The Pitfalls of Virtualization Adoption

Anybody and everybody can write about virtualization technology. However you will see that quite a few of my posts are about the aspect of virtualization consulting. Not to get all Zen (not to be confused with Xen) but let’s face it virtualization isn’t easy. It has just as many cultural impacts as it does technological. Needless to say the paradigm shift is more of a state of mind than a technology rollout. So I have put some of the things I make sure to educate and overcome in any virtualization engagement. So here are some of the pitfalls of virtualization adoption and strategy. 

(In no particular order)

• Religious Battles – Avoid the Linux vs. Windows or Linux vs. Unix or even the Windows vs. Everyone battles. Choose your platforms based on needs and standards that can be created. Be aware of the virtualization ecosystem. The ecosystem is more important than the hypervisor itself.
• Procurement & Business Changes – How many times has a business owner came and said, “I need 6 servers for my new app by tomorrow.” After you are done laughing on the inside, putting on a sympathectic look and saying, “It takes 6 – 12 weeks to get a server racked and stacked in the datacenter.” Take a look on how that conversation should go. Train the business they don’t need servers. They need resources. They need memory, processing power, network bandwidth and storage. Then always (when applicable) virtualize first. Offer to say I can get you a physical server and storage for X thousands of dollars with a 3 year lease or I can get you virtual assets that we can bill month to month based on utilization and scale up or down depending on when you need to scale up or down. See what they say then….
• Myopic Virtualization Strategy – Server Consolidation is where virtualization begins – not ends. Remember the 5 facets of virtualization (Server, Desktop, Application, Presentation and Management)
• Physical Cost Recovery Models – Modify your internal cost recovery models for virtual assets. Use real numbers and resource based cost recovery. Use a base building block as a cost unit and go from there. For example a 1CPU x 1GB VM is a great base block.
• Physical Asset Based Security  – Virtualization opens some new security challenges. But don’t forget it also solves some too.
• Stagnant Support Programs – Use virtualization to support virtualization.  Use snapshots, have a support VM farm for troubleshooting. (Just don’t do snapshots with domain controllers)
• Virtualization Paranoia – I can’t remember how many times I’ve heard that some app won’t be supported on a virtual platform. There are only two reasons that statement can be true. First pure performance. If a VM needs 128GB of RAM and 32 Cores - Dont virtualize it! Second some hardware requirement. Today at least you cannot virtualize something that requires a specific PCI card for example.
• HA / DR Strategies – First HA and DR are different. (But that a whole other blog entry) Secondarily Offline VMs, VM Libraries and many other things can be used to create more supportable and cost effective high availability or disaster recovery scenarios. Don’t forget business continuity either!
• Over-Virtualization – Anyone that says VIRTUALIZE EVERYTHING is WRONG. Not everything should be virtualized. Virtualization is NOT a silver bullet. Virtualization coupled with a great ecosystem, proper operational processes, and organization constructs can be.