US government has 12,166 data centers and most are inefficient: Can hybrid save the day?
The US Federal government hasn't come close to consolidating its data center footprint, failed to optimize operations and has utilized hybrid computing in only limited fashion, according to a General Accountability Office report.
GAO's report, which weighs in as a 98 page PDF, is an indictment of US government technology operations. While it notes a few consolidation and optimization wins, the US government's data center footprint is as bloated as ever. In 2010, the OMB launched a Federal data center consolidation effort with 2018 savings targets. While there were some successes, the Feds have failed with their data center efforts.
According to the GAO report:
Federal data center consolidation (DCOI) efforts have been underway since 2010 and OMB's fiscal year 2018 targets provided clear and transparent goals that helped define the tangible benefits that DCOI was expected to provide. However, most agencies continue to report mixed progress against those targets. Although agencies have taken action to close about half of the data centers in their combined inventories, 11 agencies did not plan to meet all of their closure targets. Further, the data center closures were expected to drive cost savings and avoidances and, to the agencies' credit, the closures have led to more than $2.37 billion in planned and achieved cost savings and avoidances from fiscal years 2016 through 2018. However, five agencies did not plan to meet their cost savings targets.
Until agencies consolidate the data centers required to meet their targets, as well as identify and report the associated cost savings, they will be challenged to realize expected efficiencies and the full benefits of DCOI will not be fully realized.
From there, the GAO calls out commissioners and agencies failing to hit data center closure targets for more than three pages of the report.
Part of the solution to this data center conundrum could be a hybrid approach. After all, it's not realistic to assume the government is going to lift and shift everything to the cloud. In addition, the GAO noted the agencies didn't use server utilization and automated monitoring, virtualization or energy metering.
Perhaps Federal agencies can learn from a few successes. Consider:
- The Social Security Administration adopted hybrid cloud with public and privacy services as a way to modernize IT. Back office applications have gone hybrid and have allowed software and hardware systems to be simplified.
- Some agencies such as the National Oceanic and Atmospheric Administration (NOAA) have gone cloud-first over data centers. NOAA didn't have the capacity with their own data centers so wound up utilizing cloud.
- The General Services Agency (GSA) moved some of its data centers to cloud and tiered their infrastructure. As a result, GSA closed 118 data centers as of August 2018.
- Other agencies focused on closing data centers and improving optimization. The Agriculture Department said it closed 2,253 data centers through August.
- The Justice Dept. kept more efficient data centers and optimized via virtualization and shuttered inefficient facilities. GAO said the Justice Dept. closed 84 out of 110 of its data centers.
According to the GAO, the agencies that hit their data center closure targets had the following best practices.
- Obtaining executive leadership support for consolidation and optimization activities;
- Using experiences and lessons learned to refine consolidation planning;
- Increasing the use of cloud and shared services to consolidate or optimize data center operations;
- Emphasizing closing data centers to meet OMB targets and achieve cost savings;
- Increasing the use of virtualization to optimize data centers; and
- Employing an organization-wide communications plan to facilitate adoption of consolidation and optimization activities.
Photos: Inside vast abandoned mine set to be world's biggest data center
Related stories:
- Top cloud providers 2019: AWS, Microsoft Azure, Google Cloud
- Vendor comparison: Azure, Amazon AWS, and Google Cloud Tech Pro Research
- Data centers: The future is software-defined
- Survey: Despite increases in data, companies reluctant to automate data center functions
- Eight signs you could be automating more of your data center