3dmentat - Fotolia
LAS VEGAS -- Servers are breaking free from the data center, finding a home close to the data that is forecasted to start gushing from internet of things devices.
The concept is being billed by Hewlett Packard Enterprise (HPE) as a new product category, called converged systems for the internet of things (IoT). Doing analytics at the edge is not a new concept -- Cisco and IBM have products targeting this -- and it will increasingly become a big play for servers and compute technology in the future, said Kuba Stolarski, research director at IDC.
"The capability will always be deeper at the core than the edge, but you will be able to do a lot of analytics at the edge," he said about new edge servers. "Whether that is servers or not, something needs to be deployed there."
The Edgeline EL1000 and Edgeline EL4000 combine x86 compute with data capture, control and storage in an appliance that sits outside the data center, and at the edge for analytics and insights for real-time decision making.
"This is not a router running modified analytics; this is not an ARM core -- this is true, battle-hardened, enterprise-class compute," said Tom Bradicich, vice president and general manager for servers and IoT systems at HPE.
On the show floor at HPE Discover, one of the Edgeline servers was running in conjunction with a pump, illustrating a likely edge computing scenario of a machine that needs monitoring and maintenance, noted Jarom Petrich, platform engineer at Brigham Young University in Provo, Utah. While he didn't see any immediate use for edge servers at the university, as an IT pro, he sees likely uses across industries from soft drink bottling to bicycle riding.
Jarom Petrichplatform engineer at Brigham Young University
"When you need data instantly, [in] real time, I think having an edge server makes sense," he said. "There is always going to be latency, even if it is an edge server, but the latency will practically be zero."
This HPE product strategy focuses on the idea that data doesn't always have to be moved back to the core to be analyzed, and decisions need to be made with every data point.
As internet of things devices become more common, the majority of data will stay at the edge, with maybe only 2% making its way back to the core for deeper analysis, Stolarski said.
Doing analytics at the edge makes sense, even if it may not be that deep, such as event monitoring based around existing triggers, marketing something to a customer in a store or fixing something that's broken.
Big organizations with edge locations, such as retail stores, elevators, oil rigs and more, "will be critical -- not everything can be deployed in the cloud, something has to reside there," he said. "It may not be gangbusters, but certainly there are more use cases emerging of edge location deployment."
Edge analytics: Get out of my house
Nick East, CEO and cofounder of Zynstra Ltd., a hybrid IT provider in Bath, U.K., has a retail customer in England with more than 2,000 locations where edge servers could fit.
"We went in and the guy says, 'The last thing we want is to have servers in all our stores; now, we have all these great IoT use cases and they are throwing off so much data -- we know it is going to be too much,'" East said.
The company knows it needs edge compute, but still had a problem with multiple workloads that would need truck rolls. East called it classic management of servers across a distributed state.
"We want a template, we branch out everyone and manage across the cloud, à la distributed Wi-Fi networks," where admins don't need to log in to every device, East said.
It is better to act on the edge, rather than clog up the core, Stolarski said.
"IoT has been around for several years, but it is coming up in more places," and is driving infrastructure builds both on premises and some in the cloud, he said.
Robert Gates covers data centers, data center strategies, server technologies, converged and hyper-converged infrastructure and open source operating systems for SearchDataCenter. Follow him on Twitter @RBGatesTT or email him at [email protected].
Read up on the future of IoT
Get your data center ready for the IoT era
IoT devices introduce new data center security concerns