Recently, at GSMA’s Mobile World Live Unwrapped, I spoke at a session focused on edge-computing. (You can watch the replay here with free registration.)
The theme of the discussion focused on whether MNOs care about edge cloud against a backdrop of slow rollouts and reports that edge computing is low on priority lists.
From Rakuten Mobile’s perspective, mobile operators are today’s largest users of edge computing in the world. Traditionally edge clouds are hidden, bundled inside radio processing appliances. Rakuten Symphony has built an edge cloud for Rakuten Mobile where the network itself is the first tenant, with the opportunity for more tenants to follow.
As end-to-end cloud and microservice architecture is introduced into 5G standards, radio software becomes another application to deploy rather than anything special and standalone. The opportunity is for mobile operators to embrace this reality and care deeply about engineering a highly-performant automated edge. They can then be their first customer. It represents the future architecture of radio software delivery, the same way cloud has become the established architecture for delivery for all IT applications.
Let’s talk about why that is and why it matters, starting with the basics.
It’s easiest to define edge computing by first understanding how it differs from cloud computing.
When something runs in the cloud, you don’t really care where it runs. When something runs at the edge, you do. It is 100% location specific and often, the closer to the end user, it is potentially better. Edge computing deployments might be defined by round trip latency, proximity to high bandwidth data sources and data privacy and security policies. The easiest way to think about it: if location matters, it is edge. If location does not matter, it is cloud.
The real definition of edge is application specific, where does it need to sit in order to perform the best?
"Nobody in any company wakes up in the morning thinking, 'thank goodness for 5G and telecom, I need to buy lots of that TODAY.' They do wake up thinking, 'I have X number of massive problems and opportunities and need to know how to best address them.'”
Telecom is doing what telecom does.
It has invested a huge amount of money in 5G and has yet to see a return on investment. So of course, the logical answer to these revenue woes is enterprise, especially if it can be 5G-driven.
Telecom’s goal is to fill the 5G consumer revenue gap by extracting money from enterprises. To put it bluntly, this strategy and perspective is wrong and will fail.
Nobody in any company wakes up in the morning thinking, “thank goodness for 5G and telecom, I need to buy lots of that TODAY.” They do wake up thinking, “I have X number of massive problems and opportunities and need to know how to best address them.”
Telecom and 5G are part of a complete breakfast, but definitely not the whole breakfast. As an industry, we need to make it our business to take part in solving outcomes and being relevant where it makes sense.
"Stop testing [edge computing operations] in the lab. That is where good ideas go to die. Get it deployed, learn for real, stop avoiding the future and it will not avoid you."
During the panel discussion, Muneyb Minhazuddin from Intel rightly pointed out, “operators need to earn trust from enterprises.”
There have been many enterprise forays from the telecom community in the past, in the space of cloud, that no longer exist. If we see them as an ATM, we lose. Rather, we need to view them as a market we can create value within and get the rewards we deserve.
“Rather than saying 5G is the answer, ask how 5G can be an ingredient in a recipe to make these companies work better?”
With respect to skill set challenges, as Jillian Kaplan from Dell pointed out, “we have uplifted skill sets in the industry before and we can do it again.”
As the center of gravity in telecom moves from hardware to software, there is an absolute shift in understanding and capability requirements. The industry is making this shift, and my personal reflection is that it needs to happen even faster.
This means we need to become our own first customer of the next-gen network advancements we tout.
To be clear, the first customer for edge computing needs to be mobile operators ourselves and the first killer use case needs to be the network. We have to run our own mission critical and performance sensitive workloads like radio before we can prove to others it is capable of also addressing their network needs.
Once we do this, two things happen:
We know this because we live it. Today, Rakuten Mobile operates the world’s largest Open RAN deployment and mobile operator edge deployment, with approximately 1,000 edge servers in Japan. Both are foundations for enabling future experiences.
Ask yourself this very hard question: Why would an enterprise trust an operator’s edge computing platform if the operators themselves did not trust it for their own business?
In closing, all participants referred back to the opening presentation and the GSMA statistic that 7% of mobile operators have live commercial edge computing operations and 51% are testing it in the lab. Stop testing it in the lab. That is where good ideas go to die. Get it deployed, learn for real, stop avoiding the future and it will not avoid you.
And full disclosure, the webinar is a better and more balanced watch. Check it out now!