There's no doubt that data storage has come a long way in recent years. With the rise of virtualization, each storage array is now serving more applications than ever. Then came the virtualization of the storage array itself, which, when combined with virtualization on the server side, is helping to make storage systems more efficient by driving utilization rates much higher.
Now the goal, says Craig Nunes, vice president of storage marketing for Hewlett-Packard, should be to take those advances to the next logical conclusion by federating all the storage across the data center so it can be managed as one logical entity. In that scenario, Nunes says Peer Motion software from HP functions as a distributed volume manager that automatically manages the placement of data across HP's 3Par and Lefthand Networks storage arrays.
This capability is becoming increasingly important because as application workloads running on top of virtual machines move about the data center, IT organizations need a way to optimize the allocation of data across multiple storage systems to maintain performance levels in data center environments that are becoming more unpredictable. Otherwise, says Nunes, each time a virtual machine moves, it creates the probability of introducing more latency.
Peer Motion is one of the latest examples of the convergence of data and storage management. As virtualization becomes the norm rather than the exception, the time to rethink how storage is managed is at hand. Rather than dedicated storage systems to specific servers and applications, the storage systems need to be a shared resource available to every application in the data center, especially application workloads that are likely to move without a moment's notice. This creates a requirement for a more federated approach to managing data that gives the underlying storage systems a lot more visibility into what's happening at the application layer.
On one level, because of virtualization, the job of managing storage has never been more complex. In fact, it's starting to move out of the realm of a function that can be managed manually. As such, IT organizations are going to have to start finding storage systems that can increasingly manage themselves.