Doing a Google search on “virtualization tipping point” will pull up many results around virtualization adoption and maturity across the board.  While the topic is not at all new, after visiting several clients last week that have adopted virtualization into their environments, it became clear that the message needs reinforcement.  I always see the numbers, 17% of servers virtualized, 24% by 2010, etc., etc.  These are typical numbers with the progression of a technology such as virtualization and while they work great for marketing materials, we have to look under the covers to really see what is happening.  We want our clients to see virtualization as the transformative technology it is and not treat it just like another operating system. It needs to break the mold of how we have treated the physical 1:1 workload environments.   

The conversation with the clients covered very similar topics, whether they were leveraging virtualization heavily or just for a few workloads.  Most deployments are still stuck away in departmental silos, used only in test/dev, didn’t fall under a global management scheme, etc.  The list goes on and on.  Virtualization health checks have been one of the most sought after services we provide.  Many times these are clients wanting to validate where they are and to ensure they are getting the return they signed up for.  This is turn in the right direction, and gives virtualization ambassadors (us) a chance to further drive the technology.

Organizational Readiness – the “people and process”

Let’s take the process side first.  Virtualization’s highest value should be realized when the technology is ingrained in the processes that IT and business use day to day.  Change, Project, Risk and Operations Management need to be modified or even recreated to fit into a virtualized landscape.  Too many times the provisioning model is hampered by leveraging the existing physical server process.  Too many times the traditional backup process is followed when there are more efficient and effective ways to get our data safe with virtualization.  

From the people side of things, training is always a must.  This might be in a classroom or delivered at the client site with workshops and proof of concept deployments.  But it should not stop at just the technical team, the Windows or UNIX team which normally is in charge of the  virtualization deployment in the beginning.  Application teams, change management, executives and line of business owners all need awareness of the technology and how it affects them during the lifecycle of the project.  This is always a challenge but needs to be tackled to help accelerate the overall acceptance.  Each group will most likely need their own custom content.

Environment Rightsizing

Most organizations we run across did some initial capacity planning, used some conservative scenarios and then cut the consolidation ratios in half to play it safe.  They might have obtained some savings and met their initial goals, they might only be realizing a small percentage of the benefit.  Capacity planning is important for the foundational design and planning, but capacity management is a function that needs ongoing attention.  Setting up guidelines for the physical to virtual conversion of CPU, memory and IO resources is a good start.  This will be more successful if the application teams understand why they don’t have four CPUs now in their server, hence “people and process”.  

Leveraging tools to monitor and track performance and trends is a next step.  If we have the ability to leverage existing resources to virtualize more without adding more hardware and/or software it is a big win in this economic climate.  To know if this is possible, we need to be tracking this with some automated tools. These tools also can help us build rules inside of features such as VMware’s Dynamic Resource Scheduler (DRS).  While we like to see clients use the fully automated function with DRS, many organizations have too many policies in place to allow that much freedom.   The tools can help build a case for some partially automated action and might eventually move to fully automated.

Further consolidation of services also needs to be discussed.  This means looking at ways to not just strink the server footprint, but also the services on each virtual workload such as backup and security tools.  Removing  those backup agents from the workloads and backing up at a consolidated level helps to increase performance on the virtual workloads and hosts.  Storage plays a big part in this, so there might need to be changes in the existing structure.  Firewalls and Malware protection is next in line, but the technologies are still maturing and are more of a play with desktop virtualization as of today.

Road-mapping – what is next?

Since many virtualization deployments start at a departmental level or in a silo, it doesn’t always make the master roadmap of the organization.  There isn’t anything wrong with starting like this, we want to see virtualization adopted and this can be a natural step in adoption.  To make sure adoption continues a solid roadmap is necessary.  Roadmaps for business adoption, technology adoption, people and process and how health check what we have.  Once the technology has proven itself, it is time to weave it into the overall plan and virtualization is a very big item to weave in.  Making sure to have the end to end solution for servers in place, build the foundation and practices then start proofing out automation tools, desktop/application virtualization and building a business cases if the tools suit the business.

As consultants focused on virtualization, we get a chance to see many different cases.  Whiles these are fairly broad topics, we feel they are starting points to allow organizations to get over the tipping point with the technology and allow it to transform them as it was meant to do.