How DBS HK Became the First DBS Franchise to Decommission Its Mainframe
Introduction
In June 2019, DBS Bank Hong Kong (DBS HK) successfully decommissioned its mainframe — becoming the first DBS franchise to do so. My name is William Lam and I was the project lead on this endeavour. I’m proud to have had such a strong team along with me for this journey. In addition to eliminating significant licensing and infrastructure costs, the Mainframe Decommission Project marked a pivotal moment in our Bank’s transformation to become a fully-fledged technology company.
In this blog post, I’d like to share a bit about how our team approached this project, as well as discuss the critical enablers and lessons learnt.
Moving on from the Mainframe
The DBS Hong Kong mainframe, which had been around since the 1960s, had served the Bank incredibly well. However, the software that had accrued on it over the past 15 years needed replacing if we were to follow through on our goal of transforming the Bank to make it future-ready. The mainframe limited our ability to offer our customers more complex and personalised digital journeys.
Approximately 60% of what constituted the mainframe had been our legacy core banking system. This was migrated off the mainframe between 2013 and 2019 during DBS HK’s Core Banking Transformation Project (2013–2019). The project bundled together not only the migration of our core banking systems to a new hybrid stack, but also the revamp of our front-end channels — internet banking, mobile banking, phone banking and ATM. So, with the majority of our applications now on a new hybrid stack, there was no longer a need for DBS HK to have a mainframe.
The residual 40% of what sat on the mainframe included data storage and was in essence a clean-up job. The challenge, however, was that we didn’t quite know what was left. It was therefore a process of discovery — akin to going through archaeological layers — to find out what those residual applications were really doing.
So, how did we do it?
When I joined DBS Bank Hong Kong in 2015 as the Team Lead for Core Systems Technology, the Core Banking Transformation Project had already begun. Then, in early 2018, Nimish Panchmatia (Head of T&O for Hong Kong & China, 2016–2019) and Siew Choo Soh (Group Head of C2E, 2015 to present) approached me with the task of decommissioning the remaining applications and systems. Our team submitted the business case in March 2018 with the target of shutting down all the applications by June 2019.
The first question we had to ask ourselves was, “Do we understand enough about what it is on the mainframe in order to switch it off?” In fact, there are many companies who are still running their mainframes, even after making the move to digital. This is because they haven’t been able to understand what is on their mainframe.
To answer this question, Nimish put two conflicting teams on the task to understand what was on the mainframe — myself on the software side and our infrastructure team lead on the other side. Working with the infrastructure folks, we quickly discovered more and more of the applications which were still running on the mainframe. These applications hadn’t been visible to us as we had been working from an applications perspective and had not been able to monitor the utilisation of CPU. We were then able to perform a complete stock take and identify a go-to state for each application on the mainframe.
The next question we had to answer was which applications were covered by regional programmes. While the majority of legacy applications on the mainframe were already covered by regional programmes, we still needed to do our due diligence before we decommissioned them to make sure all functionalities were being migrated. We found that around a dozen of the applications’ functions — such as some local data applications — were not handled by the regional team. We therefore mapped out these applications by functionality and re-wrote each functionality into an open system.
Every time we moved the data, we saw the overall amount storage on the mainframe go down. Similarly, every time we switched off an application, the infrastructure team monitored the CPU usage to see whether it would also go down. We continued this process until we eventually saw both CPU and storage drop to zero. I think having that ‘friendly conflict’ between my team in applications and the infrastructure team helped to hold us accountable. It enabled us to find hidden legacy applications on the mainframe that we might have otherwise missed.
For example, there were times when our team thought we had switched something off, but the infrastructure team would come back and say that the same amount of CPU was being used and there were still terabytes of data sitting there. As we had been working from an application perspective, this information was not visible to us. This collaboration was really a demonstration of our OneTeam spirit in T&O — working towards one shared goal.
What were the critical success factors in this journey?
1. Empowering the team to take ownership
· The key success factor in this project was giving ownership to people within my team — just like Nimish and Siew Choo had given ownership to myself and our infrastructure lead. I had been a mainframe developer previously, so I was familiar with the job at hand. However, I still had to get up to speed on the underlying infrastructure and hidden legacy modules.
· Giving ownership to people who I knew had more experience with some certain legacy modules, enabled them to identify & implement more effective solutions than I would have done. Had I not empowered my team and sought their support, we would not have been able to accomplish this task. The quality of our team in Hong Kong is incredible.
· What’s more, whenever anything went wrong, Nimish and Siew Choo were there to step in and help us out. Just as Nimish and Siew Choo checked in with me regularly, I would also check in with my team .
2. Keeping an eye on the bigger picture
· As relayed earlier, while the Core Banking Transformation Project was a landmark technology project for DBS Bank Hong Kong, the Mainframe Decommission was in essence a clean-up job. It would have been very easy for the Bank’s management to turn its attention to other more exciting projects. What’s more, it would’ve also been easy for my team to feel this was a project for which they would not be recognised. However, everyone in the Bank — from working level to the programme steering committee — realised that the Mainframe Decommission was an essential component to making the Bank future-ready.
A Pivotal Moment for the Bank
The Mainframe Decommission project is a journey for which I look back on with a lot of pride. The highlight of the journey was undoubtedly in June 2019 when we stepped into the Hong Kong Data Centre with all our team members, our DBS HK CEO Sebastian Paredes and our then Group CIO Dave Gledhill to flip the switch on the mainframe. We made a big deal of the occasion and celebrated with champagne. We also invited IBM — the mainframe provider — to join us for the occasion. Funnily enough, IBM shared with us that it was the first time one of their customers had ever decommissioned their mainframe!
Even now, people still ask us, how did you manage to switch the mainframe off? That is a testament to the complex nature of the task, and an achievement for which I will always be very proud of my team.
