Server virtualization has been a driving force in the IT datacenter for the last few years as customers have moved many of their workloads from physical to virtual servers. Hypervisors have provided extensive management to coordinate between the virtual and physical server resources but there is still a gap between the virtual servers and the storage resources.
Now managing virtualization and storage in context is becoming critical and is helping to break down the strong silos between virtualization and storage that have persisted from the physical server and storage days. The new landscape that has developed around the virtual environment is driving convergence and helping bring virtualization and storage closer together.
The speed and demands of a typical VMware (or other hypervisor) environment has increased substantially. In the days of physical servers, changes could take weeks or even months allowing similar time frames for the storage team to adjust.
With VMs being built and deployed in minutes these days, storage needs to be able to adjust much faster. As a result, there is closer communication and more real-time data sharing between virtualization and storage. In the simplest form, this may be the VMware admin and the storage admin just talking more frequently but increasingly it takes the form of tools that enable real-time visibility back and forth between the virtual and the storage systems. The virtualization team wants to understand when there are performance issues in the storage environment while the storage team wants to understand what VMs and data stores are pounding their storage the hardest. Increasingly there are vendor management tools that can provide visibility from the VM all the way down to the storage spindle.
Advances in hypervisor technology have continued to pull in some basic management of various domains into the hypervisor tools. This is not just storage where the hypervisor will automatically allocate preconfigured LUNs and are major users of thin provisioning of storage.
Similarly, virtual switches are like gateways to the networking team. These gateways provide increased linkages and domain knowledge (or at least understanding) to virtualization or server management team that can help improve the level of interactions with the domain specialists.
While applications running on physical hardware clearly exist, the bulk of the IT spend over the last few years has been on infrastructure supporting new or expanded virtual environments. Often this has meant that there is a single organization and team that owns the applications, virtualization software, physical servers, storage hardware and networking components needed to deploy their resources.
That doesn’t necessarily mean that the individual teams that were set up to manage the existing physical resources have been disbanded but often the two teams run in parallel. Under this type of organization, all the team members typically will have a common manager or executive that can help resolve conflicts and motivate good behavior across resource skills.
We shouldn't get too giddy about the progress that has occurred; there is still plenty of room for improvement. But virtualization should get some credit – it has been a driver for more efficient cross-domain operations that has done more than decades of management consultants before it.
Mike Thompson is Director of Business Strategy for Virtualization and Storage at SolarWinds.