Architecting a Cloud-to-Edge Computing Strategy
In the meantime, organizations are largely on their own to manage edge data effectively. With edge devices expected to number in the billions over the next decade, it would be cost-prohibitive to store and process all the data they generate in the cloud. But the ability to analyze huge amounts of data in the cloud may prove lucrative in leveraging artificial intelligence and in improving and developing processes, products and services. So the challenge lies in determining what data to move to the cloud and how to manage it.
Not surprisingly, large cloud services are eager to help, as they have a stake in ensuring that their services encompass as much edge data as possible. AWS and Microsoft Azure have both introduced cloud-related technology that can be employed on premise in edge data centers, on edge devices such as gateways, in 5G networks, and even down to microcontrollers that operate individual IoT sensors.
“We offer a plethora tools and services which have one thing in common,” said Dirk Didascalou, AWS vice president with responsibility for IoT services, “that is in principle the way you write your application is the same: You only have to write it once,” whether deployment is in the cloud, on the factory floor, or in a 5G wireless network. “If you had to write applications for all those different scenarios, you’d create more complexity than you have today,” he added.
Other providers such as Google, IBM, VMware and Oracle have also implemented strategies to extend their cloud services to the edge.
Matthew Wopata, a U.S.-based senior analyst focused on industrial technologies for IoT Analytics of Hamburg, Germany, said that companies must evaluate whether they want to be locked-in to a single architectural stack, from the cloud all the way down to edge devices. “It’s a broader question of whether locked-in is the way to go because of the benefits you get from not having to support multiple systems, or is it a better way to go with a multicloud and multi-edge architecture and compatibility across all these providers using virtualization,” Wopata said.
Edge and cloud computing are far from mutually exclusive, but they represent competing interests.
“The systems need to work together but edge is its own environment,” said Kilton Hopkins, CEO and co-founder of Edgeworx, a developer of an open source edge software platform. “It’s almost like a republic, where edge is the local government and cloud is this federated entity which oversees everything, but needs to allow the edge to independently operate. I think that really is not palatable to an AWS or a Microsoft Azure.”
The stakes are high. In 2019, according to Synergy Research Group, spending on cloud infrastructure services—including infrastructure as a service, platform-as-a-service, and private cloud—for the first time exceeded spending on data center hardware and software, demonstrating how centralized computing has become the dominant force.