For server virtualisation, Microsoft bets against VMotion

In competing against VMware, Microsoft seems to be betting that at the low to middle end of the market, it makes more sense to utilise planned downtime when data migration is necessary instead of paying the additional cost of VMotion or Storage VMotion.

Two industry giants -- VMware and Microsoft -- are vying for supremacy in server virtualisation. Currently VMware is the Goliath, with sales of server virtualisation software far outstripping those of its competitors. But industry powerhouse Microsoft is wading in with Hyper-V.

The emergence of Hyper-V (after its rather disappointing predecessor) as a far cheaper alternative to the VMware suite has caused many companies to reassess their virtualisation strategy.

While maturity of product is often touted as VMware's biggest strength, cost is -- wrongly or rightly -- often at the forefront of the decision-making process. Storage managers wishing to take advantage of a virtualised environment should not only ask the cost of deployment and what ROI they could expect to achieve, but also how flexible is the virtualised architecture, and do we understand the value/priority to the business of the applications we are likely to virtualise?

The functionality afforded by VMware's ability to virtualise storage and leverage disk as a pooled resource, rather than a fixed device, has long been regarded as an ideal goal for storage administrators. This has also been something that many vendors have strived to achieve for years. But with each vendor touting a different product and method for virtualisation, no clear standard has been defined, allowing those in the virtualisation space to pick up this functionality at the operating system level.

Through the use of VMFS, VMware allows a storage administrator to utilise presented disk, and provision from that to its virtual machines instantaneously, with full flexible management from the Virtual Centre server. Combined with this, Storage VMotion allows the administrator to relocate virtual machine disk files between and across shared storage locations whilst maintaining continuous service, thereby allowing you to maintain your environment (server or storage) with less service-affecting downtime. It would also allow far easier migrations from old to new technology, an infrequent but high-cost exercise.

Despite the marketing, most companies going through the process of creating a virtual environment realise that those servers/applications best suited to being virtualised sit in the low to middle tier. Web servers, infrastructure servers, low-profile database servers and administration servers are all typical choices, and they are all in the tier that Microsoft is most comfortable, and already have massive pre-virtualisation presence.

In this, Microsoft may have the key. Microsoft's bet is that the additional functionality that brings VMware to the fore may not be best placed for this type of application, and that in this area it makes more sense to utilise planned downtime when migration is necessary instead of paying the additional cost of VMotion or Storage VMotion.

And so the lines are drawn. Hyper-V is getting good reviews (even from those pundits that are typically anti-Microsoft), but to obtain optimum performance, you need to be running Windows Server 2008. The deciding factor may be the placement of Virtualisation in a market that is moving fast to adopt it. The first tentative years of adoption are over, and companies are beginning to understand where virtualisation sits, where it works, and where it needs work to progress. The key to VMware's success lies in its ability to progress virtualisation in to the space where Microsoft finds it difficult to compete: the mid to high-end arena. With the backing of its parent company EMC (a company whose products are more than comfortable in that space), and investors such as Cisco and Intel, it may just be able to do that.

So the battle, once again, comes down to the age-old argument of cost over functionality. The savvy adopter would do well to understand themselves, before investing in others.

About the author: Allaster Finke is a senior consultant with more than eight years experience in the design and deployment of complex storage networks, backup, high availability and disaster recovery solutions. Prior to working with Glasshouse, he was a senior enterprise storage consultant at Unisys.

Read more on Virtualisation and storage