Home5GAI, automation and the rise of cloud-out network architecture (IBM on 2022)

AI, automation and the rise of cloud-out network architecture (IBM on 2022)

In the last two years, we have seen an explosion in the number of cloud providers used by enterprises. What started-out with deployments in one or two clouds has now morphed into five or six, especially when SaaS providers are taken into consideration. We are truly operating in a hybrid, multi-cloud world, and this is having a profound effect on application and network usage.

IBM’s global study on cloud transformation found only three percent of respondents reported using a single private or public cloud in 2021, down from 29 percent in 2019. Applications are now often spread – some would say sprawled – across multiple clouds. For example, a web front end can be in one cloud, main processing hosted on another, with databases to track customers, stocks, and enterprise resource planning (ERP) across a host of different clouds and SaaS solutions as well.

Coward – scaling 5G services in days, not months

In other words, network management has also gotten significantly more complex in only a few years. Get a few network managers together today, and talk will quickly become group therapy as they swap horror stories of latency, poor performance and unplanned down-time pinned on them, even when these outages are the result of a business decision made with no regard for geographic distance or bandwidth.

Without tools to monitor, control or secure multi-cloud networking, network managers are left with a do-it-yourself stitching of network connectivity between the clouds. So, what does this all mean for 2022? Simply put, it means a profound shift in how we think about networking.

AI and automation helps drive 5G adoption

As communications service providers (CSPs) work to help deliver on the promise of 5G, they will look to AI-powered automation to stand-up and manage networks quickly, across a wide range of environments, and scale new services in days, rather than months. New capabilities will help them bring together advanced analytics, machine learning and AIOps to discover hidden patterns and trends in networking data and continually optimize network operations and performance with minimal human intervention.

An emphasis on security 

If networks of the last decade meant building a global physical network focused around bringing traffic to corporate data centers, the next decade is about building secure virtual networks in and between cloud providers and private clouds. This is nothing short of a Copernicus moment for networking, where we realize the corporate data center is no longer the center of our universe. Instead, we will realize we now live in a heliocentric universe which revolves around managing traffic between an array of different cloud providers.

Better tooling

As new products and vendors emerge, expect to see tools that enable network managers to design networks that flow in and between cloud providers, and can be enabled on a global basis within minutes, rather than months. These tools can use AI and automation to help providers reduce network-related unplanned downtime, speed the deployment of network equipment, and lower their overall IT infrastructure costs.

Blurring IT models

One trend we saw in the early 2020s was the blurring of traditional IT models like secure access service edge, or SASE, and software-defined wide-area networks (SD-WAN) that rely on a software-defined network. I expect the blurring of SASE and SD-WAN products to combine into multi-cloud access and security solutions.

The rise of cloud-out architecture

We expect the shift to a “cloud-out architecture” will require adaptions. We have found that network orchestration, fault and performance management work best when they are hosted in the cloud and managed outwards toward the edge – as opposed to housing tools in a corporate data center that reach into the cloud. This can make it easier to manage cloud-to-cloud data exchanges, cloud-to-edge applications, and end-user experience.

This distinction is more than semantic. Up to this point, most network architectures viewed the cloud as something to connect to and not something to connect from. This is a gravitational shift, as the enormous volume of enterprise traffic begins to flow in and between multiple cloud and SaaS providers without even touching the corporate network.

Evidence suggests that network managers are increasingly empowered within their organizations. Back in 2011, only one in five CIOs surveyed considered themselves a critical enabler of a business’s vision. The pandemic obliterated this perception and made network management a critical priority. In fact, 77 percent of the CIOs in a recent survey report that their teams played a vital role in their organizations’ response to the pandemic.

Not only that, but they are also playing a critical role in innovating in areas ranging from automation to sustainability. This new, more highly visible role for network managers can help them make the changes needed to continue adapting and delivering more reliable service to their customers in the years to come.

There is good reason to hope that thanks to these improvements, network managers will be able to shift from commiseration to celebration when they gather in 2022.

smart meters
Previous post
Telefónica Tech to deploy 130,000 smart meters for Canal de Isabel II
digital twins
Next post
Spending on industrial digital twins to reach $33.9bn in 2030: ABI Research