This is an opportune time to put a stake in the ground – to talk about the ongoing costs of operating the U.S. federal government's data centers.
Currently, agencies are working to comply with the Data Center Consolidation Initiative (DCOI), which requires them to count the number of datacenters they operate (the total number is in the thousands) and to also count servers, associated hardware and software, plus the types of IT solutions being supported. Agencies are required to track the amount being spent on various components within their datacenters, and they need to make plans to reduce those costs.
However, as we focus on potential cost reductions, IDC does not anticipate that many agencies will see an immediate decrease in data center spending. Here's why: It can cost money to migrate current solutions to new platforms, and it's not always easy to simply turn out the lights in a data center, even if it is slated for closure.
Also, the DCOI does not have any special funds set aside to help with the closures, so agencies will be tackling the effort as best they can with available budgets. Our experience has shown us that the government's legacy systems have a tendency to linger longer than anticipated when agencies work to move them to new platforms, especially when they lack the funds to make the change quickly.
Where the Money Goes
IDC believes that data center spending in the U.S. Federal government will experience slow growth between now and 2020. We do not expect to see broader, multi-agency cost reductions until after that timeframe. The total amount spent is expected to grow from $5.5 billion in FY15 to $5.8 billion in FY 2020. (Note: The numbers do not include telecommunication costs or connectivity fees for internet and other network access.)
The anticipated federal data center spending shows a relatively low-growth CAGR of 1.3% for the five-year period. A few years ago, agencies often averaged 5% to 6% growth for their annual IT budgets, so a lower CAGR is, at least, a bit of progress.
Fiscal Year 2017 has ended for the federal government, and we anticipate, once the numbers are tallied, that data spending will be slightly lower than FY 2016. But we don't expect that level of cost cutting to continue over the next three years.
In the long term, the DCOI effort will mean fewer datacenters and reduced management costs. It also will be effective in clearing out stand-alone servers that tend to be dedicated to a single project or department. Nearly all of these types of servers will be forced to move to shared datacenters and virtualized servers.
Details on the IDC Data Center Forecast
The newly published forecast focuses on overall U.S. federal spending related to government datacenters, with a specific break-out on costs related to hardware, software, labor, consumption of electricity, and facilities development or management. Cost are expected to vary considerably, agency by agency, through 2020. (The forecast follows 27 different major federal agencies or Departments.) We expect to see the growth average out as follows across all agencies.
Data center labor costs will see the lowest growth rate, with a .6% CAGR for the five-year period.
Facilities management should see the highest growth rate, as the data centers themselves will see significant changes – as surviving facilities grow and others are shuttered. A 2.8% CAGR is expected across all agencies.
Software spending should see growth measurable growth, with a 2.2% CAGR
Hardware spending should see a CAGR of 1.3%, driven in part by investments related to higher-end virtualized systems and more storage, to help handle larger data sets.
Electricity spending will see a CAGR of just 1%. Electric consumption will go up, but will be offset by cost reductions related to renewable energy.
Based on current consumption trends, it is likely that overall datacenter spending can be reduced substantially, starting around 2021. However, as the federal government moves more heavily into the Internet of Things, more data will be generated, and we expect there will be future investments in additional storage, databases, and data analytics.