2010 Prediction: Jay Litkey, Embotics
2010: A year of complexity
As I predicted last year, despite the economy, 2009 saw an increase in the overall level of virtualization in the datacenter, and the emergence of “virtual first” policies. This will continue in 2010, and even accelerate as the economy starts to recover. As the size of the virtual datacenter increases, sprawl will become a bigger and bigger problem, and a larger percentage of VMs will be in production environments rather than in development or lab environments.
As penetration increases, so will competition for VMware. We are already seeing new entrants (i.e., Red Hat), and we should see a push into the market from folks like Oracle. Microsoft has been slow off the mark here, but expect to see them become more and more of a force in 2010. This increased competition (many of them willing to buy their way into the market), will almost certainly put downward pressure on virtualization price points, and one can expect to see some significant pricing action from VMware before the end of the year in an effort to lock in key accounts and market footprint. This competition will also mean more organizations adopting heterogeneous virtualization platforms, making environments more complex and compartmentalized. Cloud computing will continue to be tested by most organizations but without any real commitment and only with non-critical applications.
The increased datacenter penetration will also have an impact on datacenter management tools. Once a datacenter is more than approximately 25 percent virtualized, virtualization management becomes more and more important. There will be more internal emphasis on not only virtualization specific management, but also on “management convergence,” i.e., management across multiple virtualization platforms as well as across the physical and virtual data center management. This means ITIL will move over to the virtual world, and integration and cross platform management will become key.
Reacting to this, the main (physical) datacenter management system providers will make more claims about their ability to manage both environments from a single platform. Whether they can back up these claims is another issue.
Virtual sprawl is inevitable given the current state of management incorporated into the virtualization providers’ platforms. In many ways sprawl is good for them. More VMs = more hosts = more license revenue, so don’t expect to see any serious effort on their behalf to fix it. This means that organizations will need to add additional management layers over the base management provided by the virtualization platforms.
But with the capital markets remaining tight, there will likely be some consolidation in this virtualization management market, and more focus on optimization, scalability and simplicity. As virtualization penetration increased in the datacenter in 2009, headcounts generally did not. Consequently, virtualization administration teams simply do not have the time for complexity. At the same time, they are realizing that point tools will not scale to where they need them, and therefore we will see a shift towards a more systemic and scalable approach to management, with the focus on ease of deployment, i.e., no forklifts!
