Schneider Electric is betting on a burgeoning market for tools to reliably manage edge resources, at massive scale.
A chain is no stronger than its weakest link. Nor is an IT service, and with edge computing facilities playing an ever increasing role in the delivery of IT services of all kinds, including IoT, there needs to be just as much effort devoted to the reliability and availability of edge resources as to the mega data centres.
Often that is not the case. And that can be a problem, a point made forcibly by senior vice president and CTO of Schneider Electric’s Secure Power division at the company’s recent ‘Life at the Edge’ press event in Singapore.
The Uptime Institute’s Tier Standard is the globally recognised measure of data centre reliability and overall performance. It defines four tiers for data centre performance and specifies tests for conformance with these definitions.
The Institute does not publish uptime as a percentage for any of these tiers, but Brown quoted some figures he said he had acquired from other sources: Tier I: 99.67 percent availability; Tier II 99.74; Tier III 99.98: Tier IV 99.995. These percentages translate, respectively, to 28.8, 22.7, 1.6 hours and 25 minutes of downtime per year.
But there is little consolation in having a Tier III data centre rated at only 1.6 hours of downtime per year if it is but one link in a chain of systems where what matters is the overall availability of services delivered by that chain, and that level drops dramatically with an even slightly less reliable edge computing facility, Brown pointed out.
Combining a Tier III data centre (99.98 percent availability) with an edge computing facility rated at 99.67 percent availability would create a system with end-to-end availability of 99.65 percent, or 30.7 hours downtime per year.
He argued that this end-to-end reliability is what matters to customer, but is very much a new approach.
“The industry has historically only worried about the centralised data centre and only focused on generally the result of this tier three standard,” Brown said.
“It’s not focused on the customer experience. … The industry only cared about the gear insidethis big data centre. That’s misleading. Because if the focus is the customer experience, or the employee experience, you have to consider what this mobile edge sector is like.”
Brown said the industry needed improvement in three key areas to make the edge resilient: an integrated ecosystem; management tools, analytics and AI to aid staff managing edge resources.
Managing the distributed edge
He argued that conventional management tools were not up to the challenges of managing a large number of edge resources, resources that might be distributed across two thousand retail premises.
“We believe there are four key ingredients: a secure scalable, robust cloud architecture, a data lake with massive amounts of normalised data; a pool of subject matter experts with deep knowledge of system behaviour; access to machine learning algorithm expertise.”
Cloud architecture is where Schneider Electric is placing its bets. “We don’t see a wayto solve this without leveraging cloud technologies,” Brown said.
“We’re not going to be able to manage thousands of stores in different geographic areas around the world, unless you leverage a different type of software. We’re pretty excited about the foundation we’ve got for analytics and AI.”
The author attended Schneider Electric’s ‘Life at the Edge’ event as a guest of the company.
