The AIOps Venn Dilemma

Whilst preparing our recently launched paper, “AIOps of the Future: A Definitive Guide, ” one thing came through loud and clear from the carriers we spoke with – assembling the skills-mix required to support AIOps tools / programmes. The carriers simply don’t seem to be confident in being able to find, or keep, people with the right skills to drive the AIOps tools and related programmes that they will invest heavily in over coming years.

They have a really valid point about how hard it is to adequately resource AIOps initiatives. I equate it to the Venn Diagram below.

  1. Network Skills (gold) – It has been difficult to find people with deep networking skills, especially in some of the less common domains. Luckily for the telco industry, Cisco has prepared comprehensive training courses like CCIE, CCNA, etc that have been a win-win for them and the broader industry, across IP-centric networks anyway. Network skills are arguably the easiest skill to find on this Venn diagram, but I’m certainly not implying that it’s easy (far from it!)
  2. Automation / Programming Skills (blue) – There are many coders available on the market and therefore lots who can create automations. Low/no-code automation tools are also coming onto the market, to reduce the skills barrier somewhat. The challenge here is actually finding coders who understand networks well enough to write network automations (ie overlap point #3).As networks become more software-defined and more closed-loop assurance mechanisms are sought by telcos, resources with overlap point #3 become ever more vital (but really hard to find)
  3. Data Science Skills (tan) – Larger telcos already tend to have many data scientists on their teams, so you could argue that the tan ring is rare, but not exceedingly so. However, if you’ve ever dealt with telco data science teams, you’ll have noticed that there are very few who really understand telco data. They’re brilliant at preparing queries and data science algorithms, but in my experience at least, few really understand the business / operations implications of the data they’re working with. Without a deep understanding of networks, it becomes challenging for data scientists to really do what’s necessary to support AIOps programmes (ie overlap point #1). In turn, this means few data scientists also have the skills to create network automations (ie overlap point #2)

With AIOps tools still in their relative infancy, the tools can’t yet supplant the need for the above mentioned skills. The tools are great frameworks, but they still need deep skills to tune them, customise them and get them delivering their expected benefits.

So, what’s the solution?

We can’t just complain that the skills aren’t available. Nor can we just complain that the tools don’t deliver everything we need to overcome the skills-mix shortfall.  The tools simply aren’t the “remove all ops head-count” panacea that’s often desired. Complaining is just a “head-in-the-sand” approach that doesn’t solve anything.

In the meantime whilst AIOps solutions mature, we have no other choice than to develop the skills we need. Then we must keep the roles interesting enough to retain the talent rather than churning to other industries that have equal need for these valuable skills / resources too.

There are a few challenges that need to be overcome here though too:

  1. Historically, telcos have always been adept at establishing master / apprentice streams – for the network domains anyway. Unfortunately many of the network “apprenticeships” have been outsourced for the last couple of decades. This is even more applicable for the IT / developer type roles within telcos. There’s been a hollowing-out of “masters” available to teach the apprentices in many telcos in recent times
  2. For Data Science and Software / Coding  skills, there also typically aren’t enough in-house masters to train the next generation
  3. Overlap area #4 is almost impossible to find, especially since each ring is a career-path in its own right. Those “masters” are likely to be in such high demand and so valuable that they’re unlikely to be available to train apprentices. Therefore, significant fore-thought must be given to finding the right balance of short and long-term benefits

The telcos simply have to invest in the overlapping areas, either as individuals, or  by developing high-functioning teams that work together well to fill the void.

The only alternative they seem to have is to continue to put their faith in suppliers having the right skills-mix to support the telco industry needs. That seems to be delegating responsibility for something that just seems too important to me.

I’d love to hear your thoughts. How have you overcome the AIOps skills-mix challenge, or what mitigations are you already putting in place?

If this article was helpful, subscribe to the Passionate About OSS Blog to get each new post sent directly to your inbox. 100% free of charge and free of spam.

Our Solutions

Share:

Most Recent Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.