Optimizing costs within the Data Center

Let me begin by way of introduction - I am a strategic financial analyst with Intel IT Finance organization focused on data center strategy and efficiency efforts.  This is my maiden voyage into the world of blogging, so I hope the topic is relevant and interesting to the audience.

Similar to many organizations, Intel IT is focused on constantly improving the cost of keeping the business running while not sacrificing the level of support required by customers.  With industry and technology solutions evolving at an increasing pace, choosing the most appropriate place and time to invest is paramount to driving down infrastructure costs.  Budget constraints in this economic climate and the make implementing efficiency efforts all the more daunting.

In 2008, Intel IT initiated a Design Server Refresh strategy where the basic premise was to leverage server performance improvements to respond to increasing compute requirements without growing data center capacity at a corresponding rate.  In 2008, we were able to remove 20,000 single core servers from our production environment, allowing us to realize approximately $45M savings through avoiding data center additions and server operating costs.  However, even with this strategy driving significant near term results, the 2009 operating environment forced us to pause and re-evaluate the merits of continuing execution to the strategy.

This re-evaluation concluded that this was an investment that couldn't be deferred due to the need for incremental growth and the high utilization of our existing data centers.  In addition, based on a average 10:1 consolidation, the refresh of single core servers would generate significant operating savings and clear more headroom than seen historically.  The details of this analysis are included in the White Paper:  /javascript:;

Questions for the readers: Do others have a refresh strategy or guideline? Are others seeing this type of impact/results and the challenges in implementation?