VMware’s Cloud Strategy: Neither Deadwood Nor Sandalwood But More Than A Collection Of Driftwoods

Last week VMworld happened and some of the Clouderati were busy hanging out in the halls of Moscone Center at San Francisco. I was planning to attend the event but had to cancel due to personal reasons. But a steady stream of tweets and blog posts kept me updated about everything from the keynotes to the happenings at various booths on the Expo floor. Based on these third party reports, my takeaway from the event is “Gimme More”. Even though Project Redwood was not deadwood, as some expected, but it was not what I expected either. Of course, it has some exciting pieces including the ability to secure the virtualized environment but it didn’t give me the “aha” moment I thought is coming this year. There were many interesting announcements from VMworld 2010 but I am going to focus on the release of vCloud Director and vFabric announcement.

What is vCloud Director and vFabric, BTW?

vCloud Director is what some considered as deadwood, their cloud management platform for infrastructure running on VMware vSphere 4.1. Essentially, vCloud Director helps businesses transform their existing IT infrastructure into a secure private cloud by pooling their existing virtual resources. It also allows businesses to extend their private cloud infrastructure to public clouds supporting vCloud APIs (like the handful of cloud providers they showcased during the beta testing). vCloud Director also integrates with VMware’s vShield offering, thereby, bringing the much needed security solutions to fill some of the security and compliance gaps left in the virtualized infrastructure. In short, it is a pretty neat cloud management system which works seamlessly within the VMware environment (which are already there in many of the enterprises, anyway)

vFabric is VMware’s attempt to tap into the Springsource and other acquisitions to offer an application development platform for the cloud. VMware clearly realizes that enterprises are not going to let go of their existing investments on applications in their quest to move to the cloud. They also realize that the modern applications will be built for the cloud and not for the internal datacenters of yesterday. They are trying to position vFabric in such a way that enterprises could do both. In short, it makes it easy to set up a private PaaS for the enterprises. Knowing very well PaaS is the future of cloud services, VMware is taking a two pronged approach. On one side, they are partnering with the likes of Google and Salesforce.com to offer public PaaS solutions to enterprises hooked on to the Spring Framework. On the other side, they are developing a powerful platform which meshes well with VMware’s private/hybrid cloud strategy for enterprises.

Dude, what more do you want then?

Ever since Project Redwood details were leaked to the web, I was having great expectations about VMware delivering a cloud solution which even my grandma can deploy without my help. Well, I am a bit exaggerating here but you get the message. In fact, Carl Brooks of Techtarget also shares the same disappointment and articulates the problem with VMware’s vCloud Director announcement well.

But it turns out vCloud Director won’t do all that. It’s actually going to take a collection of disparate parts and third-party software vendors all working together to provide an IT shop with the full panoply of cloud features, and the requirements are both strict and strange.

First, to get the entire package, vCloud Director requires that users run a full-fledged Oracle database as the back end. Users are required to have VMware vSphere 4.0 Enterprise Plus with Update 2 or better. Additionally, vCloud Director is not virtualized; it’s a native application that only runs on 64-bit Red Hat Enterprise Linux 5; development started before VMware inked its deal with Novell to make SUSE the standard Linux distribution for its product family.

Definitely, the integration of vShield, offering application, network and end point security for the cloud, is a big step. As Chris Hoff points out, it fills some of the security and compliance gaps that existed previously. This will definitely make (v)cloud palatable to many more enterprises. However, as Hoff also points out in one of the two VMworld posts he did last week, the solution is not complete. For example, it doesn’t even have IPV6 support even though it is not difficult to add it at a later stage. For me, this appears more of a rushed package of cloud related tools than the one I was expecting.

However, I do want to point out that they are offering promise with their vFabric announcement. In fact, Windows Azure was so attractive to me because of the potential it offered to move our computing needs away from virtual machine chunks to tapping compute power from the seemingly infinite compute cloud. In my initial reaction to Microsoft’s Azure appliance announcement, I was talking about how Microsoft is missing out an opportunity which VMware is well positioned to grab.

The reason I was excited by the Windows Azure version of private cloud is because of my strong belief that PaaS is the future of Cloud Services. With the Windows Server-Hyper-V-System Center Approach, they could only offer the private cloud equivalent of IaaS and not PaaS. Windows Azure is very well suited for the PaaS future. By not playing this game properly, they are in danger of losing out to other players. Especially, VMware is playing their cards right on this one. Their acquisition of Springsource and beyond, their partnership with Salesforce and Google and their leaked plans for the cloudy future are all indicative of VMware making the right moves to capture the future marketshare. Microsoft faces a really tough fight and they have to actually get it right to stay really competitive.

Well, their vFabric announcement shows that they are getting ready to do just that. In fact, if they had released vFabric during the VMworld, it would have shook the cloud world and pushed many vendors down the drains. Their delay is definitely going to cost them some marketshare.

Conclusion

Even though their vCloud Director is not an ultimate game changer, it will definitely help them tighten their hold on the existing customers. Their vFabric may end up being a game changer provided they execute it well and, more importantly, do it fast. The battle lines are drawn in the enterprise cloud market with VMware on one side and everyone else on the other side. But one thing that could severely hurt VMware in this race is the absence of a credible public cloud strategy. What do you think?

Related articles by Zemanta

• VMWare launches six new vCloud products (zdnet.com)
• VMworld: VMware Unveils Deals, Partners (datacenterknowledge.com)
• VMworld: VMware pushes IT-as-a-service (v3.co.uk)
• VMware floats into developer cloud services (go.theregister.com)
• VMware’s Maritz: Innovation efforts will focus on IT as a service (zdnet.com)