Answers on artificial intelligence, IoT & analytics workloads at the edge

Aug. 18, 2020
A chat with an IBM expert on next-gen edge-computing techniques.

There is increasing demand to deploy and manage artificial intelligence, IoT and analytics workloads at the edge. And, of course, there are nagging questions and concerns as we mature these fields and try to move more data and apps between data centers, hybrid multicloud environments and the edge.

Here we connect with Evaristus Mainsah, IBM’s general manager for cloud, cloud pak and edge ecosystem, to explore some of the latest techniques. Take a look…  

Smart Industry: How do you define a "hybrid cloud offering"?

Evaristus: By hybrid cloud, we mean the ability to deploy and run workloads in both private and public clouds at the same time. In survey after survey, we find that the overwhelming majority of clients will need to operate both on-prem and multi-cloud workloads for the foreseeable future. Those clients are searching for solutions that deliver security, scale and flexibility, and the ability to manage that infrastructure as one, which makes an enterprise-grade open, secure architecture critical.

By adopting a hybrid approach, enterprises can run their solutions everywhere—from the data center, across several public and private clouds, all the way to the edge. Because our solutions are built on open source and Red Hat OpenShift, they enable the client to use edge devices from multiple providers while still experiencing a seamless and robust edge experience. This unique ability helps enterprises manage workloads across a large volume of devices from a multitude of vendors—and manage workloads from the “cloud out” to “edge in,” using the same open horizontal architecture.

Smart Industry: How is 5G changing edge computing? 

Evaristus: The rollout of 5G enables manufacturers to more quickly and efficiently automate and analyze real-time data at the edge. IDC predicts that by 2023, half of the newly deployed on-premises infrastructure will be in critical edge locations rather than corporate data centers—and we’re at less than 10%  today.

While edge computing helps address issues stemming from today’s ever-increasing amounts of data, 5G networks enable lightning speeds and reliability in even more connected devices. It allows enterprises to tap into operations from even the most remote locations while retaining a high level of operational resilience. More importantly, it enables enterprises, industries and even whole countries to address new use cases and unleash new levels of innovation.

But 5G does not change the laws of physics—latency and roundtrip times still matter and require deployment of computing to the edge to deliver those use cases. In addition, there will continue to be the need to analyze data where it is created for cost, legal and other reasons. We see huge potential in our partner ecosystem to jointly deliver a 5G-enabled enterprise-compute platform to the edge. We expect to see more and more use cases in areas from telco through retail to Industry 4.0.

Smart Industry: What impact is COVID having on edge computing?

Evaristus: We believe it is critical to our economy that companies are able to resume their work operations while keeping their employees and customers safe. We also believe companies need to be able to do this without incurring unpredictable capital costs, and without exposing personal and private information.

In our new normal of a COVID-19 world, the potential of edge computing is more important than ever. Computing at the edge enables companies to stay on top of industry safety requirements and balance risks resulting from COVID-19 interruptions. Edge computing also lowers the latency to making critical decisions, reduces the bandwidth costs otherwise required to transmit large quantities of data to the cloud, and terminates the flow and exposure of potentially sensitive information.

Additionally, enterprises are looking to edge computing to keep their business operating and allow workers to perform tasks from anywhere. For example, the proliferation of Industrial IoT devices, combined with edge computing, can lead to smarter manufacturing and supply chains, equipping them to handle disruption of all kinds. Edge technology can enhance and expand the performance of drones used for a myriad of uses—from disinfection and diagnosis to deliveries. And edge computing combined with 5G can help address bandwidth, speed and security issues for networks experiencing sudden and ongoing traffic surges.

The pandemic has accelerated the rate of adoption for these technologies and has highlighted the potential that companies in multiple industries can leverage.

Smart Industry: What most excites you about the near future of edge computing?

Evaristus: From retail to banking to manufacturing, we believe enterprises across industries will be impacted by a new wave of innovation through working at the edge. In our new normal of remote work, edge computing makes it easier than ever to work with, process, analyze and gain insights from data where it is created—and to do so at scale.

However, if a hybrid cloud environment is complex because of the different types of compute, storage and networking, that complexity is exponential with equipment provided by different OT or device manufacturers, different server providers, different IT protocols and file formats, etc. In this type of environment, the need to work with an ecosystem is paramount, and the need to work with an open horizontal architecture is acute. That is why our commitment to open is so valued by our clients and partners. It is also why we take a distinctly ecosystem approach to working on the edge opportunity.

We have already seen impressive work from our ecosystem partners. These companies are helping their clients take advantage of 5G while capitalizing on insights closer to where data is created by people, places and things. And they do so while working freely with other ecosystem partners, tapping into innovation wherever it exists.