In 2010, HP conceded that it saw unified storage as being inadequate to future demands and announced its move to a new strategy it calls “converged storage.”
Compared to unified and monolithic (dedicated stack) solutions, the company claims that converged storage delivers benefits including:
- up to 41% faster performance
- up to 90% lower storage management time
- up to 50% less power, cooling, and floor space
- reduction of IT service delivery time to minutes
According to Brad Parks, converged infrastructure marketing strategist for HP Storage, the new strategy divides into three key points: standardization on x86 hardware, adoption of scale-out software for up to 40% faster deployment, and converged management.
“Typically, when you’re having to do hand-offs between a network guy, a server guy, and a storage guy, an average storage application deployment can take days if not weeks,” says Parks. “Because of the automation we’re doing on the management side, we see new apps coming up in a matter of minutes, with full automation from apps into servers and then into storage.”
The advantages seen in energy, cooling, and footprint arrive from the transition to a very dense blade architecture in which servers, storage, and networking all exist within a single enclosure. For instance, a single HP BladeServer c7000 might hold eight server blades, eight storage blades, and two HP Virtual Connect modules. That’s at least 16 CPU cores and up to 114 storage drives, or roughly 100 TB of capacity. Instead of spreading these resources out over a host of 3U boxes or towers plus two or three JBODs plus networking components, everything now fits into one box with all components sharing common power and cooling.
Scale-out storage infers a design that can not only rapidly expand in capacity but also throughput, file size, and file volume, either independently or concurrently. This is much easier to achieve on a blade-type infrastructure, where buyers might start with two nodes and scale overnight to five or six. Meanwhile, software such as VMware plug-ins span both server and storage virtualization at the same time, again denoting the “converged” design. Such plug-ins can allow for systemic health monitoring, management, and much more.
“We also have some higher-level management tools,” says Parks. “For example, we have a product called Cloud System Matrix, which is actually a bit above the mid-market, more into private cloud deployments. But you can go to a single portal and automatically provision storage from soup to nuts by using a common set of orchestration tools because we built in the plumbing that ties all this together. Because of the automation we’re doing on the management side, we see new apps coming up in a matter of minutes, with full automation from apps into servers and then into storage. ”
Naturally, converged storage solutions come in all shapes and sizes, but the underlying principle remains the same. As Russ Fellows notes, “You don’t necessarily need to virtualize your storage to get efficiencies, but what you do need is storage that’s optimized for a virtual server environments.” Storage benefits from converged environments. We’ll be back soon with a look at converged architectures, which will wrap an even bigger context around storage and next-gen IT strategies.