Edge computing is shifting workloads, data, processing, and business value away from the cloud and toward global and distributed practices for low-latency and real-time processing.
While the cloud provides massive centralization, economies of scale, and self-service, it simply cannot keep up with the growing compute data and resources needed by businesses to effectively handle real-time interactions and decisions. Waiting on a data center miles away isn’t going to work much longer.
Five leaders in the edge technology space, including Ajay Kapur, Edgio’s CTO, sat together on a panel at TechEx’s Edge Computing Expo to discuss their perspectives on the current edge landscape. The discussion covered the common challenges businesses face when implementing an edge-first strategy, as well as the required solutions to make the edge transformation journey as successful as possible.
One common thread shared between panelists is that edge-native solutions can, should, and will enable global, scaled, and entirely automated workloads.
Panel: Enabling Your Organization’s Edge Computing Transformation Journey
Grant Challenger, Director – Edge Computing, VMware
Blake Kerrigan, Senior Director of ThinkEdge Business Group, ISG, Lenovo
Tom Reyes, Chief Product Officer, Stackpath
Rob High, IBM Fellow, Vice President and CTO, IBM Networking and Edge Computing Software, IBM Edge Computing
Ajay Kapur, Chief Technology Officer, Edgio
How do you see the edge landscape shaping and how do you see your organization’s position in it?
Ajay Kapur (Edgio): When we talk about the edge it must be global and a service. If it’s not, you’re just hosting compute – and that alone will not compete with AWS. If you have a truly global service, for example, if you can flip on security globally and get it running across all your hybrid cloud environments – that’s the differentiating value. The edge is on its way to being global and as a service, as opposed to getting servers in your local region.
Rob High (IBM): To unlock the value of edge computing, you need to apply it to solving a real business problem. Whether it’s industrial problems such as production optimization, safety, distribution and logistics, or retail optimizations – clients want to leverage the right technology at scale. Today, edge computing takes advantage of the growing compute capacity to harness solutions, no matter your industry.
Tom Reyes (Stackpath): In urban dense areas, we are seeing customers lose interest in business value in a matter of seconds. We will graduate to the next phase of the edge if we provide developers with the tools that allow them to fine-tune their applications to the edge. Once they can spin up what they need and when they need it – and deploy online content immediately worldwide – we will graduate from the “fast-walk” stage to the “run” stage of edge maturity. Growth lies within mid-market level players who have a legitimate use case that isn’t currently being solved at hyper-scale levels.
Blake Kerrigan (Lenovo): With edge computing, I see two common themes. The first is enterprise workload consolidation is happening quickly. The second is that accelerated workloads need engineering excellence. We must solve these things to see edge compute mature.
Grant Challenger (VMware): The future of the cloud will be the edge. You don’t make a car, bottles, or food, in the cloud. You make it local, and those people who make it consume compute resources. As things become more software-defined, people will consume data at the same time as raw material. As a result, they’ll need compute and cloud-like functions at the edge: cloud-native to edge native. Edge native is the automated scaling and distribution of workloads where they’re needed, regardless of location. The cloud is the edge.
What is the biggest challenge in the edge computing movement?
Rob High (IBM): The biggest challenge will be adapting native development practices into edge native practices. We have to recognize that we’ve been through this journey with distributed computing – and almost every time we’ve failed on the issue of how we manage and administer these environments so our apps are out there at the right place and time, securely. Client server and SOA failed on this. We’ve seen this succeed when the administration problem moves to the end user, so they’re responsible for getting apps on devices, keeping them up to date, etc. But that doesn’t work in a commercial enterprise business because you don’t always have someone at those devices – so solving this problem is essential.
Blake Kerrigan (Lenovo): To me, the biggest challenge is for a customer to identify ‘what is the proof of value for an implementation?’ And, ‘how do you measure what type of value a particular piece of technology creates for your customer?’ These are the questions that every retailer, end user, midde-ware provider has to ask themselves along this journey. I also think collaboration is an important challenge to solve. Somehow we have to do a better job in the last go-round with IoT with collaboration – especially with open source.
Grant Challenger (VMware): The major challenge with edge is that OT (Operational technology) is not in charge and doesn’t own the budget for deploying edge. If they did, we’d have plenty of compute at the edge and use cases – IT would deploy all the time. But we have a fundamental disconnect between IT and OT and the requirements needed to deploy edge for specific use cases.
Question from the audience: From my point of view I see the collection, aggregation, and privacy of data as one of the biggest challenges of edge computing. Does anyone else?
Rob High (IBM): I’d like to share a story of how the edge has been an advantage to the data problem. One client was using facial recognition software to send an image of customers to the cloud to assist with in-store experience. The risk was transmitting a customer’s personal information over the network. Once that data leaves the premises, you’re increasing the potential for someone to hijack this data. With edge, you can do the AI inferencing in the store, ID it, and immediately throw it away. We see the same thing in manufacturing when we see AI being used for quality inspection, workers safety, etc. One of the advantages that edge computing brings is the ability to protect that data in ways you can’t if you’re shipping it all back to AWS.
Grant Challenger (VMware): I challenge this. The notion of GDPR and data-sensitive topics get resolved by the owner of that data. You have to design for addressing it. The real problem is the volume of the data, the ability to understand what it is, and what you have to do to protect it. You must know how to handle and interpret that data in its real-time, binary state. If you don’t design a solution that deals with data sovereignty, you’re not going to live in the edge domain.
Ajay Kapur (Edgio): Solving the data problem at these different levels and for various use cases is hard – and it’s where the edge can bring real value. Certain platforms make this really easy to do, so customers aren’t stuck trying to figure this out for themselves. Global, consistent, and distributed key-value storage is really important if you’re trying to build a global application. If you’re going back to the core to get your data, the efficiency and security that you’re doing at the edge are wiped out.
What do you think is a crucial action or measure to make the edge transformation journey successful and sustainable?
Blake Kerrigan (Lenovo): We’ll never be successful by only focusing on outcomes. There’s a two-prong problem that’s happening faster than data can be generated or captured: consolidation of workloads and acceleration. My big leave behind is that once you’ve zeroed in on the outcome you’re trying to achieve, then scale should be your next priority. The goal is to to end in that nirvana state of managed, secure, and reliable.
Rob High (IBM): You must identify what the business outcome is that you’re trying to produce. At the end of the day latency matters, cost of transmission matters, business resiliency matters – and edge compute lends a hand to all of these to unlock business growth. However, these benefits can be superseded by the cost of managing the edge. If we don’t standardize, operationalize or create edge-native development practices that are efficient, scalable, not tied to a particular vendor – the cost can easily implode the value. Our responsibility as technologists is to solve those problems so businesses can carry on with their goals. Tom Reyes (Stackpath): It’s all about how customers can laser focus on solving their business problems. The hero use cases we see in the space are distributed business processing, data and event reporting, network optimization, and content delivery that allows customers to fine tune for their specific requirements. We’ll see more enduring hero use cases in 2023 and 2024 revolving around latency, data, and total cost ownership.
Ajay Kapur (Edgio): Security is a perfect use case of the edge. If you want to secure your application and you’re in the transition of being hybrid cloud or multi-cloud, the only place you can do it is at the perimeter outside all these things. That way security is agnostic to anything else you’re doing. A given app might have 2-3 backends and data might be in a hybrid environment or multiple clouds, so the only place you can protect everything is at the edge. Our approach provides WAF, DDoS, bot mitigation, and API protection at the edge layer within milliseconds of the end user.
Grant Challenger (VMware): My suggestion is to build with the cloud in mind – edge native. Build with the IT organization or operations in mind. Those folks are generally in need of these solutions. Build on a platform that is multi-cloud because you’ll need management at scale. And change your perspective on the adoption of new security models at the edge because today’s solutions will not enable scale.
Interested in learning more about how edge-enabled technology can positively impact your business? Speak to our edge experts to learn more about the innovations from Edgio. Edgio is the world’s fastest and most programmable edge network.