Microsoft wants to reduce latency with underwater data centers

Speaking at the company's Future Decoded conference in London, Nadella said that underwater data center deployments are the way Microsoft will think about regions and data center expansion. He cites proximity as a particular advantage: Approximately 50 percent of the world's population lives within 120 miles of the coast.

Underwater data centers are Microsoft's project

Placing servers in the ocean means that they can be close to population centers, which in turn ensures lower latencies. Low latencies are particularly important for real-time services, including Microsoft's upcoming Xcloud game streaming service.

We recommend reading our article on Microsoft preparing modular controls for Project xCloud

Microsoft has already been experimenting with underwater servers for some time. The Natick Project placed an underwater server pod off the California coast in 2016. Naturally, the capsule uses chilled water and pours residual heat into the surrounding ocean. It is designed as a sealed unit, deployed for five years before returning to the surface and replacing it. Since then, Microsoft has deployed a larger capsule off the coast of Scotland.

The other great advantage Nadella cited from underwater data centers is the speed with which servers can be deployed in this manner. Without the need to build a real data center, he said that from start to finish, the Scottish pod took just 90 days to build and deploy. This shorter time to market means the company can be reactive, adding additional server capacity close to where it is needed on demand. This is in contrast to terrestrial data centers, where the business has to guess what the future demand will be and therefore how large a site is required.

The Scottish deployment is powered by wind energy. As the cost of offshore wind generation continues to decline, you can even imagine that these offshore data centers could be combined with offshore wind farms.

Neowin font