Looking to the future of any industry through the lens of its foundational technology is very exciting. Yet getting from the idealistic vision to a workable reality requires a major leap – an operational management layer.
Retail is an example. How will edge computing revolutionise the retail industry? Examples range from the theoretical to the tangible. Brick and mortar retail stores contain a lot of data such as inventory, supply chain, transaction, and customer data. Decades of computer science taught us that compute should reside as close as possible to the data to ensure better execution at lower latency and lower cost.
What can this look like in practice? Accenture offers examples: real-time personalised recommendations and offers to customers’ devices as they walk around the store; shop staff offering information and potentially upselling; and ‘virtual try-ons’ of clothing and apparel through augmented reality.
While such use cases sound attractive, implementation and operations can get very complex. How can you make sure latency and reliability issues do not occur? Is the investment worth it if you are having to hire expensive technicians if something goes wrong? How can you manage thousands of sites reliably and efficiently? How can you deploy applications to each site remotely and automatically?
This is where a company such as Platform9 comes in. The company’s edge offerings – including those specific to retail, telco, manufacturing, industrial tech and others – aim to answer these questions through zero touch operations. Whether your organisation’s workloads are cloud-native, container- or VM-based, Platform9 will get you from bare metal to a managed platform remotely and dynamically.
The company’s vision, dating back to its founding in 2013, remains consistent. It is built on three pillars of an open, distributed cloud architecture; open source service endpoints, ‘fully managed’ by a SaaS management plane, with any choice of public, private, or edge infrastructure.
Ron Haberman, market CTO at Platform9, explains how the vision is “reinforced” by edge through what the company calls always-on assurance. “The vision was to have a centralised management plane with an always-on assurance that helps your operations run smoother, and take the hard problems and solve them while you sleep,” he notes. “And this vision gets stronger with edge because as systems get more distributed, the more monitoring, and support you will need and managed solutions become important.”
Haberman’s role is primarily working with customers to build solutions. Previously, he was CTO at Nokia Software, where in part his team built a cloud platform for telco. Given his experience with cloud as that technology progressed from concept to commoditisation, the natural question was: where are we at comparatively with edge today?
“It’s somewhere beyond the halfway point, I would say,” offers Haberman, but with a caveat. “I will say that edge hasn’t really been defined as a market. In my previous career, we used to talk about the edge, the far edge, and the customer edge, just to name three. Overall, there are more edges than that.
“Retailers are now running highly secure in-store applications including point of sale, AI/ML inference workloads, video surveillance, order fulfilment, and ability to take online and mobile orders. All of those applications can now run on a single platform, which means that developers can push software much more frequently into all of their store locations. A lot of businesses realise this is the way it’s going to go.”Ron Haberman, market CTO, Platform9
“That said, I think the purpose of why you would choose one type of edge over another has been defined quite well,” Haberman adds. “If you have applications that require a lot of upstream bandwidth, such as video; if you have data that is sensitive; if you have any kind of sovereignty issues around this data, you might need to be in a particular edge – an edge where compute and data are close to each and in a secure location.”
Much as the cycle from cloud to edge marks a change from centralised to decentralised computing, the thinking around implementation moves along the cycle too, as Haberman explains.
“It was always easy to start with a vertical solution,” he says. “You bring something that is fully integrated, and you drop it wherever you need it, and it works. Then you add the second one. By the time you add the third one, you start thinking ‘maybe I don’t need three times of everything. Maybe I should have some more generic compute here that can take use cases as they come.’
“When I say [edge] is beyond the halfway point, especially with retail, this is exactly what I mean,” Haberman adds. “Retailers are now running highly secure in-store applications including point of sale, AI/ML inference workloads, video surveillance, order fulfilment, and ability to take online and mobile orders. And all of these applications can now run on a single platform, which means that developers can push software much more frequently into all of their store locations.
“That’s already happening, and a lot of businesses realise this is the way it’s going to go.”
From an executive’s perspective, the obvious selling point is in these ‘business-impacting outcomes’, from customer happiness and user experience, to faster and more effective execution. Yet cost must not be overlooked.
“We’re in the new economy, and saving money on compute and how efficiently you deliver software, is critical,” says Haberman. “So we see a bit of a combination. Companies now realise that they need to optimize costs of their old VM-type environments.
“They want to pay less both for the licence costs, and, more critically, for the ongoing operating costs of the environment,” adds Haberman. “And they’re looking for help from a company like us to solve operational challenges for them.”
So what does the future hold? The killer applications for edge will emerge with availability of edge cloud platforms that can easily be managed and run centrally, Haberman notes, “Some edge applications that require large amount of data processing and low latency are already here, including streaming video, big data analytics, AI/ML, and real-time store operations,” says Haberman.
“Having such platforms now empowers developers in various industries to rapidly implement and deploy new use cases, opening up new capabilities, and transform end customer experiences.”
Photo by Nita
Want to learn more about edge computing from industry leaders? Check out Edge Computing Expo taking place in Amsterdam, California and London.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.