SANs Stall Server Virtualization

But is the SAN itself really the problem?

Awareness around the importance of deploying networked storage with server virtualization has became fairly prevalent. Most IT shops we speak with get the value of virtual machine mobility, high availability, disaster recovery, etc. and tie that to the fact that all these goodies require networked storage. Also the SAN storage vendors absolutely love the fact that server virtualization has helped keep the boat afloat over the last few years and continues to drive sales and services. Nice!

Next comes the fact that the majority of workloads that have migrated onto a virtual machine are IT-owned, low hanging fruit, minimal performance requirement applications. It has been easy for IT staffs to convince themselves of the value of server virtualization. Server virtualization has allowed them to perform maintenance during operating hours, saved them from having to make late night and weekend visits to the data center, improved service levels, and just plain increased morale inside IT. Thus far, the success of server virtualization has been truly phenomenal and has created heroes inside IT. Nice!

Now, IT shops that have made investments in virtualization and migrated the low hanging fruit are finding themselves screeching to a halt as they size up the next harvest of potential workloads that are ripe for the virtualized environment. There are many reasons this is happening, but one that keeps boiling to the top is the SAN. This is not to say that the SAN itself is the problem, rather the problem could be misconfiguration, lack of virtualization skills, inexperience with the technology, and poor planning. Not so nice!

As a result, guess who gets blamed? That’s right, the server virtualization providers: VMware, Citrix, Microsoft, Parallels, Oracle. The majority of the time, it has nothing to do with the server virtualization solution, it is almost always a misconfigured environment. The bad news for these vendors is that they don’t want to ever be perceived as the bad guy in the stack, so they are finding themselves having to become SAN experts to support their customers. Crazy!

Here is an all-too-common true story:

IT sells the business on the value of server virtualization and calculates the ROI on the back of a napkin during a lunch meeting. They get the green light. Server virtualization is still a relatively new technology, but how hard can it be? And look at all these great things it does! IT begins by either sizing up the environment themselves, working with an integration partner, or subbing out the project so they can confidently get started off on the right foot. Some choose to do some type of virtualization assessment analysis (most do not) and they purchase some gear. The gear arrives and everyone giant smiles on their faces as they rack and stack it into place. IT starts virtualizing print servers, file servers, and small custom applications and are giddy with success to my point above. Confidence is high and they start to target the next tier of applications, such as Microsoft Exchange, and suddenly realize that the 7200RPM SATA drives they purchased to support their entire virtualization deployment may not cut the mustard. They fear that installing a VM running Exchange 2007 might be sluggish based on the slower 7200RPM drives. They reached back out to the original consultant that helped them launch the project and the consultants say that most of the read/writes will be done in cache mode and will be fine. Of course they do. What are they going to say? No, we screwed up and it won’t work!?!?

The fear is that they drop this new Microsoft Exchange VM in place and start having major performance issues. Their instincts are spot on and now they are faced with the challenge of how to migrate Exchange to the virtualized environment on their existing virtualization investment that has the compute horsepower and storage capacity available, but not the storage performance. Yikes! “Hey Mr. Bossman, you know that virtualization project you signed off on 6 months ago and all that new gear we just spun up? Well, looks like we may need some more because what we originally purchased really isn’t going to work for the Exchange migration you just tasked me with.”

As a result of all this and other similar scenarios, server virtualization deployments are stalling. While server virtualization has made giant strides over the past decade, persistent challenges related to application requirements, security, networking, infrastructure considerations, and organizational conflict are creating confusion--and inertia--for IT shops. Both the percentage of virtualized workloads and virtual-machines-per-physical-server ratio remain disappointingly low.

There is a huge lack of virtualization skills and expertise that still needs to be found within IT and amongst the vendor ecosystem. Server virtualization is at risk of continued success if it can only reach the low hanging fruit.

Join the discussion
Be the first to comment on this article. Our Commenting Policies