We're surrounded by disruption. Telecom, data centers, networking in general -- everything in the Light Reading universe seems on the verge of major changes that are still taking shape.
It's our job to guide readers through the maze. And we think we've found one word to encapsulate a lot of what's going on: automation.
We've even made "Automation" a category on Light Reading. It's got its own news page just like 5G does. You'll find it here.
Automation is the next step after virtualization. Virtualization was about turning functions into software. Those functions can then be housed almost anywhere on plain old servers. Carriers' data centers become one vast bank of processors where network functions blink into and out of existence all the time.
But this won't work if we have to wait for human intervention. The network must be automated.
This isn't a new concept. But in our stories, as we dig into the minutiae of NFV and survey the hundreds of services available on public clouds, we sometimes lose sight of the endgame. It's time to recognize the big picture and call out automation as the driving force behind many technology initiatives. That includes:
Software-defined networking (SDN). In a sense, automation has long been the thesis here. SDN started out as a way to apply different types of routing to a small network -- specifically the Stanford University network, if you want to go way back to the beginning. By the time SDN went commercial, we were talking about a network intelligently reconfiguring its switches to respond to traffic conditions. What's more "automation" than that?
Network functions virtualization (NFV). This is the big one. NFV has become a story about automation -- about using cloud-like technologies to bring network functions into being on an as-needed basis. It's also about moving those functions as necessary, due to fiber cuts, traffic congestion or plain old economics.
Artificial intelligence and machine learning. This one's fuzzier, because AI can be used to "automate" anything. The AI stories in our automation bucket will gravitate toward network-centric tasks such as provisioning, configuration and the feedback loop of telemetry. Not as glamorous as helping spinal patients walk, but still potentially world-changing for telcos.
Big data. Speaking of telemetry. Automation doesn't work unless a carrier builds that now-famous feedback loop, tapping the network to find out what's working and what isn't. I've spoken recently to CenturyLink, where they consider big data to be a necessary early step toward automation. That means creating a distributed data lake so that a global orchestrator can grab information about any part of the network at any time. Big data is the legwork behind orchestration, so we intend to keep an eye on how carriers use this technology.
That's the foundation. I expect our automation focus to widen over time to include pieces of security and IoT. Eventually, automation will become so ubiquitous that the "automation" keyword will be redundant.
Of course, we'll continue covering the other major topics around telecom and networking, such as 5G, smart cities and the impact of the public cloud. We'll continue to draw from sister sites Enterprise Cloud News and Security Now to keep abreast of major advancements in communications. And as we cover this "automation" stuff, we won't lose sight of the potential impact on jobs.
These are all pieces of one very complex puzzle. I've described how we see it through the Light Reading lens. I'd love to hear what other points of view might be out there.
Cool writeup.. I believe that the telecom / network management line of business has the scale, the scope, and the staff knowledge to lead the way with real time machine learning and fast analytics from big data. Many areas like healthcare have the need for batch machine learning but network management will be driving the need for real time. keep up the great articles.
A long time coming Yes, automation is the key. In 1990-1991, I worked with some talented architects at Bell Labs on the "7 Selfs" - self-aware, self-inventory, self-healing, etc. Our goal was a network that required human intervention only if something had to be physically changed. And we did it, with prototype control on real 5ESS, SLC-5, etc. and the hooks for automated OSS/BSS, routing, and more. (I personally demo'ed it to Arno at Chester.) I look at NFV/SDN pictures now and they look a lot like our diagrams from then. That tells me it's the coming thing.
But 27 years and still "coming"? That's a long time!
Re: The name of the game: Automation It's a good and fair point. A lot of the industry discussion of Automation presupposes that the input is already the "right answer" - it's just a question of execution. In fact operators need to bridge the huge gap between the goal of getting a profitable outcome (as you say) and changing the network. Virtualization is a new tool for the latter, but does not necessarily determine the former. Automation can industrialize the delivery of unprofitable services just as effectively as profitable ones!
Not only... but also... Great to see LightReading calling this out, and looking for industry progress on the goal of automating more. I'm hoping this won't be limited only to the automation of SDN and NFV-centric activities. As I think the industry is rapidly appreciating, SDN and NFV are themselves only partial enablers of the greater automation journey. To that end - I'd take propose a change to the headline of this piece: NFV, SDN and AI (rather than Big Data) are the trio of technologies that together are the requisite mix for business-transforming Automation in telecom.
The name of the game: Automation Yes, Automation is the "name of the game". However, the end game is and should be: new services, speed, agility, time-to-revenue, and other business related, in the end the main purpose is to have a profitable business. Automation, like Virtualization, are ways to achive those objectives. There has been a misconfusion on very important concepts, regarding the purpose of doing something and the ways to achieve that purpose.
Re: It's also about scale It's worth noting with respect to 5G that the 3GPP now has a dedicated "Network Operations" development/specification track that is, in effect, about automation. Also in mobile, the NGMN Initiative has long campaigned (I think that's the right word) for better operations and automation. This not linked only to NFV/Cloud but to the entire network.
It's also about scale A major goal of NFV/SDN and almost every other aspect of the current transformation process is to enable network operators to deliver massive amounts of bandwidth at relatively low cost - something current architectures won't enable. Automation is critical to achieving that goal and always has been - thus companies such as AT&T, Orange and Vodafone have been talking about it for years.
It's great to zero in on that focus because without achieving a level of automation, network operators will not achieve that fundamental goal.
Re: The term 'de jour' Automation truly is the end game. We've seen it in the enterprise and now service providers are adopting open source, virtualization and standards, we are increasingly see the drive toward automation in this industry. In fact, AT&T's Chris Rice stressed that point in a recent interview with UBB2020 - AT&T'S Rice: Open Source Paves the Path to Automation
The term 'de jour' Automation isn't new but it's now emerged as the 'term' to encapsulate the next big step, though it's clear there are also other terms being used for manifestations of automation eg the self-driving network.
And now we can point out, over the next few years, all the things that need to be put in place before NFV and network/service/process automation can become a reality and actually impact top and bottom lines of operator financials.