IBM discusses how cloud and telecommunications, along with AI, are converging
Early 5G deployments are limited in coverage scope and capacity but, as operators fine tune their spectrum allocation, scale service offerings and the technology continues to mature, expect much more than simply an enhanced mobile broadband experience. Indeed, the long view on 5G envisions low latency and ultra reliable applications delivered to enterprises and industries of all sorts via bespoke network slices. But to get from where we are today to where we want to be, operators and ecosystem partners will need to continue investments in virtualizing network functions and distributing cloud compute infrastructure out to the edge of the network.
IBM’s Steve Canepa, managing director of the company’s Global Telecommunications, Media and Entertainment Industry business, said IBM followed trends within the telco sector–namely convergence of content creation/ownership and distribution, to better serve the end-to-end value chain in a unified go-to-market organization. As that paradigm is emboldened by the advent of 5G, “What we’ve been working on is kind of a movement toward a hybrid, multi-cloud world,” he told RCR Wireless News in a recent interview. “These workloads that are running in different places have to be stitched together to create value.”
What this means is putting the right network and compute resources in the right places. As data flows through a 5G network, telcos, enterprises and industries are essentially generating and analyzing data to make better decisions. To make low-latency applications–automated industrial control, for instance, valuable, data needs to be processed physically close to a machine, which speaks to edge computing and artificial intelligence. IBM and Vodafone recently entered into a partnership geared toward enabling enterprises to take advantage of these technologies. The goal is cloud interconnectivity in a simplified, secure manner.
Canepa explained that telco environments, largely based on proprietary equipment, are now being abstracted by using an open approach to virtualize network functions in a “common, open cloud platform…all underscored by automation and security. In the enterprise world, when we combine these kind of characteristics, we can see a platform developing that sits inside that telco environment that marries network services with cloud services.”
This concept of openness is exemplified by operators embracing open-source orchestration platforms like ONAP, as well as in the radio access network with the O-RAN Alliance drawing together ecosystem stakeholders to develop interoperability standards for the radio network. This seachange in the telco world mirrors an evolution in the IT world, which is fitting as 5G is marked by a much for IT-centric approach to network architecture than previous generations of cellular.
“A lot of the old IT stacks used to be very proprietary and locked together and you saw those stacks open up over the years,” Canepa said. “And you saw value shift. I would envision a very similar evolution happening. As that network kind of opens and network functions get freed from the underlying appliance, they become services that can be deployed wherever they should be deployed. Once you have the ability to do that…it becomes much more of an automation and integration challenge. The way we do that is we’ve been spending years and years and years putting AI under the cover.”
And in the 5G world, getting away from proprietary networks in favor of open networks allows operators to drive the pace of innovation rather than wait around for their vendors to drive change. “There’s a trade-off,” he said. While there are clear advantages to a completely integrated, pre-tested stack from a single supplier, “Then again, that locks you into the rate and pace of the development curve of that supplier. That can end up becoming a big inhibitor. This unbundling really becomes a market-driven phenomenon.”