The Hidden (Human) Cost of Automation
Automation is the next big thing for the communications industry, but it's set to have an equally profound effect on automobile manufacturers. In the car world, they use the term "autonomous," rather than "automated." But in both industries, the end point will be the same: massive reductions in the human cost of running a business.
That's demonstrably the end goal for a company like Uber, which has shown over and over again that it really doesn't hold human beings in very high regard in the first place. The opportunity to use automation to eliminate all those pesky human drivers, currently devouring 80 cents of each dollar it collects from carrying passengers, explains why Uber has invested heavily and repeatedly in developing autonomous tech with its auto partners.
Telecom players, meanwhile, are getting equally excited about automation (essentially, automation is the end game for virtualizing network infrastructure and services using NFV and SDN).
In a virtualized and automated network, things happen, well, automatically, and carriers are talking up the benefits that this delivers to their customers via making services more reliable (by eliminating human error) and faster (by eliminating the need for their employees to do network and service "stuff").
What they don't tend to talk about is potentially the biggest benefit of all. It's one that benefits the service providers themselves, rather than their customers, in the form of huge savings in operational expenditure (opex) from using automation to eliminate staff currently employed to manually manage and administrate their networks and services.
How much money are we talking about?
AT&T Inc. (NYSE: T) currently employs 246,000 staff. Moving to a software-based, virtualized and automated network could potentially enable it to eliminate 30% of its staff (or 78,000 employees) according to one source. Assuming an average salary of $70,000, not including benefits, that's an annual savings for AT&T of $5,166,000,000 -- or a scoot over $5 billion -- but also, obviously, results in a major human cost to those who could lose their jobs.
Of course, we're in back-of-the-envelope guesstimate territory here, and my math doesn't factor in redundancy payments, or investment in the software infrastructure required to make the downsizing possible or the cost of hiring programmers and other software types required to make the whole thing hum. But whichever way you cut it, we're talking about the opportunity for AT&T and others to save huge amounts of money, while simultaneously delivering dramatic improvements in service quality and reliability and creating an infrastructure capable of delivering all of the whizzy new services and applications consumers want in the 21st century.
So how far along the path to automation are carriers?
In terms of noise, 2017 is turning into the year of automation. Recent announcements by Cisco, with its Intuitive Networks initiative, and SK Telekom's T-MANO announcement, are both symptomatic of this trend. Both are relative newcomers, however. It was AT&T, with ECOMP, that first anticipated the shift to automation -- and the work that John Donovan, Andre Fuetsch and their teams have been conducting in this space, and which started all the way back in 2012 with the Domain 2.0 whitepaper, is today looking extraordinarily prescient.
However, underneath the buzz, all of these companies still face an absolutely huge task getting the tech to work. Automation takes telecom companies into a place well outside of their traditional comfort zone -- one that has much more to do with arcane computer science than communications (a challenge that was explicitly acknowledged by AT&T when it took its ECOMP program and opened it up to the Open Source community).
Indeed, there is more than one way to skin the automation cat (meow). Industry initiatives such as Open Network Automation Platform (ONAP) and ETSI's OSM rely on the approach of getting all of the companies in the comms supply chain to agree to a common set of rules (including APIs) which govern how virtualization products talk to each other. An alternative, being implemented by EnterpriseWeb, is to use a middleware package sitting between virtualization solutions from different vendors to handle the integration. This approach frees third-party software providers to innovate without having to conform to an agreed set of "rules," but potentially locks the carrier into a middleware solution from one company.
However carriers choose to go about automating their networks, I'm happy to point out right now, nice and early, that the job is going to be substantially harder than the announcements currently being made make out. That's why, as Craig Matsumoto pointed out in his column last week, Light Reading will be focusing on the automation space with religious fervor this year.
It's also the focus of a major new collaboration between LR, The New IP Agency, EANTC and many of the world's leading next-gen communications companies to create the industry's first comprehensive hierarchical taxonomy of next-generation communications -- including automation. This six-month project will result in a unique and exhaustive blueprint that will simplify and accelerate the design and construction and monetization of virtualized, automated networks (hit me up with an email at firstname.lastname@example.org if you're interested in helping us with this historic work).
And if you're interested in learning about the role that automation is playing in the parallel universe of the automotive industry, please do check out our new community: The Connected Car.
— Stephen Saunders, Founder and CEO, Light Reading