Amazon Web Services (AWS) plans to supplement its ‘mega’ datacentre regions with smaller cloud data processing hubs in major cities to support enterprises requiring low-latency connections to its compute and storage resources
Amazon Web Services (AWS) is planning to build a series of hyper-local datacentre hubs close to where its users are in major cities, as it guns to get enterprises to shift more of their low-latency-demanding workloads to the cloud.
The hubs, dubbed Local Zones, will enable users to request access to AWS compute, storage and database resources sited closer to where they are when they have specific applications or workloads requiring single-digit millisecond latency connections to function properly.
The hubs will be kitted out with a variant of the AWS Outpost hardware, which enables enterprises to run public cloud workloads in their private datacentres using hardware that Amazon favours in its own server farms. As of 2 December 2019, the technology is on general release.
AWS CEO Andy Jassy also announced Local Zones during the second day keynote of the cloud giant’s annual Re:Invent user and developer conference in Las Vegas, where he positioned Local Zones as a means of providing users with a route to moving low-latency workloads off-premise.
“If you’re a media company in L.A or you do content creation in video games, those workloads need single digit, milli-second latency. Take a company in New York or Switzerland, in financial services that have to be close to market data. They need that single digit millisecond latency,” said Jassy.
“So we thought about it – is there a different type of construct we can provide you with that solves this issue at scale?”
Because such latency-sensitive workloads may be ill-suited to being hosted within one of the company’s “mega” datacentre regions around the world, the firm is building out smaller-scale AWS server farms to accommodate them.
“It’s very expensive to launch these mega regions, although we do have a lot more coming,” added Jassy.
The first Local Zone is live in Los Angeles, and will be accessible to users there via its US West AWS datacentre region, who can select resources hosted there through the AWS Management Console.
For other less latency-sensitive applications and workloads users might have, customers can continue to host them in AWS’s existing datacentre regions worldwide.
During a post-keynote Q&A with the press, AWS technical advisor, Joshua Burgin, said the firm has chosen Los Angeles as the launch site for Local Zones, and follow-on locations will be selected and announced in due course in response to customer demand.
“We have 12 more availability zones under development right now, four more datacentre regions around the world. And what you’ll see over time is that, as you get more customer demand, you’ll see a roadmap for Local Zones, much like the one we have right now for availability zones and regions,” he said.
The company has made frequent references during its most recent run of financial results about how it is ramping up its investment in its infrastructure to support new and emerging customer workloads and demands, and Local Zones appears to be one of the results of that.
The Re:Invent keynote also provided some additional clues about where else the firm has been concentrating its infrastructure investments, with AWS setting out plans to provide developers with access to 5G-based edge computing environments, through its newly announced AWS Wavelength offering.
Similar to Local Zones, Wavelength will enable developers to access AWS compute and storage resources at the edge of 5G networks hosted by the firm’s telecommunications partners, including Vodafone in Europe and Verizon in the US.
Corey Quinn, cloud economist at AWS monitoring firm The Duckbill Group, told Computer Weekly the creation of Wavelength is a savvy move on the company’s part, as it positions it at the forefront of 5G app development with a relatively low-level of outlay.
“AWS gets in on the ground floor of 5G with a minimum of investment on their part. It’s suddenly become the de-facto cloud provider for anyone building a 5G application, without having to attempt to predict what that’s going to look like in the market,” he said.
“It’s also bought itself access to the folks who are building those applications along the way, which will almost certainly inform their product roadmap going forward. Whatever happens with 5G, AWS will definitely be on the same wavelength.”