IRS – just what the SDN Goldilocks is looking for?
Most current SDNish tools are too cumbersome for everyday use: OpenFlow is too granular (the controller interacts directly with the FIB or TCAM), and NETCONF is too coarse (it works on the device configuration level and thus cannot be used to implement anything the networking device can’t already do). In many cases, we’d like an external application to interact with the device’s routing table or routing protocols (similar to tracked static routes available in Cisco IOS, but without the configuration hassle).
Internet Routing System (with an unfortunate acronym) is a new initiative that should provide just what we might need in those cases. The details are still sketchy (after all, this baby is only a few weeks old), but promising. To learn more about IRS, you might want to read the problem statement and framework drafts, view the slides presented at IETF84, or even join the irs-discuss mailing list.
Even if you don’t want to know those details, but consider yourself a person interested in routing and routing protocols, do read two excellent e-mails written by Russ White: in the first one he explained how IRS might appear as yet another routing protocol and benefit from the existing routing-table-related infrastructure (including admin distance and route redistribution), in the second one he described several interesting use cases.
Is IRS the SDN porridge we're looking for? It’s way too early to tell (we need to see more than an initial attempt to define the problem and the framework), but the idea is definitely promising.
- it's being developed in the IETF and is designed in consideration of existing networking standards (and operations)
- it's not limited to low-level FIB programming (the networking equivalent of assembly programming)
- allows a programmable interface into the control plane that isn't limited to configuration (potentially allows programming of various types of networking state)
IRS is basically the IETF trying to tell the industry that "you have to think bigger" than just low-level hardware programming. Exactly /what/ it programs is still up for (active and lively) debate. Note that it's currently just a framework draft(really mostly a problem definition). Just /how/ you communicate to an IRS agent is yet undefined. Likewise, what you can tell it (schema) is also undefined.
It's certainly attractive from the point of view of a networking vendor, as it would open up the kind of abstract programming interface that Juniper's vision outlines (closing the application/network divide).
It will be interesting to see how this plays with the ONF. They could see IRS as competition -- particularly if it starts to include both FIB-level as well as control-plane level state programming (as I suspect it will).