Hewlett-Packard thinks that we are on the edge of a major change in infrastructure, which will see the break-down of the traditional IT silos of storage, server, and network. In its vision, infrastructure will soon be sold, configured, and consumed in unified logical building-blocks. And it is acting on this vision by selling infrastructure-in-a-box – standard pre-integrated and packaged IT blocks that, in the words of HP EVP David Donatelli, can be drop-shipped to a customer location, plugged in, and turned on to provide a complete data center or a unified expansion of an existing computing environment.
The logic behind this strategy comes from two directions and is compelling. On the one side, virtualization, which for the first time allows applications to effectively share resources across the infrastructure, using whatever resources they need at a given moment, makes it possible to house multiple core applications on a logical infrastructure block without having to tune specific servers and storage to the needs of an individual application to the detriment of the others.
From that standpoint the attraction of pre-configured logical infrastructure building-blocks is simplicity – it extends the concept of “plug-and-play” to core infrastructure. An entire data center can be delivered off a truck and turned on as soon as it is plugged in, without all the configuration and integration and fine tuning of traditional piece-part architecture. And maintenance becomes equally simple. When the entire infrastructure is a single SKU, the vendor, in this case HP, delivers comprehensive upgrades, maintenance, and, when needed, service, that are pretested and guaranteed to be load-and-run, with no manual intervention at the customer site needed.
On the other side, private- and hybrid-cloud computing and virtualization, the twin trends transforming the data center, require a more integrated infrastructure with server, storage, and network groups working together. While islands of tight integration can be built piece-meal, creating an integrated infrastructure from the typical “one-of-everything” environment in many data centers, is challenging at best and at worst a recipe for a “storm cloud” that delivers more lightning bolts than rainbows. Buying the infrastructure pre-integrated from a single vendor or tightly knit consortium can save money in the long run by avoiding hundreds of person-hours trying to integrate the pieces lying around the data center while making IT look like a hero to business executives demanding higher service at lower overall cost.
“it's all about making things simple for the customer,” argues HP VP of Storage Marketing Craig Nunes in an interview on SiliconAngle.tv from SNW in early April 2011 and transcribed on Wikibon.com. “If you can reduce 30 SKUs to one and make it stupidly simple to manage, it makes things much simpler and answers the customer's number one question about cloud computing – how do I get there.”
Action Item: Buying infrastructure in logical blocks can solve many of the problems of realizing the compelling advantages of the new virtualized, hybrid-cloud architecture. It may seem expensive at first glance to replace perfectly good technology and in some cases overbuying in some categories, but the incredible efficiency boost of virtualizing and moving from 15% to 80% utilization more than makes up for those concerns. And the total cost of ownership of these logical blocks, which virtually eliminate the need for customization and the complexities of upgrading and servicing those highly customized environments, will be much lower.