From micro-climate to big cloud
Last year TechPro magazine tracked the discussion of edge computing, revealing a guarded optimism familiar to any enthusiast of emerging technology. Has anything changed in the past year to justify the hype, or are we still playing a game of Buzzword Bingo?
According to the 2018 Gartner list of top 10 strategic digital trends, edge computing is rubbing shoulders with AR/VR, blockchain, AI, conversational platforms, and continuous adaptive risk and trust – a mix of emerging technologies from those embedded in the zeitgeist to those that would power it but remain invisible. Research firm IDC expects spend on connected devices expected to reach $1.3 trillion by 2020. Sound Promising? To be sure but there is a problem of definition as to what we should be getting excited about. According to the IEEE Standards Association edge (aka mesh) “places applications, data and processing at the logical extremes of a network… reducing the volume and distance that data must be moved”. A much more detailed vendor-driven definition to be found in the glossary of the Linux Foundation runs to nearly 100 words adding cost reduction, the deployment of additional resources along the “path” from the edge to centralised clouds.
What we can be certain of is that edge has been with us a lot longer than the term itself, manifested in the form of hybrid solutions dividing up which applications should be kept on-premise and which can live off-premise, which devices can be a source of actionable data and the kind of technology required to provide secure, low-latency, high-bandwidth infrastructure to facilitate remote interconnection.
Outcomes
Catherine Doyle, regional sales director, enterprise, Dell EMC, says edge computing comes into its own when looked at from the perspective of selling a solution over hardware alone. “What we do is look at what the application is. In a lot of situations there would be some part to play for Pivotal and specifically Pivotal Cloud Foundry. We would tend to look at that as a starting point, look at the application and see what we can do to help the customer and create an outcome based on that. We wouldn’t go straight to the hardware – it’s the result of the engagement. We would start with the app, the analytics and ‘what are you trying to do’.”
Doyle sees 5G as an essential technology for taking edge to the mainstream, particularly in automation and self-driving cars where sensors demand incredibly low latency together and react on data in real time. This inevitably leads to a situation where estates will rely on a hybrid model. “I think most companies will use a multi-cloud structure to service their business,” she says. “Some applications are perfect for the cloud and some are not. They will also take software as a service and either keep their data on-site or onto a public cloud depending on its sensitivity. Multicloud is a way of providing a service to the edge rather than edge being an enabler of it.”
Assembling the pieces
For data centre operators the solutions-based approach opens up a range of possibilities. According to the second Equinix Interconnection Index, by 2021 some 48% of global interconnection bandwidth will be accounted for by connections between people, things, locations, cloud and data. That translates into more than 8,200Tb/s of interconnection bandwidth and 33 zettabytes exchanged – about 10 times more than the Internet, which will account for 26% of interconnection bandwidth. The report looks towards interconnectivity as the key to organisations creating their own ecosystems.
Equinix attributes these trends to a number of factors such as businesses scaling up, increased urbanisation and better standards compliance (particularly in EMEA thanks to GDPR). New applications are also springing up in energy, healthcare, retail, logistics or anywhere you will find a touch point logging data.
Kevin O’Connor, Equinix channel manager, Ireland, believes companies should be looking towards developing ecosystems as a long-term strategy. He cites Platform Equinix as a solution for designing, implementing and managing solutions.
“Rather than a company deciding it must be cloud-first or think ‘I’m reticent about the cloud, I want to keep it on-premise’, it’s connecting services that are suitable to use in the cloud with services that are on-premise,” he says. “As an example you may have customers that have their heritage, IT systems or even newer key IT systems in a data centre, but then through the service we would offer, like Cloud Connection Solutions, Cloud Exchange Fabric, you could connect them. That’s not just for the hyperscalers, which we offer, but also through a variety of SaaS providers like Workday, SAP HANA, Salesforce etc.”
O’Connor believes organisations are looking at a multi-cloud approach on an application-by-application basis. “I think it’s a more grown-up approach in terms of ‘well I have certain services on Azure like Office 365 or I have production systems on other systems’.”
The biggest threat to edge computing comes, once again, from a problem of definition. If edge makes sense as a solution to getting the most out of data faster, you might like to consider fog computing, a more secure alternative where processing occurs not on the device but within the local area network using the same equipment only with a greater emphasis on security at the expense of latency. That’s an argument for another day – or maybe another year.
Subscribers 0
Fans 0
Followers 0
Followers