RFC 1092

EGP and Policy Based Routing in the New NSFNET Backbone 🎀

Okay bestie, let's talk about RFC 1092 🌸. It's from the Network Working Group, written by J. Rekhter (Jacob Rekhter in the address block) at the T. J. Watson Research Center, and dated February 1989 ✨. The original doesn't list a category, and it doesn't say it obsoletes or updates anything, so there's nothing to chase there 💅.

Status of This Memo

Okay so this memo covers implementation decisions for routing issues in the NSFNET, especially the NSFNET Backbone 🦋. The big concern is restricting routing info so only the best route, as set by a policy decision, gets advertized 👑. Distribution of this memo is unlimited 🌈.

Introduction

The NSFNET backbone basically routes packets between the Regionals Networks it's connected to. So a packet that arrives at a backbone entry node gets routed to an exit node 🚀. Two things decide how packets travel through the network:

• the NSFNET backbone routing protocol/algorithm, and
• additional information about the externally connected networks.

This paper is about how reachability info gets exchanged between the external networks and the NSFNET backbone, so packets can reach the right destination using a reasonable path 🧭.

EGP as reachability protocol

The EGP (Exterior Gateway Protocol) routing method is going to swap reachability info between the NSFNET backbone and the regional networks 🧚. But ngl, there are a bunch of problems with using EGP as a reachability protocol in a meshed environment 💀. Some EGP components need extra definitions for backbone-regional network interactions. It's also worth noting that EGP is only viewed as an interim measure until better inter autonomous system protocols get defined and widely deployed for gateways used by regional networks 🫶.

Here's the list of some EGP problems and issues:

• The EGP model assumes an engineered spanning tree topology. The NSFNET doesn't fit that model 🙈, because of backdoor routes. In the NSFNET, the same network may be advertized as reachable by more than one regional network. Beyond the NSFNET not fitting a spanning tree, there are serious concerns about the "core" concept (central to EGP) and its obvious deficiencies 😭.
• EGP is going to isolate intra-Regional routing from intra-NSFNET-Backbone routing, but it doesn't address false info that regional networks might supply. EGP by itself doesn't protect a particular network from unwanted and unsolicited representation by some regional network 🚫. For example, if network N1 is reachable through regional network R1 and also through regional network R2, EGP has no way to specify one path as primary and one as secondary. That's because there isn't a generally accepted interpretation of EGP metrics today. Also, nothing in EGP can stop one or more regional networks from advertizing other networks (especially ones that belong to other regional networks) as reachable with zero distance. That could create a "black hole", or at least suboptimal IP routing 🕳️.
• EGP by itself has no provisions to guarantee that routes through the NSFNET Backbone will be preferred over routes through the backdoor routers, or vice versa 🙅.

Policy Based Routing

Looking at the problems above, new factors like autonomy and mutual trust become obvious 👑. While we chase the routing functionality the new NSFNET backbone needs, we should realize that one of our primary concerns has to be accommodating those factors 💅.

So some kind of rudimentary Policy Based Routing method becomes imperative 😤. We want to emphasize, though, that we're not talking about complete Policy Based Routing. We're more concerned with supporting a minimum subset of policy functionality as an initial solution to the problems above. This requires support and cooperation between the management of each network connected to the NSFNET backbone 🤝.

We need to support a particular network N, which belongs to one of the regional networks, establishing a bilateral agreement with one or more regional networks. The agreement is of the type "network N can be reached via one or more regional networks (RN1, RN2, ... RNx)". This lets each network select one or more representatives at the regional network level 🎀. Once the agreement is established, the info will be available to:

• The network which initiated the agreement.
• The management of the regional network(s) with whom this agreement has been established.
• The NSFNET backbone Network Operation Center, where it will be entered into the Routing Policy Data Base, which will be available through the NSFNET information services.

Supporting multiple routes to the NSFNET core requires a guarantee that for a certain network N, no regional network other than the one(s) selected by N will advertize N as reachable. That means the NSFNET core will ignore unauthorized advertisements for network N 🚨.

EGP and Rudimentary Policy Based Routing

Each network that belongs to the NSFNET will select a specific regional network as its primary representative to the NSFNET core 🌟. It does this through a bilateral agreement with the management of that same regional network, as well as the NSFNET backbone management. The same network can also select an arbitrary number of other regional networks as their secondary, tertiary, etc. representatives. It does that by establishing bilateral agreements with the management of those regional networks and the NSFNET backbone management 🧩.

Reachability info supplied by each regional network gets distributed to all other NSS nodes of the NSFNET Backbone. We want to emphasize that we're not going to flood EGP packets inside the backbone. Instead, we use the learned info for the interior gateway protocol, which uses the ANSI IS-IS protocol 🌀.

The implementation allows a defined regional network to advertize a particular leaf network in EGP NR packets with a distance of zero. Secondary representatives may advertize the same network with distance one or higher. If the path through the primary regional representative is available, all secondary paths get ignored. If the path through the primary regional representative goes down (which gets discovered via the EGP NR info), the next path with the lowest available EGP metric gets used 🔁.

We'll also be able to detect and report unsolicited representations 🔍. This happens by examining, on a periodic basis, all reachability info obtained via EGP. The result gets compared against the Routing Policy Data Base, which holds info about all bilateral agreements between networks and their regional representatives. Any mismatch causes an alarm to the Network Operations Center 🚨. For example, network N established a bilateral agreement with regional network R1, electing it as its primary representative. The EGP NR record received from regional network R5 advertizes network N as reachable with distance zero. By comparing the Routing Policy Data Base entry for network N with that EGP NR record, a mismatch gets detected, and an alarm is forwarded to the Network Operation Center 📣.

Since the whole scheme is based on a combination of the network number and the autonomous system number, it's also important to insure the correctness of the autonomous system numbers advertized by the regional networks to the NSFNET core, to allow for further verification 🧐.

The autonomous system number validation for each regional network will be performed at the NSS that connects the particular leaf network to the NSFNET backbone. All discrepancies will be reported to the Network Operations Center 📝.

The NSFNET backbone will be considered a separate Autonomous System with its own autonomous system number 👑.

Backbone versus Backdoor Routes

Sometimes regional networks prefer paths through some backdoor route over paths through the NSFNET backbone 🚪. So the reachability info that the NSFNET core advertizes to regional networks (via EGP NR records) will always use a fixed metric of 128 for all routes. This may aid in encouraging traffic to flow through backdoors, if desired and available 🦋.

Regional networks can use a variety of techniques to determine how they route traffic for any particular network, at their own option 💖.

Heads up: the fixed metric of 128 is used for all routes the NSFNET core advertizes to regional networks, every time.

What do we expect from the Regional Networks

Each regional network should get its own Autonomous System number 💜.

The connection between regional networks and the NSFNET backbone will be done via EGP. It's the responsibility of the regional backbone to provide EGP functionality through the attachment to the E-PSP dedicated to the regional network 🔌.

EGP functionality may require translating network numbers in and out of the regional network. In any case, the NSFNET backbone expects individual network numbers for the leaf networks of the regional network, as long as they should be advertised. It will also announce individual networks known to the NSFNET core to the regional network 📡.

EGP support should include the ability to configure EGP metrics from some statically definable configuration table. If EGP metrics can't be defined, or if they aren't fixed, the metric determination will be done by the NSFNET backbone routers themselves, as taken from their databases. In that case, it's the responsibility of the regional network to provide the NSFNET backbone management with the metric data, so the metrics can be used properly 🧮.

We also expect each regional network to handle all bilateral agreements with its leaf networks regarding Policy Based Routing. It should also supply a copy of those agreements to the NSFNET backbone management 📑.

Acknowledgements

Big thanks to Barry Appelman (T.J. Watson Research Center, IBM Corp.) and Hans-Werner Braun (Merit) for their contributions to this document 🫶.

Author's Address 💌

Jacob Rekhter
T.J. Watson Research Center
IBM Corporation
P.O. Box 218
Yorktown Heights, NY 10598

Phone: (914) 945-3896

Email: YAKOV@IBM.COM