While the dust isn't even close to settling, I've had some time to ponder VMware's acquisition of Nicira and discussed the deal with ESG networking guru Bob Laliberte.
First, here are the positives:
- VMware is showing a lot of chutzpah by putting its money where its mouth is. The vision is to virtualize the network just as VMware virtualized the server. VMware has been pushing this message on its own for years and now it owns the company that has the most street cred in this area.
- VMware can use Nicira to double down in the service provider space, Nicira's sweet spot. A tightly integrated VMware/Nicira bundle may be a great foundation for big commercial cloud deployments.
- After the Cisco OPEN and COPE announcements, it seemed the air was slowly leaking from the Software-defined Networking balloon. VMware's acquisition of Nicira moves the industry past this speed bump. All of a sudden, SDN has deep pockets and an industry behemoth as a leader.
- On a more practical side, Nicira supports competitive hypervisors. This can help VMware maintain control while supporting other vendors from a networking and likely a management perspective. Very important, as ESG Research indicates that 70% of enterprises use multiple virtualization technologies -- Xen for desktop virtualization, Hyper-V for Windows workloads, Oracle for databases, etc.
Here are the challenges:
- From an enterprise perspective, virtual networking has had marginal success. ESG Research has pointed to a clear pattern over the last few years: Large organizations consider the virtual switch a basic access gateway to the "real" network. Other than VLANs, many organizations eschew the L2 functionality of vSwitches entirely. The connection to the physical network is so strong that many companies replace the VMware vSwitch with virtual switches from networking vendors like the Cisco Nexus 1000v. Why? So they have common configuration and policy management across virtual and physical networks. Nicira won't help here.
- I get the vision of replicating server virtualization in the networking space but we are talking about apples and oranges here. Unlike most x86 servers, the network is already a shared space with lots of crazy changing traffic patterns. You can certainly replace access switches with virtual switches but what about core switching and routing? I don't see many conservative enterprises lining up for this model anytime soon.
- As a startup, Nicira was full of bravado about replacing physical networks and freeing the masses from the tyranny of the old networking world. VMware has a different agenda and really depends upon partners for support. Can VMware pitch the Nicira world domination message without alienating its networking partners? This is especially true with regard to Cisco and its role in VCE.
- Some have written that VMware lacked networking knowledge and experience. Nonsense I say! I've always found the VMware networking team to be extremely knowledgeable in terms of networking technology. While VMware has very good technology chops, I find that it is sometimes lacking in terms of IT empathy and appreciation. In other words, VMware tends to think like a CTO not a CIO. Nicira doesn't help here.
- I hate to sound like Chicken Little, but one of the biggest obstacle to SDN, server virtualization, and cloud computing is a tremendous shortage of IT professionals with skills in these areas. Whatever VMware does with Nicira, it should double down on virtual networking training, certifications, and professional services.
- Nicira has had some success with big service provider deals but my sources tell me that each deal included a ton of professional services and custom coding. This is antithetical to the VMware transaction/platform model. There's a lot of work to do here.
One final comment. I'm no financial wizard but paying $1 billion+ for Nicira would be like going to your local convenience store and paying $100 for a coke. Very smart guys and a good vision but I can't imagine there was more than $50 million in real revenue though I'm sure the pipeline was pretty phat. The whole SDN market can't be much more that $100 million. Brocade does over $2 billion in revenue, has over a half billion in cash and has a market cap of just over $2 billion. Is Nicira's vision, revenue, and SDN image really worth half as much as Brocade?
VMware just made a very big bet. It could be a bold investment that more than pays for itself as it changes the face of networking. Then again it could be a very expensive distraction that serves to isolate VMware from the networking industry and IT professionals. We'll see.