If there’s one thing people in the tech industry are talking about—other than AI—it’s data centers. The critical infrastructure is becoming even more important in an era where computing technology relies on information and developers are rapidly building out capacity.
Despite that increase in production, however, data centers are also being decommissioned at a rapid rate. In large part, this is due to an increased reliance on the usage of offsite hyperscale data centers, which has resulted in the closure of on-premises sites. But with the rise of larger off-prem centers comes the risk that overproduction could lead to a spate of decommissioning in the future.
Process counts. Decommissioning involves migration of system data and workloads and wiping, or sanitizing, data on older hardware. Migration is a relatively simple process, F5 CTO of Government Solutions Bill Church said, because the risk involved is tempered by how much it’s spread out.
“It’s almost inconceivable that we would design an application today that would not have some sort of fallback or backup capacity,” Church told IT Brew.
But overall, decommissioning is a procedure that carries with it no small amount of security risk, as NCC Group Senior Advisor and Director of Security Tim Rawlins told IT Brew. Hardware that’s been cleared of data often retains information.
“Time and time again, it’s been shown that if you go out onto eBay and buy second-hand computing equipment, there’s a very good chance that it hasn’t been properly wiped, that there is still data on that system,” Rawlins said.
Complication station. It’s not cheap. Kiersten Todt, former CISA chief of staff and current president of communications firm Wondros, noted that the cost of decommissioning often starts with moving assets and data to the cloud. Transitioning to a new facility is frequently the reason for decommissioning; managing, minimizing, and mitigating costs are the priorities for organizations.
Top insights for IT pros
From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.
“As this data center boom continues, and as we’re seeing more of this—both from an energy perspective, but also physical—cloud first are easy to scale, and it helps to sync and create the efficiencies across them,” Todt said.
The process of decommissioning comes with other concerns. Energy and infrastructural hook-ups to the grid are frequently important, whether you’re seeing strict decommissioning or augmentation, Energy Exemplar CEO David Wilson told IT Brew. Scaling up existing sites that are operating on older generation technology is more frequently seen, meaning that when these sites are closed by one organization, another is only too happy to pick up the slack. That’s not unique.
“Typically, what we’re seeing in the US is…whenever a large load is disconnected from the grid, whether that’s a data center or an aluminum smelter or some other industrial process, there’s a list of candidates who are quite happy to pick up that load,” Wilson said.
Energizer. Disconnecting the center from the grid is easier than connecting it. Energy research firm Rystad Energy found in February that in excess of 2 terawatts of potential projects are awaiting connection approval, but “historically, the approval rate for queued projects has hovered around 15%.”
Even if the centers are augmented, rather than being decommissioned outright, the infrastructure has a finite lifespan. Eventually the assets in the centers will become obsolete, and with the technology moving forward rapidly, the capacity and scale of existing facilities is often already behind the times.
“They just become uncompetitive because they’re sub-scale, and they will retire,” Wilson said.