This morning Oracle announced the acquisition of Xsigo Systems and there has been a frenzy of media coverage around it, primarily because it follows VMware’s acquisition of Nicira, creating speculation of future acquisitions and many implications to the future of IT. However, some of the positioning of this acquisition was just flat out wrong.
I first want to dispel some of the myths about what Xsigo does and the type of vendor it is. I’ve seen it compared to other SDN vendors, such as Nicira and Big Switch, but Xsigo is not an SDN vendor. From a very high level I would agree that Xsigo creates virtual network connections through software, so I guess it may be technically correct, but it isn’t an OpenFlow-based software controller that can automate the configuration of network changes, which is what “SDN”s have largely become. Xsigo is virtual I/O. The company sits between servers, storage and network infrastructure and virtualizes I/O connections, and does a great job of it.
RELATED: Oracle, Xsigo set network virtualization consolidation in motion
The Day VMware ate Cisco
For example, if you had four servers and wanted to connect them to four storage devices you would need four HBAs in each server and 16 total cables. With Xsigo, you have one Ethernet connection into the Xsigo infrastructure so each storage device connects once and the systems administrator creates virtual I/O connections. It’s a very cool, well-thought-out product that can save administrators a boatload of time and cut down significantly on human errors. Interesting product, but they don’t do what the SDN controller vendors do.
The second myth was that this was somewhat competitive with Cisco or VMware. Make no mistake: Oracle will do what Oracle does and will use this as part of its pre-configured, pre-integrated solutions Exadata and Exalogic. These are 100% Oracle solutions from the network stack all the way through the applications. It’s the enterprise IT equivalent of Apple. The thesis being, why cobble together a bunch of IT components to build a cloud when you can just buy everything pre-tested and know it works? Sure it’s expensive, and by expensive, it’s crazy expensive, but it works.
This is similar to what Apple does with its PCs, as opposed to cobbling together a bunch of parts for a Windows machine. Apple is expensive, but it gives a superior experience to Microsoft. Another analogy is to think of Oracle as the VCE alliance on steroids. While there is some loose competitiveness here with VMware and Cisco for part of the infrastructure, I really can’t see Oracle using this to attack VMware or Cisco. I think this is a bigger threat to the likes of HP, Dell or even some of the other application vendors.
Overall, I think this is a good move for Oracle. The company got some great technology that should slip into its integrated solutions nicely, which will strengthen Oracle’s data center positioning. However, this isn’t the acquisition that creates an all-out SDN war.