At VMworld 2010 Paul Martiz presented VMware’s strategy as a new stack of software which addresses the Data Center, the Cloud, Applications Platforms, End User Access to applications and how all of this is going to managed and automated. This is a full articulation of how broadly and deeply VMware intends to change the systems software industry, and why one can now credibly argue that VMware has become (instead of Microsoft and Red Hat) the most important systems software vendor in the world.
The interesting part of the VMware stack, is how each member of the stack is in fact a different (but highly related and integrated) business at VMware. Let’s take a look at the stack, and each of the businesses that correspond to each layer in the stack.
The VMware Stack
Data Center Infrastructure (Private Cloud and Hybrid Cloud) – vSphere
This is the maturation of VMware’s traditional server consolidation business. It really represents three distinct elements:
- A new device driver to the hardware in the data center. This is probably the most understood but least appreciated part of the VMware stack. Many people downplay the importance of the hypervisor presuming that it is just a thin layer of software that is destined to be free. Whether it ends up being free or not is besides the point. The point is that once a vendor owns the layer of software that is the interface to all of the hardware in the data center, then that vendor is in a powerful position to provide additional value added layers of software on top of the base layer (as VMware has done with features like VMotion, DRS, FT, etc.) and that this layer of software is very sticky. In other words, once it is installed and working and has multiple layers of software running on top of it, it is not easily replaced.
- The value added services that VMware bundles with the Hypervisor to create the virtualization platform. The best way to view the Virtualization Platform is as all of the features in VMware vSphere which are too considerable in number to list here. But the point is that VMware is not selling just a hypervisor, it is selling a virtualization platform that contains a great deal of value beyond the hypervisor, and that is itself a platform for a variety of add-on products.
- Speaking of add-on products this starts with vCenter Server, the management console for vSphere and extends into the rest of the VMware product line.
VMware is currently the clear market share leader and the clear leader on the fronts of product feature set, product maturity, stability, and performance. VMware has done a masterful job of evolving this platform so that it has grown in capability and maturity as the use cases for virtualization have progressed from test/dev through production tactical applications, and now on to business critical and performance critical Tier 1 applications. While it might be possible to use competing and less expensive virtualization platforms to address trailing edge use cases (the ones VMware addressed two and three years ago), the vast majority of global enterprises are relying upon vSphere to address the leading edge and demanding use cases that characterize the applications that have yet to be virtualized.
From the business standpoint, this is the engine that is pulling along the other four businesses. It would be fair to characterize this business and the other four as one engine that is currently pulling four boxcars. It is also important to note that this business represents one of two arrows aimed at the heart of the Microsoft Windows franchise. An operating system only does three things; it talks to hardware, it schedules hardware resources to workloads, and it provides services to applications. vSphere already does the first two of these three things.
The Public Cloud Platform – vCloud
vCloud is simply the version of vSphere that VMware sells to service providers (companies formerly known as being in the Hosting and Co-Location businesses) who want to now become public cloud vendors. The allure of the vCloud platform is a simple yet compelling one. The vast majority of the enterprise workloads that are candidates to run in a public cloud run on vSphere today. If you are a Cloud provider, don’t you want to make it easy for the customer to migrate their workload to your cloud? Of course you do, which is why you want the cloud platform that is the most compatible platform with what those enterprises are using.
From a business standpoint VMware’s biggest challenge here is that hosting companies and co-location companies are notoriously cost conscious as they realize that they have to have a cost advantage over the enterprise data centers from whom they would like to steal workloads. Therefore in this business VMware faces two cost pressures. The first is that VMware cannot charge “enterprise software” kinds of prices for vCloud as it does with vSphere as the Cloud vendors simply will not agree to pay those kinds of prices. The second is that there are numerous open source alternatives to vCloud that are on the face of much less expensive that an enterprise license for vSphere.
From a business standpoint, it is likely that the cloud providers are the long term strategic channel for VMware just as the PC and server vendors were the long term strategic channel to Microsoft. Over the long term as workloads migrate into VMware based public clouds, VMware will get paid as those workloads show up. Therefore while it is unlikely that this box car is going to morph into an engine in the next year or two, it may indeed long term be one of the most important revenue engines for VMware as it gives VMware the opportunity to participate in every computing transaction done between every vCloud provider and their customers.
The Application Platform Business – vFabric
At VMworld 2010 VMware announced vFabric, the combination of the Spring TC Server with the Rabbit MQ messaging software and the GemFire data handling technologies that have come through recent acquisitions. This is a hugely significant initiative on the part of VMware for a variety of reasons. The first is that this represents the third arrow targeted directly at the heart of the Microsoft Windows franchise. Microsoft did not make Windows successful by itself – Windows was made successful by legions of loyal developers who built products and custom applications to the Windows platform. vFabric gives VMware an application platform of its own. It also represents a direct attack on the Weblogic revenue stream at Oracle, and the WebSphere revenue stream at IBM. These two revenue steams alone are larger than all of VMware is currently so it is easy to see both a huge revenue and strategic upside for VMware here. This boxcar is likely to become an engine sometime in the next two or three years. vFabric is also strategically significant as it is the application platform on top of the VMForce cloud that is a joint venture of VMware and Salesforce.com, and Google has also agreed that it will be the basis of the Java support in the Google AppEngine cloud. VMware is therefore promising developers something that Microsoft has for years, and something that Microsoft has never promised or delivered. The common promise is benefits from tight integration of the VMware infrastructure platforms (vSphere and vCloud) with the VMware application platform (vFabric). The new promise is that of applications portability to places where the VMware infrastructure might not be present (VMForce and AppEngine). The deal is that you get the portability, and if you run in the VMware infrastructure you get extra benefits as well.
The Management Software Stack – vCenter
VMware made a very significant announcement at VMworld in this area – specifically VMware vCloud Director. This is the multi-tenant “son of Lifecycle Manager” which has been withdrawn from the market. vCloud Director is the layer of software that either an enterprise or a cloud provider can use to build a multi-tenant self-service environment on top of vSphere or vCloud. With this announcement VMware has defined a new category of management software for virtualization – specifically the layer of management needed to allow IT to behave like a service provider to its business constituents, and the layer of software that a service provide (a cloud vendor) needs to provide self-service provisioning for its public cloud customers.
This is an enormously important layer of software, and VMware is entering a market here that has to date been entirely served by leading edge startups like ManageIQ, DynamicOps, Embotics, Eucalyptus, Newscale, Platform Computing, rPath, Reflex Systems, and now with the acquisition of Surgient, Quest Software. vCloud Director will now give some shape and momentum to this marketplace, and inevitably invite comparisons between the VMware offerings and those from third parties. It is too early to say how this will play out but one of the key criteria that everyone is already considering is the trade-off between the degree of VMware integration that vCloud will be able to offer vs the degree of lock-in to vSphere and vCloud that it implies when compared with the third party vendors who will be positioning themselves as a virtualization platform and cloud platform agnostic management layer.
VMware also has several other management initiatives. In the performance and capacity management area, VMware has CapacityIQ and Hyperic. The positions of these products relative to alternatives in the third party ecosystem were covered in VMware’s Systems Management Strategy – VMworld Update. VMware also announced the acquisition of Integrien in this area which was covered in The Significance of the VMware Integrien Acquisition. VMware announced three new security offerings at the show, the impact of which upon the market and the third party ecosystem were covered in Virtualization Security Strategy from VMworld 2010. Given VMware activity in all of these areas what was surprising was the lack of activity in the backup area. This area continues to be dominated by Veeam, Quest (now that Vizioncore has been folded into Quest), and to a lesser extent Phd Virtual. The really interesting vendor in the backup space is Pancetera who has come out with a solution that turns the entire virtualized data store into an addressable file share – which in turn allows any backup product that can talk to a D: drive to be used to back up virtual machines.
Of all of VMware’s businesses, the management software layer is the most difficult to assess. Other than vCenter which is really a console for vShere, VMware does not have an outstanding success on the management software front to point to. AppSpeed appears to have lost almost all focus and was not even the subject of a single session at VMworld. CapacityIQ is a nice product but faces aggressive and competent competition from the like of Quest (vFloglight), Hyper9, Akorri and many other monitoring vendors. The products that VMware purchased from Ionix earlier this year (discussed in depth in this article), added some important building blocks to the mix, but have yet to be integrated into a coherent management stack.
It is also clear that for the most part VMware is focusing its management offerings on managing its own platform. The view here apparently seems to be that sooner or later everything will be running on vSphere, so all you will need is a vSphere management platform. While that may end up being true for a small number of customers. It begs the question of who is going to monitor and manage the physical infrastructure that underlies vSphere and vCloud as well has manage the cross virtualization platform environments. This seems like an opportunity ripe for a next generation monitoring platform vendor like Zenoss or SevOne.
The End User Computing Business – VMware View
At VMworld, VMware announced the delivery of VMware View 4.5 and a slightly tweaked End User Computing strategy which we discussed in VMware’s Desktop Strategy – Doubling Down on VDI. The most significant new feature of View 4.5 is that the ability to check out your VM from the data center, copy it over to your laptop and run it on a Type 2 hypervisor on top of your existing Windows OS is now officially supported. VMware also announced an any app any where strategy that consists of putting the View client on devices like the iPhone, iPad and Google Android devices. But all of this begs the most important questions which is who and or what is going to really reinvent how we manage desktop computing. If your enterprise has 50% users with laptops who travel some of the time and 50% users with desktops who only need access from their desks what is the right strategy? Are you really going to build out a VMware View back end data center for all of your users when only 50% of them are going to really want to use their VM’s running in the back end environment? It seems that a much better strategy would be one that recognizes that users are going to carry devices around with them, and that figures out how to provide IT with the management that IT needs, and users with the flexbilty that they need. This is the path that Virtual Computer, Unidesk, MokaFive, Ringcube, Wanova and other startups are blazing – a path that recognizes that the portable fat client Windows devices is probably going to be with us for quite a long time. On the front of the future of VDI one has to look at best of show winner Kaviza who has created a way to do VDI that avoids much of the cost and complexity of the VMware View SAN dependent approach.
As a business this would have rank as the least important of the five to VMware. The other four businesses all represent significant increament revenue opportunities. Cloud platforms, management software, and applications platforms all probably each represent a revenue opportunity for VMware that is at least as large as VMware will be in total this year. On the desktop side as long as VMware remains tied to an approach that is dependent upon the centralization theme, VMware will be fighting trench warfare with Citrix who has owned and dominated the business of centrally managing and delivering application to users for years and who is simply going to turn VDI (XenDesktop) into a feature of its application delivery suite.
VMware dominates the enterprise virtualization platform business with vSphere, and is poised to create a vSphere compatible public cloud ecosystem around vCloud. Layering Management software on top of these platforms is a logical progression up the value stack, as is layering an applications platform (vFabric) on top of vSphere and vCloud. VMware’s end user computing strategy seems to be too tied to VDI to be able to break out of the fundamental limitations associated with this approach, and will likely leave the larger question of how to manage the next generation desktop to the previously mentioned startups and perhaps Symantec.