As 5G networks become more widely available it’s now possible to deploy a wider range of applications capable of processing and analyzing data as it is collected in real time. Rather than having to transfer large amounts of data over a wide area network (WAN) to be processed in the cloud, it’s feasible to deploy applications on an edge computing platform.
Case in point is Siemens and IBM, which are working together to build and deploy applications on a Red Hat OpenShift platform based on Kubernetes on factory floors. Those applications will analyze data created by operational technology (OT) systems as part of an effort to bridge a divide that has historically existed between OT platforms and traditional IT applications.
That effort only becomes practical when there is enough bandwidth to collect the data needed to drive those applications, says Raymond Kok, senior vice president of cloud application solutions for Siemens Digital Industries Software.
“There is quite a lot of convergence involved in what we are doing,” says Kok.
Scaling the Edge
Naturally, providers of IT platforms are jockeying to become the providers of the edge computing platforms that run those applications. A wide range of processors are currently employed in OT environments today. Intel, however, is making a case for using industry standard processors in an edge computing platform to lower the total cost. Application code will ultimately need to shift between the edge to the local data center and the cloud. It becomes easier to achieve that goal at scale when all of the platforms employed are based on a common architecture, says Dan Rodriguez, corporate vice president and general manager for the Network Platforms Group at Intel.
“You need to invest in an architecture you can grow,” says Rodriguez.
Arguably, latency is the biggest edge computing challenge IT organizations will encounter when deploying edge computing platforms at scale. Many organizations are turning to providers of content delivery networks (CDNs) that have already deployed highly distributed infrastructure to narrow the gap between edge computing applications and the rest of their IT environment.
Also read: The Impact of 5G on Cloud Computing
The PoP Approach
In much the same way CDNs are employed to cache Web applications at the network edge to improve the network experience, organizations are starting to employ CDNs from provider such as Akamai to provide similar capabilities to edge computing applications deployed in points-of-presence (PoP) managed by the CDN service provider.
That approach makes it possible to deliver, for example, over-the-air updates to applications residing in vehicles connected to a 5G network, says Ari Weil, vice president of product and industry marketing for Akamai.
Most organizations will not have the financial resources required to build and deploy the infrastructure required to drive latency-sensitive edge computing applications at scale so it makes more sense to leverage a service, adds Weil. “You need to be able to deliver real-time response times at the edge.”
Facing Cyber Threats
Securing these applications will also become an issue as cybercriminals begin to realize that compromising a single edge computing platform could become the gateway to laterally distributed malware throughout an entire enterprise.
To address that issue, Canonical just released Ubuntu Core 20, an update to its distribution of Linux that can be deployed in a container on both Intel and Arm processors. The update adds secure boot, full disk encryption, and secure device recovery capabilities that will soon become required capabilities to meet compliance mandates, says Galem Kayo, product manager for Ubuntu Core at Canonical.
Palo Alto Networks, meanwhile, has launched a security initiative specific to 5G networks that makes use of containers to deploy security functions on edge computing platforms that will play a major role in digital business transformation initiatives.
“The bad guys see what is happening with digital business transformation,” says Anand Oswal, senior vice president and general manager for firewall as a platform at Palo Alto Networks.
Zededa has similarly launched a namesake open orchestration platform for securing both virtual machines and containers running at the network edge. The goal is to enable organizations to deploy a zero-trust framework to secure any endpoint platform or device regardless of when it was deployed, says Jason Shepherd, vice president of ecosystem for Zededa. “There are a lot of legacy edge computing platforms that need to be secured,” he says.
It’s probable that one day soon there will be more workloads thanks in part to 5G running on edge computing platforms than in the cloud. The challenge IT organizations need to gear up for now is finding a way to deploy, update, manage, and secure those workloads that will most often be physically located in some of the most inaccessible places ever imagined.