The Future of Edge Computing

A Q&A With The Linux Foundation’s Arpit Joshipura



Steven Lerner
02/08/2019

More data is being creating now than ever before, and enterprises need to analyze that data in real time. Edge computing has emerged as the latest solution to enable real-time insights of data.

In January 2019, the non-profit Linux Foundation launched a new umbrella organization aimed at providing harmonization to accelerate deployment among the rapidly growing number of edge computing devices. Named LF Edge, the organization is establishing a unified open source framework for edge computing that is vendor agonistic.

The development could help enterprises leverage the future of edge computing and the Internet of Things (IoT). It provides a horizontal framework with open APIs that have full central visibility and control.

The LF Edge will initially be comprised of five open source projects:

  • Akraino Edge Stack: an open source software stack that supports high-availability cloud services.
  • EdgeX Foundry: another project focused on an IoT framework.
  • Open Glossary of Edge Computing: a Wikipedia-style collection of information about edge computing.
  • Home Edge Project: an open source project from Samsung focused on home-based devices.
  • Project EVE: an edge virtualization project focused on complex on- and off-premise networks and applications.

For the purposes of this initiative, the Linux Foundation defines an edge application as anything that requires five to 20 milliseconds of latency. That means that an IoT application that is collecting data in a factory and sending it to a public cloud would not be defined as an edge application.

To understand more about the implications of edge computing and LF Edge, Enterprise Mobility Exchange spoke exclusively with Arpit Joshipura, general manager of the Linux Foundation.

Enterprise Mobility Exchange: How should enterprises prepare for edge computing now that the LF Edge has been announced?

Arpit Joshipura: First, look at the vertical specific environments. Generally speaking, industrial manufacturing, energy, and retail are all already moving in the right direction. I would say a couple of things are near on the horizon. Specifically when it comes to manufacturing, you have a whole new set of open hardware that's coming out (that's ARM, Intel, and Qualcomm-based) that will have virtualization engines built in. If you have factories and if you have data centers that you're using, now you can put them in and automatically connect to your cloud, so that you're not dependent on just one cloud. You should prepare for a cloud agnostic and vendor agnostic solution. It translates into cost savings and simplicity. The second thing the enterprise should be aware of is that telecoms are starting to put test gear in several of these edge data centers. The advantage that they have is the real estate that is physically close to your business. When you have telecom companies providing you with services, you now have an option to take their SLAs, get better responsiveness and get quicker turnarounds on applications that are important to you in the edge space. Of course, 5G is a big trend here with a whole new set of applications coming in because you have lower latency. Now the next phase of that are the applications and the workloads that are now latency sensitive and time sensitive. They can now be housed closer to your enterprise, and that's what brings tremendous value to the enterprise.

EME: What challenges does LF Edge try to solve for the enterprise?

AJ: Tech companies are selling products, services, and solutions to the enterprise. Now, what they can do, if they can utilize LF Edge and the code that gets created in the community, obviously through their participation and contribution, is provide products and services to an enterprise that was otherwise not possible. The challenge is again going to be: does the enterprise invest in open source directly or do they go through the ecosystem to purchase, but get more control? And I think we'll see a little bit of both. Some of the largest enterprises will just adopt directly. [Cloud-heavy enterprises] are definitely going to adopt directly. But some of the, you know, midsize enterprises will definitely look at the big tech companies of the world to help them move the agenda forward.

EME: What does the future of edge computing look like in the enterprise?

AJ: There are several use cases, and again, and I hate to bring it down to a vertical but there's no one common answer. We believe that there's a very strong case if you assume that these technologies are here to stay. The first and the most important impact is that applications that an enterprise created are now written in a CI/CD development world with microservices and kubernetes. They're extremely portable and lightweight. They can go across clouds, they can go across edge clouds, and they can go anywhere they want. The future of an enterprise really is an extremely well-connected network. Historically, today if you ask an IT person, “Where's your data center?” they might say, “Oh, it's downstairs,” or “It's at Amazon.” Now, the answer will be, “what app do you want to know about and what service do you want?” Automatically, I can just figure out whether you want to put it on premises, whether you want to put it at the edge, or whether you want to take it back to the core.

To discover more about this new initiative, please go to the LF Edge website.