Is Your Data Center Becoming More ICU than CPU?
I recently came a cross a study performed by the Anthesis Group and Jonathan Koomey, a Research Fellow at Stanford University, which centered on server utilization. Throughout the study they looked at the growing epidemic of what they call “comatose servers.” By their definition, a comatose server is one that has not delivered information or computing services in at least six months, but is still live.
In other words, the lights are on but nothing else is running at home. There are various reasons for why this occurs from overzealous projects that never come to fruition, redundancy of hardware within your IT sparing program, or left over hardware that was virtualized within an IT shift. Regardless of the reason, comatose IT hardware exists in the majority of data centers and IT environments throughout the world.
In fact, according to the research, as much as 30% of infrastructure environments consist of comatose servers. Maybe that does not seem like much, or just an accepted industry norm, however there are some significant impacts on business costs and productivity as a result.
Top 3 Impacts of Comatose Hardware on Your Bottom LineOpportunity Lost
If you estimate an average cost of $3,000 per server, the capital expenditure utilized to obtain IT hardware quickly becomes a sunk cost with very little return. Multiply that by 30% of your server bed and you have quite the cost to simply maintain your existing infrastructure instead of utilizing it for other IT and Engineering projects that are on the table.
Energy Saving Costs
According to the Uptime Institute’s Savings Calculator, a 1000 server environment consisting of 20% comatose servers would result in just under $100,000 in unnecessary annual energy costs. If you turned those unnecessary costs into infrastructure investment dollars for future projects and IT initiatives, your return becomes exponential.
Many times costly maintenance agreements are being carried on these comatose, non-performing servers or networking hardware. These costs are easily lost in the shuffle especially when it comes to multiple OEM agreements across multiple contracts. Add in auto-renewals, and you find yourself paying for unnecessary support 10 times over.
Above and beyond the initial energy cost and operational savings from removing comatose IT servers and networking hardware, you could increase the return on your capital investment by redeploying or reselling hardware where needed. In particular, idle IT hardware can be redeployed into other existing IT environments and utilized to upgrade older or end-of-life hardware or backfill an integrated spares program. If it is determined to have no additional internal use, there is still value that can be recouped on the open market to help offset initial and future investment costs.
Cut the cord on the data center ICU, utilize active hardware to its fullest capacity, minimize your operational costs, and invest in your future IT infrastructure.