Over the last two decades now, cloud computing has dominated the business technology space, revolutionizing how companies operate and do business when it was introduced, as computers that were hundreds of miles apart in distance are now just milliseconds away in terms of communication.
With the rise of IoT, 5G, and our never-satisfied need for speed, edge computing is slowly starting to become common practice for companies of all sizes and industries.
Speaking technically, edge computing is a distributed computing paradigm that brings computation and data storage closer to the sources of data, which – improves response times and saves bandwidth. The COVID-19 pandemic further accelerated the already rapid growth of the edge, with the global edge computing market size being valued at USD 4.68 billion in 2020.
The edge is predicted to continue to grow, with a compound annual growth rate of 37.4 percent, bringing the market size to an expected USD 43.4 billion by 2027.
With forecasts like these, many industries are now gearing up for the possibility that the edge fully overtakes the cloud sometime soon. This topic was discussed in detail during the ZEDEDA Transform 2021, a virtual event earlier this year that brought together experts from across the edge computing and IoT landscape.
The opening keynote presentation was titled “Is the Edge Really Eating the Cloud?” and included four panelists and a moderator discussing key industry trends and debunking some common myths about the future of computing.
Moderated by Christian Renaud, Global Head of 451 Advisors within S&P Global Market Intelligence, the panel consisted of Said Ouissal, Founder and CEO at ZEDEDA, Tony Shakib, General Manager of Microsoft Azure IoT at Microsoft, Muneyb Minhazuddin, VP Edge Computing at VMware, and Keith Basil, VP of Product, Cloud Native Infrastructure at SUSE.
Renaud came right out the gate, first asking the panelists if they thought the edge was really eating into the cloud today.
“Cloud is not going anywhere. Nobody’s eating the cloud,” said Basil. “Though the edge is very important because I think the Linux Foundation predicted that the edge would be four times larger than the cloud. But I see not a pendulum swing from cloud to the edge from a decentralized perspective, but just a partnership or an extension between the two.”
“I think it’s going to change the cloud,” agreed Ouissal. “And I think it’s because the earlier point made is like the amount of data that is going to come from the edge is going to drive a different architecture of the cloud in some way. The edge is an architectural change of the cloud.”
The panelists were then asked if the edge isn’t going to fully take over the cloud, what was the reason for much of the recent hype surrounding the technology.
“I think we are now seeing edge mainstreaming and becoming real. And one of the reasons I believe that is real is there’s verticals that don’t exist in the cloud.,” said Ouissal. “Financials have always existed in data centers, and online companies exist on the Internet and in the cloud. But when you are an oil and gas company, or you’re a manufacturing company or retailer, or you’re an automotive company, your business is not done in the cloud. Your business is done in the physical world.”
“And as you go and move everything to the cloud, you need to connect your assets to the cloud; then edge computing becomes a topic. And that’s what we have seen,” continued Ouissal. “That the acceleration of cloud, especially in these many traditionally more lagging verticals, is now driving the adoption of edge. And I think it is absolutely beyond the hype at this point.”
Renaud then pivoted the discussion, asking the panelists which workloads today are already gravitating away from the cloud and towards the edge.
“There are three areas where it makes more sense for some of the workloads. One is the workloads that are mission-critical. The redundancy is required that if the cord is cut for whatever reason, they can’t stop the operation,” said Shakib. “The second one is where latency is really important, like when you have robotics making real-time decisions. Companies can’t wait for the round trip to the cloud before they shut down a robotic that’s operating a machine that’s near people or even autonomous vehicles. Then the third area that we see is where the data generation just outstrips the bandwidth that could be sent out to the cloud, especially in oil and gas and process manufacturing or even in retail.”
The panelists then discussed what considerations must be taken by companies today who are looking to implement and leverage edge technologies for their own benefit.
“My call out is we have to consider edge-native requirements and cater for that which are slightly different to just taking cloud-like constructs and pushing them to the edge. We have to take into consideration all of these requirements that are slightly different,” said Minhazuddin. “So if I have no network connection, how can I autonomously run without having to take some of these cloud-native workloads they’ve built and try to run them at the edge. They’re having to re-factor platforms to take into consideration network dependencies.”
“Scalability, and obviously orchestration and manageability, always come into play. However, the primary reference computing is streaming data, like there’s constant data generated from something,” added Ouissal. “But it’s generating a tremendous amount of data, and you have to process that they can upload it to the cloud, so bandwidth is the one main problem, so instead you want to process the data elsewhere, for example, using AI at the edge.”
Renaud turned to questions from the attendees, with the most-requested question one asking the panelists to discuss Edge AI and machine learning use cases today.
“We see retail quite a bit. It’s just like there’s a lot of not just the big-box retailers, but even 7/11 gas stations, things like that that have a lot of legacy cameras and cameras generate a lot of data,” said Shakib. “We also see a lot of activities around sustainability, water flow management, smart cities, where edge computing is important.”
“The other vertical I’d like to highlight is oil and gas. It’s a vertical that’s going through significant change,” said Ouissal. “It’s always been a very data-driven vertical. The amount of processing, computing, and data analytics that happens there is quite unique. With the move to renewable, they’re also changing a lot of their operations, and that’s the other vertical energy market that has always been very early on adopting new technology regards to cloud or edge. I think renewable energy use is extremely driven by edge computing. We have deployments in wind farms, really helping to bring down the cost per watt and ultimately offering a path up to a carbon-free world.”
After the Q&A, Renaud asked the panelists for any closing thoughts they might have, to which Basil replied, “Three pillars. Cloud focus operating system, easy to use Kubernetes option in the form of K3S…, and management of scale,” said Basil “Those are your big three pillars, but I would add another one that’s feature facing to make sure your architecture can run in a fully disconnected environment.” To view the full event of demand, click here.