Despite a dire need for belt tightening among government agencies, the public sector still lags well behind the private sector when it comes to virtualization. In fact, more than 80 percent of public agencies surveyed said they aren’t using virtualization to its fullest extent, according to the CDW Server Virtualization Lifecycle report. The major stumbling block: IT resources.
Staffing and budget challenges are impeding progress on that front. Lack of staff knowledge to manage the technology was noted as the #1 barrier to adopting virtualization in federal agencies. So ironically the lack of funds is preventing agencies from realizing the significant savings possible from virtualization.
What will get these agencies over the hump? First, let’s acknowledge that virtualization alone is no panacea. If not managed, it can result in server sprawl, just as physical servers can. To truly make the most of virtualization, organizations need a way to create, manage, move and control discrete workloads (combinations of applications and operating systems), without taxing IT staff. After all, what is the purpose of embracing new virtualization technology if it’s going to result in additional burdens to IT staff? Or, if the agency lacks the skill set to manage these technologies?
Ideally, agencies need a way to automate the decision making processes involved with managing workloads. In this scenario, workloads may need to be elastic and scale on demand without manual intervention. Further, each workload would actually be managed intelligently via universally unique identifiers and run-time metadata to know how and where that is executing—whether it could be moved to a physical or virtual server, or out to a cloud. Agencies have to further factor in the location aspect from a regulatory compliance standpoint so that sensitive data does not leave the premise, and burst into a computing resource that is located outside the country. This data location aspect makes decision-making harder for agencies.
Agencies also need to consider infusing workloads with intelligence about the computing environment and the enterprise policies, so that we could automate the process of controlling these workloads. What if, for example, the workload contained a framework, akin to metadata, that allowed it to orchestrate and execute itself in a consistent way across physical, virtual and cloud infrastructures? As the workload moved through the lifecycle, it could gather additional intelligence that would be crucial to future workload management, as well as reporting necessary to prepare for regulatory audits. There would be no need for ramped up IT resources. These tasks would be handled in an automated fashion, freeing technical staff to focus on higher value services.
It is this run-time intelligence that will enable public agencies to automate their workload across physical, virtual and cloud infrastructure, and fully embrace virtualization.
Are you seeing similar roadblocks to virtualization at your agency? What do you think will help you overcome these hurdles?
Posted
Aug 03 2010, 09:45 AM
by
DiptoChakravarty