Breaking the DevOps Barrier

John Heveran, Senior Vice President & Chief Information Officer, Commercial Insurance, Liberty Mutual Insurance
144
205
48
John Heveran, Senior Vice President & Chief Information Officer, Commercial Insurance, Liberty Mutual Insurance

John Heveran, Senior Vice President & Chief Information Officer, Commercial Insurance, Liberty Mutual Insurance

When we began pursuing our DevOps and Agile journey a few years back, it seemed daunting for many of us. With several legacy systems and tenured people that have fueled our success in the years and decades before now, the idea that we could transform how we develop and deploy our most valuable software assets seems like a challenge too big to conquer. Would we be up for the challenge? Would our IT architecture allow for seamless integration? Was it even possible?

As a runner, I’m reminded of the challenges facing the late Roger Bannister in 1954 when he broke the four-minute mile barrier. At the time, everyone thought it was a barrier that was almost unbreakable. And like Mr. Bannister, once the initial barrier was broken, there was a mental barrier that was broken on that day in May of 1954, with another runner breaking the barrier just 46 days later. And in the subsequent 50+ years, over 1,000 runners have broken the four-minute mark, where before it was accomplished by Mr. Bannister, it seemed as if the record would never fall.

Our journey with DevOps has been similar, once we broke the mental barrier that existed for many of our teams, we have not looked back. We had a couple of pioneer teams and leaders, and once they showed what was possible within our environment, it has been team after team also breaking through the mental models of our perceived inability to achieve true DevOps with our systems architecture or processes. But as a developer put it recently at one of regular SAFe planning events, “the life of developers is getting a lot better around here.”

  This early success challenged us to think more creatively and aggressively, especially giving us the bravado to take on our legacy platforms 

We began an agile, cloud and DevOps journey in 2015 in earnest. Historically, developers had a huge lead time for delivery of infrastructure, middleware, databases, etc. This reinforced the waterfall model we were operating, which was also reinforced by extended lead times on credentials, secrets, pipelines, testing, and product delivery. But we set our sights on a new agile and DevOps paradigm and prepared and trained hard for the transformation.

Climbing over the first barrier

In 2016, my organization embraced the journey fully, after a couple of small-scale pilots that showed enormous promise. However, many were still skeptical since these efforts were greenfield over the bulk of our systems portfolio which is more established. These new efforts, which involved significant challenges and scope in their own right, were pacesetters for the rest of the organization. We created a completely new cloud-based platform, supporting the launch of a new business segment, in record time (seven months) and by the time we went live, we were doing multiple deployments to productions of an end-to-end insurance platform multiple times per day! The removal of barriers meant something big: we were now delivering infrastructure, middleware, databases, etc. and went from weeks and months to minutes and hours we must be better at delivering total solutions.

This early success challenged us to think more creatively and aggressively, especially giving us the bravado to take on our legacy platforms. In fact, we took on the paradigm and history at large enterprises of ‘we can’t do that here’ by establishing a team of half-dozen engineers, with some help from one of our cloud partners, to tackle some of our thorniest legacy systems.

While the team thought what we were doing was interesting, not everyone was confident it was practical or even prudent. But, by the end of the first week of a 10-week engagement, we had identified a handful of applications that we were going to tackle. In fact, by the end of the first day, we had moved from sticky-notes to coding! One of the most significant outcomes of that early event, was formalizing the ability to move delivery of key infrastructure components like servers, security, middleware, databases, etc. from weeks to days, but more importantly was the ability to automate these steps into development pipelines and provide the controls directly to our developers. My favorite sentiment from that experience was after hearing that of streamlined automation and orchestration of build and deploy pipelines, a team member exclaimed, “You mean I don’t have to create tickets and wait weeks to move forward? This is radical–good radical!”

What we’ve learned

So, what we have learned with these experiences is that taking on aggressive goals and getting the folks from your team engaged in solving remains the key to all success. Like those who have since followed Roger Bannister in breaking the four-minute mile barrier, I now have an entire organization breaking barriers. Today, we have almost 97 Percent of our new builds in the cloud-based environments. And 18 percent of our workloads, including a lot of legacy workloads, have transitioned to the public cloud successfully, with the teams being on track to double that in 2018.

But in the end, the above are just finish line benchmarks for our broader agile and DevOps journey. As one member of my team described the secure-cloud development pattern we have established, “Before this was available, we’d create a repo, then submit several tickets to get it all configured for our workflow (prevent deletion of certain branches, require pull requests, etc.). Now we set everything up immediately after creating the repo, no more waiting for tickets to be completed before we start coding.”

If you about to embark on a similar journey, I recommend these steps:

Be Bold–Establish clear, and aggressive, goals but don’t miss out by being too incremental
Be comprehensive–DevOps relies on the power of agile, cloud-based paradigms, migrate to modern architecture patterns, and automate everything
Engage your teams–Top talent is motivated by bold moves where they can play a role
Test and learn–You will make mistakes, we did, embrace them and learn from them

About Liberty Mutual Insurance

Liberty Mutual’s purpose is to help people embrace today and confidently pursue tomorrow. Keeping this promise means we are there when our policyholders throughout the world need us most.

In business since 1912, and headquartered in Boston, Mass, today we are a leading global insurer with operations in 30 countries and economies around the world. We are the third largest property and casualty insurer in the U.S. based on 2017 direct written premium data as reported by the National Association of Insurance Commissioners. We also rank 75th on the Fortune 100 list of largest corporations in the U.S. based on 2016 revenue. As of December 31, 2017, we had $39.4 billion in annual consolidated revenue.

We employ more than 50,000 people in over 800 offices throughout the world. We offer a wide range of insurance products and services, including personal automobile, homeowners, commercial multi-peril, property, general liability, commercial automobile, surety, workers compensation, specialty lines, and reinsurance.

Read Also

Operational Visibility in the SDDC

Operational Visibility in the SDDC

Dirk Wallerstorfer, Technology Lead for SDN and OpenStack, Dynatrace
Ask Your Managers What They Really Need

Ask Your Managers What They Really Need

AshLea Allberry, VP-Operations, Nanonation, Inc.
The DevOps Culture Change Goes Way Beyond IT

The DevOps Culture Change Goes Way Beyond IT

Raman Sapra, VP & Global Leader, Dell Digital Business Services
When Do-IT-Yourself is Not Good Enough?

When Do-IT-Yourself is Not Good Enough?

Scott Cleland, Senior Director of Product Marketing at HGST, a Western Digital Corporation brand