Lloyds makes the jump from service virtualization to full DevOps
Bank winds down service virtualization centre of excellence declaring programme a success, having achieved cost savings, faster delivery times, and improved app quality. (picture credit: Simon Dawson/Bloomberg)
Lloyds Bank is closing a pioneering test service virtualization centre of excellence so it can focus on a full shift to DevOps and continuous delivery, according to Paul Downing, senior manager of QA environments for the Lloyds Banking Group.
The bank established the centre three years ago to accelerate the use of virtualization across the enterprise. According to Downing, the result was a 300% return-on-investment from the project, faster delivery times and improved app quality.
The centre served as a centralised team that helped with training and implementation of service virtualization throughout the bank’s IT department. CA Technologies provided the bank’s service virtualization tooling.
The key objective of the project, said Downing, was building self-sustaining service virtualization capacity throughout the bank so that the technology could be used independently of the centre. Downing’s team supported and trained specialists who then went on to implement their own virtualization projects.
Service virtualization is a technology that simulates dependencies in apps. These dependencies can be inputs from other pieces of hardware, APIs, or back-end applications. For example, using service virtualization, testers can test an app that interacts with a mainframe. Once the mainframe is virtualized, testers are not reliant on accessing the actual piece of hardware, which might have restrictions on when it can be used. Apps can then be tested earlier and more often.
The Lloyds Bank test centre of excellence was composed of 12 virtualization experts, tasked with training other employees to use the technology. These experts would then go and train their own teams. Downing said that typically training would consist of virtualizing one constraint (a server or an API for example) for an app. That employee would then go back to his team and virtualize the rest of the service.
According to Downing, the programme had three key objectives: achieving faster delivery time, building better quality applications by being able to test sooner and more often, and saving money by discovering bugs earlier, when they are less expensive to fix.
Service virtualization helped testers at Lloyds Bank ‘shift-left’, or test earlier on in the software lifecycle, said Downing. “In finance we are working with large, complex systems, with a lot of legacy. We wanted to move faster, and that meant more testing.” He added that using service virtualization his team “Could test our applications in isolation. We could avoid traditional production-bottlenecks like having to request a data-set or requiring access to a mainframe.”
But it was not all smooth sailing. Downing said that there was an initial push back from upper management, who did not understand the technology and were unwilling to commit resources from their already overdrawn budgets.
But once Downing got a small project off the ground demand for the service snowballed. “Developers and testers started coming to us asking for the virtualization option – that is when I knew we succeeded,” said Downing. Three virtualized environments in 2013 became 20 in 2014 and 30 in 2015. Downing estimates that approximately 800 users work with virtualized environments yearly.
Service virtualization is now self-sustaining throughout Lloyds. Teams have the necessary knowledge and support to create virtualized services whenever they need them said Downing.