As a company, three years ago, we decided to head toward the public cloud — and do so in a very public and aggressive way. It was seen as an interesting move, perhaps too far ahead of its time, but an idea worth keeping an eye on.
Fast forward three years, and the idea of moving compute or storage to the public cloud is far more socialized within the telecom ecosystem. The reality, however, is that not many operators have put their data into the public cloud so far.
Yet, curiosity about how to do it and what to watch out for is sky high. And it should be. What we call “cloud economics” is a world away from telcos’ traditional cost bases. Beyond that, “cloud speed” can be a game-changer, but that needs another article of its own.
Thus, in this article, we share our hard-won insights that have come from three years of putting critical BSS solutions in the public cloud.
Do not underestimate the challenges of “telco-grade” in the cloud
Our industry has been engineering-led for decades. Through necessity, we have developed a commitment to extreme availability and uptime unlike any other industry. And we have developed skill sets that are unique to the requirements of our industry. “Telco-grade” means something; it is a commitment and a promise.
Here comes the “but.” Do not expect that to come easily from the public cloud. Few other industries have the need for our industry’s commitment to extreme availability. The closest is financial services, but even they do not have the blanket need for millisecond speeds that we do. So, unfortunately, as telco is a relatively new priority for the public cloud, there has not been a long-standing commitment to (or the tens of thousands of engineering hours to deliver) our standard levels of availability.
That is not to say that it is not possible. On the contrary, it is possible to get to telco-grade in the cloud through an additional software and services layer. However, operators overestimate what the public cloud can deliver “out of the box” and will need guidance on how to get there.
Get practical about data sovereignty
There is no escaping legislation that has cropped up worldwide regarding data sovereignty, meaning where data is physically stored, in which data center, in which country. For example, legislation exists in many countries necessitating that data of its citizens needs to be stored within the confines of its national borders.
Beyond data privacy concerns, this has taken on a new significance in the light of recent events in Eastern Europe. If anything, we can expect controls to tighten as legislators grapple with the implications for their critical national infrastructure and levels of cybersecurity readiness.
Exploring your local options is not even the first step — it needs to be a preliminary step to see who your viable partners might be. For example, there is little value in choosing to work with a specific public cloud provider only to realize later that they do not have a local data center. Stringent data sovereignty laws will bar you from working with them.
End your addiction to endless customizations
Our industry is an industry of tweakers. “I like this, but could you just change …” Whole sections of your software vendor ecosystem have made the bulk of their revenues from “services” that we could impolitely characterize as endless, expensive tinkering.
Entering the public cloud needs a mindset shift, certainly when working with software partners. You will need to let those software companies be software companies, not services companies. If you choose to go with a software company that is all-in on the cloud, they will already be using DevOps and CI/CD methodologies to centrally manage and consistently improve their software, as it is.
“Lifting and shifting” software to the cloud without a true cloud-native re-architecting is ineffective. And likewise, lifting and shifting the addiction to customizations will not be effective (or cost-effective) either. This is a vital change in perspective and culture that you need to be ready for the cloud.
The add-ons are very attractive and worth exploring
The public cloud business model is based on compute and storage. That is how they make their money. However, to keep you there and happy, the major public cloud providers deliver a host of software “goodies” for little or no cost. It is like a trip to Ikea — the necessity may be new furniture, but I suspect many people go for the Swedish meatballs in the Ikea restaurant!
In the public cloud, for example, you may come across a set of AI or analytics solutions that could propel you years ahead of your competitors. And you can tap into a marketplace of add-on services that could deliver very specific solutions to problems at speeds that the telco industry is definitely not used to. The hype is real. The only caveat is that you need cloud engineering skills to make the most of what is on offer.
Public cloud players want to do business with you … now
Three years ago, public cloud vendors did not really know how to tackle telecom and were not even sure it was a strategic priority. They had been concentrating on vertical markets like financial services, healthcare and government.
That has changed dramatically. Now, telecom is seen as a top priority among the leading public cloud providers, generally within their top five target markets. New telecom divisions have been set up, new sales teams put in place, and they are ready to do business. Those divisions are in their nascent stages, and their case studies may be a little on the light side right now, but there is a willingness to play ball, maybe do some good deals and get you as one of their new telecom customers.
There is no doubt that the public cloud can deliver a host of benefits. Many of those benefits can be achieved through a cloud-native mindset shift achieved through a private cloud, while other benefits are unique to the public cloud. The net result is that however far you go in your cloud journey, the advantages are real. But please take on board these lessons, as the journey is not always straightforward.
