Communications service providers (CSPs) are embarking on a new journey.

Having trodden a bumpy path to virtualized network functions in recent years, they’re now on the road to cloud native.

The move towards 5G is driving this transformation – and disruption. A largely static telecoms marketplace is now experiencing a level of disruption in technologies and suppliers not seen for decades. The journey will involve embracing new technologies and new ways of working, but once they reach their new destination, CSPs will be able to take advantage of automation, simplification and microservices to develop and launch innovative service offerings more quickly and cost-effectively.

“The telecoms industry is right in the middle of transforming from NFV to cloud native,” says Lakshmi Mandyam, VP of Product Management, Telco and Edge Cloud business unit, at cloud computing and virtualisation software and services specialist VMware.

“Decisions are being made, procurement is active, vendors are hard at work to transition existing monolithic network functions to cloud native applications, and 5G is accelerating all this.” Something that is clearly evidenced with 5G RAN announcements from companies such as DISH.

To succeed, building on the learnings and successes of Network Functions Virtualisation (NFV), a framework for disaggregating services software from underlying hardware, and the adoption of AI/ML will be vital. Infrastructures must evolve to accommodate virtual and cloud native technologies, and the interaction between people and technology must change.

CSPs are making headway with their 5G strategies

…but it’s vital that they consider the changes that need to be implemented across their entire architectures: from core to edge to RAN.

Network operators’ adoption of artificial intelligence and machine learning technologies, and resulting automated processes, will have a direct impact on the business success of 5G, according to VMware: CSPs will be unable to develop an optimum suite of profitable 5G services without leveraging the agility and efficiencies these advanced technologies offer.

Firstly, there’s the RAN itself to consider. The network densification that comes with 5G – it could require a tenfold increase in the number of cellsites compared with previous generations of mobile technology – along with a significantly higher number of control variables, presents a major challenge for operators.

“Deploying, configuring and managing all of those cellsites is going to be an overwhelming task without AI and automation,” says Karina Dahlke, Senior Product Marketing Manager at VMware Telco and Edge Cloud business unit. The fact that non-standalone 5G relies on a 4G core means even more variables. “As more operators look to virtualize both their core network and radio access network, this further adds to the complexity, as the new technologies will need to interact with the legacy ones for some time,” she says. “AI can be leveraged to configure, manage and adapt to these continuously changing networks.”

Cost is also an issue.

“Service providers are continuously asked to do more with less. As the complexity in the network increases and the volume of the data and devices supported increases, so does the operating expense. But the average revenue per user is stagnant – and even decreasing in a lot of cases – and disruptive players come in with low-cost, flat-rate plans and that drives prices down, which further exacerbates your situation,” Dahlke says.

It’s a situation CSPs recognise well. In India, Reliance Jio reshaped the telecoms landscape in a matter of months following its launch in 2016 with a free and then low-cost service based on a legacy-free 4G network, and is now the country’s largest mobile operator with a staggering 388 million customers and counting.

The analyst community agrees that the automation afforded by AI tools will help operators to increase efficiency.

“What [CSPs are] looking to do is utilize AI to lower their overall operating expenses and [enable] better utilization of their capital. So if you can defer capital investments to some point in the future, and utilize AI to do that, especially in the planning of things like radio access network and other infrastructure, it’s a big win for communication service providers,” says Patrick Kelly, founder and principal analyst at Appledore Research.

AI is also key for operators wanting to compete effectively and ensure customer satisfaction through high-value services that come with stringent SLAs “such as predictive quality of service for enterprise applications like IoT-driven factories, or dynamic WAN slicing for low-latency applications such as self-driving cars. It’s nearly impossible to meet these requirements and strict service level agreements without data-driven AI,” VMware’s Dahlke says.

While low-latency service offerings have applications across the enterprise market, there are consumer applications too. Video gaming is a key area, particularly e-sports and cloud gaming, for which an increasing number of people are showing a willingness to pay for higher quality network connectivity.

“You need to understand first what is going on in your network at all times. Then be able to use that information to provision the right quality of service and guarantee that level of service to them – because they [the end customer] will be paying more for it,” Dahlke says. “The mobile gaming market is fast growing and high-quality 5G connectivity will play a significant role in supporting the low latency that today’s video games demand. If customers can’t get the network response time they want, they’ll go somewhere else.”

This is where AI plays a critical role, notes Ben Basler, Senior Director, Product Management Telco & Edge Cloud at VMware, in analysing and extracting intelligence from a broad range of network data points to enable and assure new revenue-generating services.

In these respects, CSPs are starting to mirror the webscale giants that have long benefitted from the agility and opportunities that automated, AI-enabled insights and software-defined operations bring, but, as Appledore’s Kelly notes, data is the fuel that drives these processes.

“The biggest obstacle [to deploying AI] is access to data… [CSPs] need to be realistic about acquiring not only the right data, but also high quality data. And they need it in large volumes, because if you can’t train the data and you can’t get a lot of data in, it’s hard to get the outcomes that you’re looking for. If you stumble here, your business case is just going to fall apart… Data drives the machine learning engine and unless you can overcome that hurdle, you really don’t have a strong business case.”

This is where VMware comes in. Its Uhana AI-based analytics platform and Telco Cloud Operations (TCO) real-time automated assurance solution help CSPs optimise their networks and applications. A customer in the Asia-Pacific region has already deployed the VMware technology in its 4G and 5G networks and learned some surprising things.

The introduction of cloud native methodologies will change the way telcos operate and behave for ever

…as they introduce continuous integration continuous delivery (CI/CD), adapt to agile and elastic processes, and embrace automation.

“The mindset has to change,” says Sumit Verdi, Senior Director Solutions Management, Telco and Edge Cloud, VMware.

Telcos have been structured in a very siloed, vertical way, with specific people responsible for specific domains: transport, access, voice, data.

“What we’re doing now is creating the management of the physical assets, virtual assets and services right across the horizontal plane,” says Verdi.

Automation is key to the shift towards cloud-type horizontal environments, and that requires a “change in the mindset of people,” says Verdi. “They really need to start trusting the machine.”

That, in turn, changes and simplifies the relationship between the operations team and the network/platform.

“What the people [within CSPs] are essentially doing is acting as consumers,” Verdi says. Simplification means building an operational model on top of the technology to make it “human-consumable and human-compatible, because if you start getting into the nuances of the technology itself, your head starts to spin. There are too many complexities and too many new constructs.”

VMware’s approach is to introduce layers of abstraction that essentially hide the technology from those consuming it.

It is a new approach for CSPs. They can rely on an automated, orchestrated workflow that remediates issues proactively: It’s more efficient for the telco and offers a better experience for customers.

“Service providers are embracing this. And they are starting to see how operationally this adds orders of efficiency,” says Verdi. “There is no need to be figuring out how the plumbing is actually going to work, or, if my container goes down, how I resuscitate it.”

Much like the webscale players: “Fast fail” is a concept made popular by the likes of Amazon, but it is a concept that has largely been out of reach for the telco community – until now. A combination of automation, containerised service deployment and a change in mindset will give telcos the tools they need to develop, launch, adapt and cull services quickly and efficiently.

“You’ve now got agile methodologies where you’re working in this ‘continuous’ mindset. Continuously innovating, continuously developing, continuously deploying and so forth,” Verdi says. “In this new world it is so much easier to create these services, deliver to different markets and then, if the application is no longer needed, it’s a lot simpler to decommission and use that capacity for other services.”

VIEW FULL VIDEO: People and Processes for Cloud Native Telcos

But this CI/CD mindset requires new skillsets.

“One difference in skillset is the software part. You want to have more software skills,” says Marcus Brunner, Head of Standardization, Open Source and Eco-System Development at Swisscom. “Working in an agile way, working in a more DevOps way, there are people who are not used to that way of working. That needs a learning curve.”

However, he notes that new hires into the company expect to be able to work in this new way and would not join without it.

Acquiring the right skills involves more than just a new hiring strategy, though: Telcos need to decide whether to do it themselves, engage external managed service providers or systems integrators, or try a combination, says VMware’s Verdi.

It’s a similar predicament with containers, “the latest kid on the block,” Verdi notes. “How do you build, deploy and manage containerized type environments in this cloud native paradigm shift? It is a daunting and time-consuming task.”

But time is of the essence and organizations, including the CSPs, have a long way to go before they acquire the right skills. According to the results of a recent survey conducted by analyst firm Omdia, 32 percent of organisations noted that a lack of cloud native skills is delaying more rapid adoption of microservices-based applications.

“Where appropriate, bringing in managed service providers can help with that stop gap until CSPs’ internal skillsets can be developed,” Verdi says. CSPs also “need to balance how much they do in-house versus leverage the best-of-breed ecosystem that exists outside of their four walls.”

And the importance of developing a strong network of partners has never been greater: According to Omdia, CSPs are missing out on significant chunks of early 5G enterprise service deals in no small part because they lack the partnership ecosystem to be able to offer the broad range of applications and solutions that enterprises need.

There are plenty of willing partners out there, should the telcos choose to engage with them. And some forward-thinking operators are already looking in that direction

“There are more components coming together to build certain services, so there needs to be a certain level of integration,” says Swisscom’s Brunner. “How complex the services get depends largely on how well the APIs are defined or even standardised between these components. At the moment there is still a lot of freeflow on that,” and little in the way of consensus, he says. “But I believe the way the technology is going, the softwareization of the telco infrastructure, will actually help, together with a lot of different partners to build these services much quicker than we used to.”

CSP customers need to deploy with minimum risk and fastest time to revenue… whether that’s the core Telco Cloud infrastructure, internal IT or enterprise-customer services such as SD-WAN or digital workspaces.

Without a clear focus on operational simplification and technology partnerships, deploying virtualisation and cloud-native technologies risks replacing one form of complexity (silos of technology and processes) with another. “We build technology for the mission, not the lab” says Mandyman, emphasising a need to focus on outcomes and day 2 operations.

“We constantly engage with both our customers and our technology partners. We want to make sure we deliver technology that works in the customers’ production environment and addresses real business needs. The result is that we deliver an application-aware digital infrastructure with the agility and flexibility to optimise capacity across IT and network services, while enabling for continuous integration and delivery of more personalised services to end customers.”

The countdown to 5G has begun.