all the potential impacts to understand its cost on the overall architecture. If effective solutions can be deployed in time to let every user community get provider-independent IPv6 addresses, the Internet can dodge the complexity, fragility, and other unforeseen problems that NAT brings 🌟
4.1. Discussion
As [RFC4924] states:
A network that does not filter or transform the data that it carries may be said to be "transparent" or "oblivious" to the content of packets. Networks that provide oblivious transport enable the deployment of new services without requiring changes to the core. It is this flexibility that is perhaps both the Internet's most essential characteristic as well as one of the most important contributors to its success.
We think end-to-end transparency, as defined above, is key to the success of the Internet 💜 Some fields (e.g., Hop Limit) are defined to be mutable, but transparency requires that fields not defined that way arrive un-transformed. Right now, the source and destination addresses are defined as immutable fields, and lots of protocols and applications rely on them being that way 🔥
Each of the three classes of solution can be defined in a way that preserves end-to-end transparency 🎀
Honestly, we don't consider IPv6 NATs to be desirable 🙈 But we get that some deployment of them is likely, unless workable solutions for avoiding renumbering, facilitating multihoming without adversely impacting routing scalability, and homogeneity are generally recognized as useful and appropriate.
So we strongly encourage the community to treat end-to-end transparency as a requirement when proposing any solution, whether it's based on tunneling, translation, or some other technique 👑 Then solutions can be compared on other aspects, like scalability and ease of deployment ✨
5. Security Considerations
okay serious moment 😭 Section 2 discusses potential privacy concerns as part of the Host Counting and Topology Hiding problems 🔒
6. IAB Members at the Time of Approval
- Marcelo Bagnulo
- Gonzalo Camarillo
- Stuart Cheshire
- Vijay Gill
- Russ Housley
- John Klensin
- Olaf Kolkman
- Gregory Lebovitz
- Andrew Malis
- Danny McPherson
- David Oran
- Jon Peterson
- Dave Thaler
7. Informative References
[Bellovin] Bellovin, S., "A Technique for Counting NATted Hosts",
Proc. Second Internet Measurement Workshop,
November 2002,
<http://www.cs.columbia.edu/~smb/papers/fnat.pdf>.
[NANOG] "Extending the Life of Layer 3 Switches in a 256k+ Route
World", NANOG 44, October 2008, <http://www.nanog.org/
meetings/nanog44/presentations/Monday/
Roisman_lightning.pdf>.
[RFC2993] Hain, T., "Architectural Implications of NAT", RFC 2993,
November 2000.
[RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston,
A., Peterson, J., Sparks, R., Handley, M., and E.
Schooler, "SIP: Session Initiation Protocol", RFC 3261,
June 2002.
[RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through
Network Address Translations (NATs)", RFC 4380,
February 2006.
[RFC4864] Van de Velde, G., Hain, T., Droms, R., Carpenter, B., and
E. Klein, "Local Network Protection for IPv6", RFC 4864,
May 2007.
[RFC4924] Aboba, B. and E. Davies, "Reflections on Internet
Transparency", RFC 4924, July 2007.
[RFC4948] Andersson, L., Davies, E., and L. Zhang, "Report from the
IAB workshop on Unwanted Traffic March 9-10, 2006",
RFC 4948, August 2007.
[RFC4960] Stewart, R., "Stream Control Transmission Protocol",
RFC 4960, September 2007.
[RFC5389] Rosenberg, J., Mahy, R., Matthews, P., and D. Wing,
"Session Traversal Utilities for NAT (STUN)", RFC 5389,
October 2008.
[RFC5887] Carpenter, B., Atkinson, R., and H. Flinck, "Renumbering
Still Needs Work", RFC 5887, May 2010.Authors' Addresses ✨
okay bestie, last stop! 🧚 these are the people who wrote the doc, and their contact info is right here if you ever need to reach out 💌
Dave Thaler 💖
Dave Thaler
Microsoft Corporation
One Microsoft Way
Redmond, WA 98052
USA
Phone: +1 425 703 8835
EMail: dthaler@microsoft.comLixia Zhang 🦋
Lixia Zhang
UCLA Computer Science Department
3713 Boelter Hall
Los Angeles, CA 90095
USA
Phone: +1 310 825 2695
EMail: lixia@cs.ucla.eduGregory Lebovitz 🌸
Gregory Lebovitz
Juniper Networks, Inc.
1194 North Mathilda Ave.
Sunnyvale, CA 94089
USA
EMail: gregory.ietf@gmail.comAnd the group that approved it, the Internet Architecture Board 👑 can be reached at iab@iab.org 🫶
That's the whole doc, bestie! 🌈✨ Iconic work reading all of it 💅
📦 where this page comes from
- ietf: rfc5902 2010-07
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms