Last week, IBM made a significant announcement that received relatively little attention. The IBM System Networking Distributed Virtual Switch 5000V is the first third party virtual switching platform for VMware environments. Of course, Cisco has the Nexus 1000v, but given the long history, investment, and VCE joint venture, it doesn’t really count as an independent third party in my book.
This is an important development because as the first tier of network switching moves into the server, the virtual switch becomes extremely strategic real-estate and control point for emerging SDN architectures. Selling virtual switches alone will likely not become a standalone big business but when coupled with external networking controllers, possibly connected over OpenFlow, a host of new architectures are possible. It also highlights that IBM is serious about networking and points to further competition between and Cisco and IBM’s networking and server groups.
To build a product like this one, IBM needs low-level access to the hypervisor, which is not available in publicly documented APIs. Translated, this means IBM needed VMware’s explicit permission and help. Which means VMware is in the cat-bird seat to play a significant role in determining the winners and losers in virtualized networking.
I learned the hard way how closely VMware guards access to the virtual switching layer while incubating a networking start-up in 2009. That company never got off the ground and threw in the towel once it realized it couldn’t build its own Vswitch for VMware. I know of other start-ups and traditional network vendors who have gotten the same cold shoulder. Consequently, many of the SDN start-ups are either targeting Xen or KVM environments because they can’t replace the virtual switch in VMware environments or are forced to try to gain control at the top of rack.
But of all companies, why IBM? The technology giant sold off its networking assets to Cisco in 1999, but got back in the game big time with its acquisition of BNT in 2010. Undoubtedly, IBM must serve as a channel for a large amount of VMware software. Could this provide IBM the leverage on the networking side of the business?
BNT (now IBM System Networking) has a solid product line of blade server and top of rack switches, but is lacking in higher density aggregation and chassis switches where Cisco dominates. Interestingly, IBM still has Cisco products on its web site and also resells Brocade and Juniper. You could call IBM’s network strategy schizophrenic or customer driven, depending on which side of the debate you wanted to take.
IBM moves are just the latest step by traditional server vendors to build out networking portfolios since Cisco’s entry into the server business shattered the historical dividing line between the two categories. The acquisitions of 3com by HP, BNT by IBM, and Force10 by Dell speak to the new competitive reality. Credit to IBM for offering something that is unique in the market and to VMware for enabling them. This product should be embraced by enterprise customers who already have a large installed base of IBM Blade Center server’s and BNT switches.
I would also hope this marks a change in philosophy at VMware and they will work to support broader ecosystems of third-party virtual switches. But I doubt it. VMware will likely continue to closely guard this strategic control point until companies pushing KVM and Xen products start to impact them in the market. But I wouldn’t anticipate that happening soon as the dividing lines between these segments are well defined with VMware in the enterprise and Xen/KVM in the cloud/web scale operators. In fairness to VMware, it does take material resources on its side to enable this type of integration, so until it feels pain in the market, it has little incentive to act further.
Lastly, this announcement was certainly heard loud and clear on Tasman Drive. Cisco can’t be happy about this as it improves IBM’s networking story in the lucrative enterprise market Cisco has historically dominated. Fascinating times in networking once again.
Alex Benik is a principal at Battery Ventures who invests in enterprise and web infrastructure start-ups. You can find him on Twitter at @abenik