Telco’s relationship with cloud? It’s complicated.
Bridging the massive gap between promise and reality has proved a constantly moving target, all while operators are moving rapidly themselves. But advanced validation approaches could finally be the key to unlocking cloud’s full potential in telco networks.
Today’s telco networks are typically a mix of 3G, 4G and emerging 5G tech. In my post on cloud migration strategies, I highlighted the most viable paths forward on transferring services, applications and functions to the cloud. Telco’s earliest efforts in cloud ultimately took paths of least resistance (yet with limited success), typically resulting in 4G hardware and software being deployed in virtual wrappers. Yes, operators had technically made it to the cloud, but the anticipated benefits were nowhere to be seen. There was no agility or interoperability, which locked purchases of virtual stacks to a single vendor. Support requirements resembled that of a traditional appliance. OpEx wasn’t reduced, and increased in some situations, as more layers and management demands were introduced.
There was an upside though: operators had an opportunity to experience cloud in their own networks, on their own terms. Vendors gained critical cloud skills and operating expertise while building the teams and knowledgebase powering today’s 5G efforts. These are all important developments as telecom seeks to effectively migrate functions to private, public and/or hybrid clouds, and edge data centers. It’s also a critical step toward making all applications cloud native once and for all. That objective, however, means different approaches for different players in this game.
The ultimate goal is for older legacy and new 5G technologies (which are designed from the get-go for the cloud) to perform as a unified environment. Legacy technologies are in different states of readiness for the unified cloud environment. 4G, for example, needs vendor’s apps to be cloud native or cloud ready. Yet, depending on the vendor, they may or may not be ready to run on the cloud. This entails a “lift and shift” approach for migrating apps to the cloud, moving an application and its associated data to a cloud platform—without redesigning the app.
Along with migrating elements of a monolithic archaic environment, such as those found in 4G, some telcos face the challenge of having as many as five different vendors with five cloud environments. If an operator wants to go into the cloud world and have a virtual environment, they must contend with vendors selling them a full and separate vertical stack, where the ultimate objective is a unified cloud stack incorporating all vendors.
If an operator wants to go into the cloud world and have a virtual environment, they must contend with vendors selling them a full and separate vertical stack, where the ultimate objective is a unified cloud stack incorporating all vendors.
For all the lessons learned from the early adopters of cloud, though, challenges remain a mainstay of 5G cloud rollouts. With hyperscalers continuing to make inroads into telco networks and 5G-driven revenues hanging in the balance, sorting these issues quickly is now a top priority. Operators are grappling with a range of realities threatening the prospects of success:
A lack of fully-defined standards and true interoperability mean truly functional vertical stacks, or carefully-defined deployment environments, remain outside the norm. As a result, with rare exceptions, the choice and flexibility of cloud remains out of reach.
Even for solutions from across the vendor ecosystem that are technically compatible from an input/output perspective, the performance quality of a particular stack is not guaranteed.
Unlike more consumer-focused 4G requirements, the considerable range of applications operators must support for 5G’s divergence of paths on cloud strategy means continuous complexity is the new common denominator for any approach.
Further compounding performance challenges is the potentially negative impact of resource-intensive functions all competing in the same environment (e.g., “noisy neighbors”) that can create unforeseen load crises when storms or major events (some planned, others unexpected) overwhelm networks.
The good news is that none of these challenges are insurmountable. The dream of agility to move fast, making updates in a matter of weeks, cutting costs with optimized resources and automation to deliver high performance, with fewer resources, is within reach. It will, however, require comprehensive and continuous testing and validation efforts driven by holistic CI/CD and DevOps best practices. These capabilities are the kind that fall considerably outside the scope of previous testing demands that operators and vendors could once address handily in-house with the right equipment and test teams.
Today, common environments must be simulated and tested under highly-realistic conditions. Consider the curveballs coming down the pike, like edge computing, Open RAN and network slicing, for a sense of how expansive testing needs will need to become. It’s a simple fact that operators without a proactive, low-touch strategy for staying on top of testing for the hundreds, and eventually thousands, of configuration possibilities, will never truly realize the full potential of cloud.
What does this look like in practice? It’s a progressive approach that starts with standalone performance-based I/O testing for cloud native software, incrementally adding more functions and situations until it emulates a full real-world live environment, putting a full load on that system and then testing again. It means emulating every network function, from every vendor update as it is released, to be able to conduct realistic timely and useful validations.
This is the new bar for 5G cloud success. Anything less means settling for a shadow of cloud’s true potential. To learn more, read the Assuring The Promise of Hybrid Cloud Services eBook.