IBM was founded in 1911 to help companies automate basic business transactions (think time clocks and punch cards). Now, 112 years later, automation is back at the heart of its story in a big way, but this time it’s focused on network automation.
The executive in charge of this mission is Andrew Coward, IBM’s general manager of software networking.
“I’ve been brought in to bring networking back to IBM,” Coward told me matter-of-factly during a recent conversation. The reality is that this is a huge shift for the company, which hasn’t been a networking vendor since Lou Gerstner brought the company back from death’s door by refocusing it on software and systems integration almost 30 years ago.
IBM has given Coward a team of around 500 and a war chest the size of a container lorry to pursue his strategy. In other words, it’s not goofing around. And neither is Coward.
Since April 2020, IBM has quietly acquired more than 30 companies, bolstering its hybrid cloud, automation and networking capabilities.
Control freak
IBM isn’t just throwing ideas at the wall to see what sticks. Underlying all its acquisitions is the goal of making cloud infrastructure a better fit for an enterprise’s business-critical traffic.
“There’s a Copernicus moment where you suddenly realize that the [public cloud] network today is not fit-for-purpose,” said Coward.
The IBM strategy is to allow telcos and enterprises to build their own automated network infrastructure on top of the public cloud infrastructure offered by from Amazon Web Services (AWS), Azure and others, to provide the functions that businesses need to entrust their mission-critical applications to the cloud.
“We overlay the infrastructure required to provide control,” said Coward.
“With some cloud providers you have no idea where your traffic is going. It could be traveling through any foreign country. Obviously, that’s not acceptable to a lot of businesses, including banks. If a regulator found out they were doing that, there would be serious repercussions,” he added.
IBM’s new network automation strategy is a prime example of the industry trend toward telco-ification of public cloud, where the hyperscalers’ “just get it done” approach to traffic delivery is upgraded using features like network slicing, load-balancing and SLA enforcement to enable telco-like levels of reliability, visibility and predictability.
It’s looking increasingly likely that this is just a precursor to a time at some point in the misty future when those categorizations of enterprise and telco become moot – not simply co-mingling but being served by a single ubiquitous cloud infrastructure.
Competitive landscape
IBM has been on an extended shopping trip, purchasing 30 companies to build out a portfolio in support of its new mission because “the problem is bigger than a single product,” Coward admitted.
IBM’s deals include Volta in 2021 (for distributed routing); SevOne in the same year 2021 (for application-layer visibility); and, this year, NS1, an intelligent DNS & traffic management vendor that enables highly sophisticated management and automation of applications and services.
To be sure, multi-cloud capability is a central tenet of IBM’s strategy and Coward’s philosophy.
“We are here to support whatever combination of cloud that the customer chooses to deploy,” he said. “All the acquisitions we are making are building towards supporting the multicloud capability of customers.”
From this writer’s perspective, it looks like IBM could use some more multi-cloud capability, but Coward said he’s not in any hurry to buy one of the two dozen or so multi-cloud start-ups that are vying for an exit currently.
“Start-ups do the technology stuff, but they can’t help with things like cloud migration,” he said. “And spending a billion dollars to buy a start-up with $10 million in revenue is a bad look for us.”
That’s not to say that IBM isn’t now heading into a space where it will compete with a slew of startup companies trying to make sense of the current multi-cloud miasma, including Aviatrix, Alkira, Cato and Prosimo.
And with the rapacious rate at which IBM is buying companies, nothing is entirely off the table.
History’s Lessons
IBM is not the first equipment company to take the build-through-acquisition route — and Coward says he is acutely aware of historical precedent for the company’s current strategy.
“Lucent tried to do this and failed,” he said.
He plans to avoid the same outcome through product lead growth (PLG); tracking which apps and services are popular with companies and prioritizing business development around them — this explains the company’s acquisition of SevOne, which provides this capability.
It doesn’t hurt IBM’s chances of success that its executives aren’t known for stupendous levels of mediocrity or using shareholder monies to build private golf courses, as Lucent’s were.
The company’s network strategy is part of a broader company re-org that included splitting out the service and support for the networks that it sells into its new managed infrastructure business unit: Kyndryl.
The Software Networking division sells directly to service providers — including Dish and Telefonica — and enterprises, Coward said. It also has reseller agreements with both telcos and the hyperscalers themselves, he added.
“We are seeing lots of enterprise sales via the telcos,” he said.
That is not especially surprising. Telcos have made a proper bollocks of monetizing 5G and have been caught on the hop by the stupendous success of public cloud networking, failing to provide a viable alternative that can provide the best of both telco and open-source worlds.
Now, along comes IBM with a big bag of network automation goodies to make them look competent.
For Coward, his appointment to IBM’s general manager of software networking is the latest leap up a ladder that he started climbing in 1994 when he started his career as an engineer at Xylogics. Tragically, that job was in Milton Keynes, U.K. He now lives in the Bay Area and is proof that sometimes nice guys do finish first.
This article was updated on March 27, 2023, at 2:36pm EST.