hero:Tech: A convergence at the Edge
– 4 min read

Tech: A convergence at the Edge


A while back, Kaj Pyyhtiä asked an excellent question when we were debating some aspects of edge-compute: Isn’t this just the old telco edge [that never got anywhere?]1 This question was, at the same time, so right and so frustratingly wrong – that I could not stop thinking about it. The quest to figure out how to describe the difference led me to a bigger fish — a plausible, maybe even probable, convergence linked to the edge-compute.

As a Danish saying often attributed to Niels Bohr goes, “It’s difficult to make predictions, especially about the future.” The phrase feels particularly true when discussing technology or the “Telco Edge.” About ten years ago, the telecoms industry had difficulty justifying the costs of the big server farms they had built for the pre-public cloud era. Renting racks, servers, or even virtual capacity no longer paid the bills – as the clients moved more and more to Amazon and Azure. But you can not upscale and downscale physical compute at will: the telecoms had to figure out how to sell this capacity or underwrite the capital expenditure used to build it.

A new concept was born to consolidate this issue: the cloud is far away in Austin, Texas2, but some solutions need low latency. What if we’d rebrand the local server farms to this new concept: the “edge[of cloud]-compute”? If Jeff could create a new business by selling the old unused computers lying in his garage, so could we.

But there was much more ongoing than Amazon having some extra capacity when Amazon Web Services was born—a convergence of many technologies emerging as a completely new tech paradigm: the cloud. Yes, AWS took the market by storm, but that storm had been brewing for a while. Among others, virtualization, clustering, and network security had matured to a point that allowed a new way of buying, selling, and using the compute.

While telcos could see multiple emerging new technologies linked to the “edge,” - the technologies were too telco-centric, or there was not enough wind in the market to get the concept flying. The result – often referenced as the Telco Edge - still exists and has enough use cases to make it viable for the niche it works for. But seeing it as an excellent opportunity for new business is challenging.

The difference between a tech convergence predictor and an actual signal. A predictor does not mean the phenomena will manifest.

The failure of Telco Edge did not take away the local compute. And where there are extra resources, the market tends to find a way to make those available. This is where things get interesting: after discussing with multiple vendors and professionals - I can see a very plausible possibility of a new storm brewing.

An illustration of tech trends converging to a megatrend

The difference between a tech convergence predictor and an actual signal. A predictor does not mean the phenomena will manifest.

Now, this is an early prediction. I’m not saying there will undoubtedly be a new paradigm, like the cloud, in 2 to 5 years. But I would not be surprised if there is. I’m also hesitant to postulate if it’s just a third way of doing cloud - or as separate from the cloud as cloud is from on-prem.

I base the prediction on multiple maturing technologies and clear business drivers that all make Edge adoption possible and desirable for companies:

  1. CDN- and cloud-provider offerings for Edge are maturing fast and have a fantastic developer experience compared to traditional cloud, like the
  2. Edge-runtimes – a new container-like concept that is cheaper and simpler to manage and run than containers or virtual machines. While these runtimes have limitations, the benefits for developer experience and running costs are likely more significant than overcoming the limits, significantly when leveraging
  3. Data handling at the Edge to significantly lower cost and simplify solution architecture. Which in turn can be combined with
  4. Security at the Edge to sidestep private and health data issues and regulations completely – without complicated anonymization and security schemes.

When security, tech, and cost benefits start accumulating – it’s tough to say that the Edge is not a rising opportunity. How big, and if it’s just another way of using cloud - are still open.

Footnotes

  1. the part in parenthesis is my interpretation; Kaj’s intended message might differ.

  2. while not all of the cloud sits in Austin, the cloud is still located in a handful of huge regional data centers. There are tens, if not hundreds, of edge-nodes for each cloud region.