In yet another sign that there’s a LOT of SDN maneuvering going on, Brocade has announced the acquisition of open-source routing/switching software platform provider Vyatta. Sometimes seen by the media as a Cisco rival, Vyatta has in fact not been a serious threat to Cisco or any other equipment vendor—until lately. With the advent of the notion of SDN and the additional driver of operator interest in offloading network functionality from devices into servers (Network Functions Virtualization or NFV) there’s suddenly a lot of good reasons to like what Vyatta could provide, and maybe like Brocade for buying them.
The big question now is whether Brocade has any of those good reasons for having done the deal. Brocade does have a cloud and SDN goal, like pretty much everyone in networking these days, and it also has the Foundry product lines which have been moderately successful and have also managed to gain some traction in data center networking. It’s not at all unreasonable to assume that Brocade sees Vyatta as its Nicira, a virtual networking play to shore up its cloud and data center strategy. And it might well do that, too. Vyatta has a broad spectrum of features, surpassing simple virtual networking and extending well into the spectrum of features that have been recently added (in the Folsom release) to Quantum. In fact, Vyatta makes a darn good framework for a complete Quantum play. It would be a help, but not a revolution, and not enough to change Brocade’s fortunes in my view.
The NFV part is Brocade’s chance for revolution. Recall that the goal of NFV is to host on servers what would otherwise have been embedded in custom appliances. Address assignment, firewalls and security, packet inspection—you name it. That’s almost exactly what Vyatta does. And SDN requires path computation, topology, addressing, and all that good stuff to be added to OpenFlow. That’s what Vyatta does too. So if Brocade adds in its own OpenFlow (which it has already committed to) and frames the Vyatta assets in the higher SDN layers, then it has not only what could well be the industry’s first full-on, desktop-to-data-center SDN, it would also have an operative version of the NFV vision. It could force bigger vendors like Alcatel-Lucent, Cisco, and Juniper to choose between actively developing an architecture that they’d see as undermining their incumbent products or letting Brocade run off with the customers. Either one could help Brocade a lot.
But this is a BIG stretch for Brocade in terms of bold initiatives. The company has been challenged in positioning its Foundry assets even in their conventional terms, and while it’s told an SDN story, the story to date hasn’t been anything more than others have articulated. Now they’re confronting not only full-on SDN, but hosted network functionality that makes the cloud INTO the network and not just a customer for it. Can they rise to that grand a vision, or will it scare everyone in management into immobility?
For Brocade, that’s the question; for the industry it may not matter. Look at Vyatta’s Network OS and you see what almost seems a point-by-point list of the functionality that the NFV operators want migrated from appliances into hosts, which is exactly where Vyatta could put it. Who could miss that connection? Which means that there’s a good chance that even if Brocade doesn’t take the Big Step and perhaps remake themselves totally (and for the better) they’ll drive someone else to do that. Not any of the Big Three in routing that I’ve named, but perhaps either another major player (Ericsson or NSN) or a smaller one (Extreme, Aruba, Riverbed, even Palo Alto). There’s a ton of open-source network code out there, derived in some cases from the early UNIX stuff. There’s also the IP of some failed equipment vendors. Anyone who likes cloud and likes SDN and likes money has got to see this as an opportunity.
This model fits into the vision of edge equipment change I blogged about yesterday, of course. You can visualize all these network features as being cloud-hosted and pushed out by my “omnipus” to basic edge devices who use OpenFlow-like forwarding stubs to link users to the features just as they could link them to services or applications. That, of course, may be the biggest barrier to vendors taking the bold path with something like Vyatta. How many, even the Brocades, would be willing to toss out the model of the past network devices, even as an option? But for players like Brocade, it’s revolutionize or perish. If the market for network equipment is commoditizing, as it surely is, the smallest players get eaten first and under the least favorable terms.
The Vyatta deal could be really big. It could be the move that ignites SDN and that unites SDN and NFV. Or it could be another missed opportunity. We’ll have to watch Brocade to see how they play this, and from that we’ll know their company future, and know more about the future of our industry.
