Fotolia

Green Flag waves farewell to Oracle legacy with cloud-native revamp

The roadside breakdown rescue service has developed a technical architecture to support an ecosystem approach to business

When Dean Keeling joined Green Flag as managing director three years ago, top of his to-do list was transforming the business and setting it on a trajectory to become the UK’s number one roadside recovery and rescue organisation.

The firm is the UK’s third biggest roadside recovery firm, behind the AA and RAC, with a £160m turnover and a customer base of around 3.5 million subscribers. 

Its parent company, insurance giant Direct Line Group, has – in Keeling’s words – ambitions to be the number one player in every business category it operates in, and rescue is no different.  

“Direct Line Group wants to be more than the biggest motor insurer in the country,” he says. “They want to be a market leader in a number of categories, and with rescue they tried to do a few things in the past and it has not quite worked, so the growth objective was what I was recruited to identify.”

This was based on his past employment experience of overseeing growth initiatives at Proctor and Gamble, Domestic and General, and British Gas.

“My approach to that was to seek to understand what the business model was and what the opportunities were to improve,” says Keeling.

There were a couple of aspects of the Green Flag business, which were already working in its favour, including the fact it is a “data-rich business” with a distinctive operating model that relies on connecting motorists with independent roadside recovery businesses local to wherever they have broken down in the UK.  “Green Flag is really like the ‘Uber’ of rescue, but it is a 49 years old business with a very, very modern operating model,” he says.

Read more about microservices

  • Ready for a migration to microservices? Here are the steps your development team can take to gradually transition your existing monolithic applications.
  • Before you build a microservices application, take a closer look at the components of the architecture and their capabilities.

Indeed, the on-premise technology infrastructure underpinning Green Flag’s operations is built around a 40-year-old Oracle database that is robust enough, but would struggle to accommodate any big changes to how Green Flag operates or the services it provides.   

“We wanted to start evolving the business into being fit for the purpose for the next 10 years, though, rather than just doing okay with what we’ve done for the last four or so decades,” he says. 

For example, rather than just being there to help customers during urgent and acute times of crisis, such as a roadside breakdown on a motorway, Keeling sees opportunities for Green Flag to move into proactive vehicle maintenance, such as allowing motorists to connect to garages where they can get a service, MOT or upgrade their vehicles cosmetically.

A question of legacy IT

However, as Shakeel Butt, technology and delivery lead at Green Flag explains, the system suffers from high running cost and offers little in the way of agility, which is typical of many legacy IT systems. “It is difficult to get access to data and hard to add new features like support AI or connected cars,” he says.

The company realised it needed to have a fundamental rethink of the technology underpinning the Green Flag business. “The challenge is what I can build to enable Dean’s ambition such that tech isn't a blocker,” says Butt.

During the initial planning stage of the project, Green Flag ran a comprehensive due diligence process looking at all the options available to modernise the IT that ran the business. “One option was to refactor and encapsulate the existing systems,” he says.

However, the legacy Oracle-based application was monolithic, which meant the whole system would have needed to be tested as and when changes had to be made. “We eventually landed with the idea to build a new system from scratch, while running the legacy IT in parallel,” says Butt.

To provide the company with the ability to move at pace, the new technical architecture is cloud-native, based on standalone microservices and is API-led (application programming interfaces).

The technical architecture offers Keeling and the business an approach that enables the company to grow to become what he describes as “a digital microservices business where we are amazing at using data in real time”.

For Keeling, the architecture gives Green Flag the ability to create an ecosystem in which, he says: “we are in charge of being able to respond to what the market and the customer is asking for”. 

Single functional building blocks

Microservices are often described as Lego bricks. An individual brick may not do very much, but as a whole, they link together to build something useful.

Those organisations looking at a microservices-based approach to developing new software need to decide how granular their microservices should be. If it is too granular, application development teams need to do a lot of work to build linking microservices together in order to build something useful; if the microservices have too much in-built functionality, they suffer from the same limitations as legacy monolithic applications, where it becomes difficult to make small code changes quickly.

Butt describes the Green Flag approach as an exciting architecture. “We started with domain-driven design and looked at four key types of users – the direct customer, member of staff, the roadside rescue network and the technical support and maintenance teams,” he says.

This approach has dictated the functionality required, which Green Flag has delivered as core platforms, built-in Python using single function microservices. For instance, as Butt explains, the rescue platform – used by the roadside assistance network – uses 12 microservices.

“By keeping the microservices small, we were able to offload the heavy lifting by consuming AWS services,” he says. Broadly, the principle driving the microservices-led development of the core platforms at Green Flag has been to use an AWS service if one exists rather than reinventing the wheel.

“If there is a native AWS service, we should absolutely consume that,” says Butt. For instance, the company is using the AWS API Gateway for API management, the AWS FarGate serverless compute engine and AWS Lambda containerised services.

Resourcing

Like many organisations, Green Flag previously used offshore resources for IT. When it came to building the new systems, Keeling says he took a decision to create in-house expertise.

“I used some internal talent for the key roles,” he says. “Then we have recruited in-house teams since the go-live, which was at the start of last year, in order to build up the various Scrum teams around our ecosystem.”

Given the ongoing skills shortage in many of the technical roles the company needed, Green Flag kicked off its transformation programme in Q4 2018 with support from Contino, the enterprise DevOps, data and cloud transformation consultancy.

Explaining the rationale, Butt says: “We had just 12 people on the ground, but needed 60 resources, so I turned to a partner to help us deliver [the people].”

Content Continues Below

Read more on Microservices

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.

SearchCIO

SearchSecurity

SearchNetworking

SearchDataCenter

SearchDataManagement

Close