In a recent blog, Robin Harris postulates that virtualization is a feature, and not a product. As a result, Harris postulates that VMware will lose.
A more useful analysis is to agree the core underlying hypervisor can be regarded as a feature. However, the rich set of management products developed by VMware to run on top of this virtualization layer are well differentiated and are in no danger of becoming features on any server manufacturers price list.
Feature vs. Product
Virtualization allows more efficient use of resources by allowing multiple operating systems types and instances to run on a single physical piece of hardware, improving utilization of resources and providing better availability, operational efficiency, and flexibility. Back in the 1960's, the first true virtualization IBM product was called CP67, which evolved into VM/370 on mainframes and eventually became the LPAR feature on mainframes and AIX machines.
However, CP67 also became CMS, in its day a very successful time-sharing end-user product, that provided a flexibility and ease-of-use ahead of its competition. CMS timesharing gave way to the PC, not to any other virtualization feature.
Just providing a hypervisor that allows multiple instances of an OS to run is relatively easy. VMware gives this function away for free. The issue for VMware is how to make a set of products on top of the virtualization feature that operating systems cannot subsume. VMware is differentiating by building software management products on top of the virtualization that solve the security, performance management, problem determination, and SLA management challenges better than the underlying operating system can manage. VMware is continuing to build a suite of software products which cover:
- Data Center & Cloud Infrastructure (vSphere, vCloud Director and others),
- Infrastructure & Operations management (vCenter and others),
- Security (vShield and others),
- Application Management (vCenter AppSpeed & Application Discovery Manager),
- Desk-top & end-user Computing (View and supporting applications),
- VMware Application Platform (vFabric and others).
The application components are weak (and these pieces are being moved out of the VMware portfolio and into EMC's Pivotal Initiative), but the operational management products are very strong and well integrated in a majority of the data centers.
Key Questions to Ask about the Future of VMware
- Can Red Hat coordinate the writing of open-source software that can provide this range of function for Xen and Linux?
- Yes, but they will be following some way behind for many years.
- Can Microsoft do the same for Hyper-V, Windows 2012, and Systems Manager?
- Yes, it will eventually produce a great product, and it will compete with VMware for Windows environments but very unlikely for Linux.
- Can Oracle VM be useful?
- That depends on Oracle investing in being inventive - some hope exists.
- Will competitive pressure from Red Hat and Microsoft put pressure on VMware prices and margins?
- Of course - users will hope for genuine competition in the marketplace and will choose the best fit of function and price for their environments.
- Will every server vendor provide the equivalent of the VMware Suite as a feature?
- Not in any conceivable timescale. These vendors may provide features that assist virtualization, but there are no known plans for them developing comprehensive management suites.
The overall objective for VMware is to provide the basis for a software managed infrastructure, a set of services that can be applied as required to applications running on commodity hardware. The biggest threat to this are the hyper-scale organizations such as Amazon, Apple, Google, Facebook, OpenStack, Yahoo, and three-letter acronym federal agencies, who would prefer a software-led infrastructure built on open-source products and open APIs.
Action Item: Investors and CXOs should not conclude that VMware will lose. VMware is currently the leading vendor of virtualization management products. VMware faces increasing competition from other software vendors, especially Microsoft & Red Hat, and many organizations have and will extend multi-hypervisor strategies. Long term, VMware faces pressure from the Hyper-scale enterprises. Senior executives with significant VMware deployment should put pressure on VMware to lower prices and drive for higher volume and greater penetration, as the best defense against a sea of troubles.