How to stay ahead in DevOps

The world is changing rapidly and telecoms are becoming the enabler we need to pivot towards the Internet of Things and connected devices.

There was a ​great keynote​ recently from our CEO Hans Vestberg where he talks about the 8 currencies of 5G – which include things like speed, latency, and reliability that all have to be in balance. You cannot sacrifice one to the detriment of others. This struck a real chord with me as it is the same with the DevOps enablers. It’s all about balance.

So how to stay ahead in the race? Part of it is delivering value faster and more efficiently, which is what digital transformation is all about.

The Lenses

Everything we do is centered around business outcomes, and Digital Transformation and Customer Experience should be viewed through multiple lenses. ​For each of these personas we have to deliver appropriate digital tools to satisfy:

  • “What” they need to do
  • “Where” they need it
  • “When” they need to do it

Customers, partners, and employees are all equally important, so let’s start with our internal talent and give them the tools they require – or as we say “let’s drink our own champagne!”

Technology enablement and community mobilisation

We realised that our most valuable resource is our people. It takes an active community to accelerate change, which is why we formed TECmob. The program’s vision was to empower all our engineers to help take us to the next level – what we affectionately term Verizon 2.0. The main components of TECmob include:

  • Innersourcing
  • Evolutionary architecture
  • Chapters & Guilds

Innersourcing

Innersourcing to me means we’re trying to look at best industry open source practices and apply these within our firm with a massive emphasis on internal collaboration, sharing, and self-governance. To support this vision of build once, build right, and reuse by many, we realised we had to adopt an industry standard repo.

We found that best engineering practices were bred when our folks were working on the bright new shiny stuff! So, to kickstart our cloud migration journey, we constructed a cloudshare repo where diverse teams could submit templates that standardized the management of our cloud infrastructure. We used categorisations, like compute or DB, to assist teams in discovering and consuming these templates.

The results were pretty impressive with around 100 templates created and consumed by over 300 teams. Another success was a data leakage and detection runner that we created and made available against all our open repos to secure us from one key aspect of security breaches.

Our vision is where every pull request gets the same set of templated services run against it. We’re striving for that same level of quality and security to be standard – table stakes. That is the only way we will have low fatigue development cycles and drama-free production releases. A good leading indicator of your DevOps maturity is the stress levels of your release manager – are they pulling their hair out or are they cool as a cucumber?

In order to decrease costs further, specifically to save more than $150M over the course of 3 years, decisions were made to limit the use of commercial software and expensive support contracts, and further embrace open source and innersource.

One area we quickly realised we needed to move from being consumers to builders was the telemetry of all our teams’ software delivery pipelines. We needed a visualisation tool that would track our flow of work from ideation all the way through to production. Teams were measuring a lot of things, but not always the right things. To fix this, we partnered with another firm to opensource product to standardise and streamline what data is collected to baseline and track our progress over time.

I think it’s great we’re maturing where we can embrace this “entersourcing.” I love the new term “frien-emy” (friend or enemy?!) – do we really know whom our customers or partners will be in 1,2 or 5 years? Gone are the days when software vendors came to us with solutions to our problems. We know best as to what our problems are so let’s collaborate in the open and solve them.

At the core is delivering value better, faster, and safer. This is the holy trinity. Remember back to the 8 currencies of 5G – you cannot just focus on one aspect. There is no point in delivering superfast into production if your quality is shocking as you will need to rework and even more importantly you will negatively impact customer credibility. This is a crucial balancing act that all of our teams are very cognisant of.

Evolutionary Architecture

We also transformed our Enterprise Architecture process from being a top-down “command and control”. Where the gods in their ivory towers pronounced “thou shall use …”

… to instead empowering the people closest to Tech – the real SMEs. We wanted them to be accountable and make the decisions.

To inspire them, we created a workflow where anyone could suggest a technology, pattern, or practice in the form of an Enterprise Decision Record (EDR), and then engage in a transparent, fact-based debate with their peers through voting, commenting, and adding a ‘thumbs up’.

This was very much about “​by the people for the people” ​and that really struck a chord with our community.

To date, we have approved over 150 EDRs, and close to 100 are in progress as new tech hits the market daily, and our thoughts and decisions are always being re-evaluated.

We’re increasing visibility, openness, and the ability to measure by leveraging a Thoughtworks’ opensource framework called the ​Techradar. ​This provides a nice visualization of our architecture decisions and allows our community to search, filter on the various quadrants to recognise what we should be adopting or retiring, and understand why this is the direction.

Our community now has a ​stage​ and a ​voice​ – we just have to convince everyone to use it!

People are our fuel – the chapters

With people at the heart of each effort, we created 12 Chapters to get our experts with similar technical passions to meet-up and lead the transformation charge. We strongly encourage this positive behaviour as we don’t want to fall back to the bad old days of silos solving the same problem over and over again. We had to step up from taking short term tactical technology decisions to thinking broader and more strategically.

Each chapter is centered around tech topics such as logging & monitoring, databases, and so on. For membership, we ran a grassroots campaign to identify the best talent throughout the organization and found leaders who sought collaborative opportunities. Chapters broadly work on 3 levels around ​decisions​, ​knowledge, and solutions​. Chapters are instrumental in our EDR architecture decision process; they are the fuel behind our technology direction.

Chapters also actively participate in our Stackoverflow providing great insights and knowledge sharing whilst solving issues, answering questions, and providing technical solutions.

Our Chapters are super active in solving “Business Use Cases”. For instance, our RPA Chapter implemented a digital network assistant – the “DNA” bot with over 100 skills, saving over 40,000 hours and giving an omnichannel experience for our field engineers and call centers supporting Network installations.

What’s great for anyone in our company is we can quickly find out key decisions and gain knowledge on any technology all because our Architecture decisions, Tech Radar, and search engines are all joined up – discoverability is the key.

Recognise excellence and celebrate it


With all of that said, we have​ m​ore work to do​ ​ some of our chapter members are still not fully active. So, we’ve tested various hypotheses about why this could be the case – there is an element of you have been “voluntold” by your manager but our latest surveys (although show some global differences) do highlight that we need to celebrate success more and show more exec level appreciation and recognition.

It should be super clear by now that it’s massively important that any effort we’ve pursued we use metrics to track outcomes. And more so we further reward and recognize leading behaviours.

The chapters are innovators who will set an example and inspire the early adopters and others to follow. We should praise our engineers who are actively engaged and contribute whether it be in EDRs, Stackoverflow reputation points, or pull requests for innersourcing. We must ask ourselves – is our community going the extra mile?

We must inspire our employees to strive towards their stretch goals. So, we’re working very closely with HR and our communication teams to ensure we acknowledge, celebrate, and reward the transformation of our engineers. Let’s remember: they are the ones delivering the world class solutions that our customers want.

Written by John Scott, Digitisation and DevOps community leader at Verzion 

 

More
articles

Menu