Datacenter Refresh Advice from Intel’s IT CTO

Gregg Wyant, Intel’s IT CTO is in the process of refreshing his data centers.  Like every other CTO, he has limited resources.  Not even he gets free CPUs and motherboards.  Gregg recently explained to me and a few others on a teleconference his Data Center Strategy.  He focused on the 8 principles that drive Intel’s strategy;

• Proactive Refresh – All data centers need refreshing, one of Gregg’s focuses is determining the appropriate timeline for this.  He derived a 4-year refresh cycle that ensures efficiency and capability gains.  Intel is finding regular server refreshes providing tangible business results, realizing huge,  value gains.
• Virtualizing Resources – New Intel Xeon 5600 server processors offer 6 cores and are used as the virtualization standard (i.e. enterprise private cloud).  Utilizing these cores for standard server tasks allows more done with less. Virtualized servers are also easily re-configured, cutting down on configuration timetables.  New Intel Xeon 7500 server processors offer 8 cores and ared used where    reliability and large memory are vital to application performance and availability.  Examples of this type of usage are large application instances in ERP and design of their silicon as part of the high performance computing platform.
• Right-sized Server Purchase – Underpowered servers can create cost overruns, requiring additional IT staff and cycles.  Likewise, overpowered servers require great cost outlays at the start of the cycles, and if under-used never create the ROI they advertised.
• Improvement of Energy Efficiency – Intel has been focusing on achieving extremely powerful yet energy efficient server processors.  This is achieved by smaller footprints and more efficient architecture around the processor as well.
• Facility Upgrades/Updates – Upgrading your facilities to the latest datacenter designs can help cut down on costs and increase efficiency.  In addition, upgrading facilities can make IT investments last longer.
• Modular Additions for Capacity – Adding capacity modularly cuts down on “Fork Lift” style updates.  The ability to just add small stacks of capability when it is needed creates flexibility and agility.
• Optimization of Networks – Managing networks can take an awful lot of computing power and rack space.  Optimized network management can reduce computing demands and IT administration needs.
• Enhancing Storage Solutions – Storage servers often take a large amount of datacenter floorspace.  Hard drives are getting more and more inexpensive, and there are many data de-duplication devices solutions on the market.  This enables datacenter designers to compress storage into less floorspace, liberating racks for other uses.

So what has dedication to these 8 principles brought Gregg?  Well he’s finding that with every refresh, they gain 250% capability, at 60% of the cost.  Through consolidation, he aims to pay off the upgrades within a year.  Intel is finding they need 45% more computing power and 35% more storage capacity every year.  Keeping up with these demands is tough, but as Gregg continues to increase capabilities, meeting needs is possible.