Insurance market in mammoth DC migration: 70b rows of data, 200 critical apps, 1 weekend...

Some 70 billion rows of data, 700 IT assets including mainframes, 200+ mission-critical applications, a team of 500: DXC Technology and partners have successfully migrated an entire data centre powering the world’s largest insurance market over to new facilities – lifting and shifting Linux, Solaris, Unix systems and more to a more modern facility over a single weekend as part of an ongoing transformation of the Lloyd’s and London markets.

Speaking to The Stack, Andy Evans, Delivery Director, DXC, said the migration had been a “team sport” that required meticulous planning – with over a dozen partners on hand in the new data centre with a “truckload of parts” in case there were any unexpected hardware issues. (There weren't and the migration went smoothly.)

“Meticulous planning, delivery execution, and exceptional collaboration with the customer community was the recipe for success,” added Rob Myers, Operations Director, Lloyd's Market Association.

The shift was a critical step in the Future at Lloyd’s programme which envisions a far more data-focused, automated, and cost-efficient future for an insurance industry that still relies heavily on paper and pdfs – with the broader industry modernisation programme being led by London Market Joint Venture comprising Lloyd’s, DXC Technology and the International Underwriting Association (IUA) in a strategic, multi-year project that will ultimately see the creation of a wide range of new (rather than re-factored) cloud-native apps running in AWS.

The migration involved a shift from an older “inherited” DXC data centre to one of its modern, resilient data centres that has high-speed interconnects to cloud providers, more modern hardware, cooling, and power supplies. The shift included porting mission-critical apps running on mainframe to newer mainframe hardware platforms as part of efforts to mitigate legacy technology risk and improve performance across the market.

The migration took place during a single weekend in February but had involved years of “meticulous stakeholder management and planning” as well as multiple dress rehearsals, Evans told The Stack this week.

He said: “This was one of our biggest and most ambitious data centre migrations: We've migrated the entire London insurance market operations, including all the data, the apps or the business processes, from one set of data centres, to the DXC strategic data centres over a single weekend. When you look at the size and scale of that it is quite extraordinary; thousands of databases; 200 applications that really fueled the market [and which] 400+ customers use in their everyday live  insurance markets… Ordinarily, what we would do with these kinds of migrations is we would do them in ‘move groups’ but the customer really wanted to minimise the disruption to the actual market and try and get it done in a single weekend to minimise the downtime," he added.

Making that happen involved a team of some 500 staff on three continents and close engagement with myriad stakeholders across the London insurance market. How did his team pull it off? “This [industry] represents a quarter of the GDP of London, so there was a little bit of stress, a little bit of pressure” he admits, “but this was a multi-year planning process… we set out right from the start with partners that they were going to be part of something that was gonna be quite special and quite big, but they had to pull in the right direction; and they had to fit in with our governance, fit in with our very, very strict change rules. They can't do anything out-of-step; they're not authorised to do any things that are not part of our kind of structured processes and changes...

“So [it involved] very, very structured planning. Each partner had an owner within DXC who did the communication with them, with very transparent rules around how they operate. Certain suppliers had to turn up at certain times; certain partners were there throughout on standby. We had partners on standby for every kind of eventuality of what might happen, including getting parts for our systems in case we had a motherboard failure or a fan failure or something. So trust and transparency and very, very clear signalling of the rules of engagement and where they fit into that runbook. We did a lot of runbook dry runs, where we would literally sit down with each of the teams and we go, ‘right, this is, this is where you fit into the plan.’

“[Operating at this scale] just requires meticulous stakeholder management and planning.”

London insurance data centre migration

https://www.youtube.com/watch?v=-GqqMCVG8Oo&t=130s

Lloyd's of London, a specialist insurance marketplace with 76 syndicates, 350 brokers, and 4,030 coverholder offices employing 47,000 people across the UK and writing gross premiums worth over $100 billion annually, continues to work towards delivering its "Blueprint 2". That's a digital transformation programme that includes plans to build a raft of new containerised applications underpinned by AWS EKS, a data reporting and analytics platform using a combination of S3, Redshift, and Qlik tools that cover data capture, data integration, report product generation, data catalogue and self-service report and API Access, coordinated by DXC.

The DC migration set the insurance market up for that by pivoting to improved connectivity and consolidated infrastructure, even if many applications continue to run on Big Iron. As Evans puts it, it was "moving all kinds platforms to newer platforms; consolidating some of the technical stacks. [For some areas] we've gone from a mainframe to a mainframe, but it's a much newer mainframe with higher availability and faster: in some cases, we're seeing upwards of seven-times faster processing capability across some of the technology stack."

Down the line, Lloyd's April 2023 FAQs on Blueprint 2 confirm, that ultimately these digital services for premium and claims are going to be "new business applications, operating on a new cloud platform. Once they are live, the existing mainframe systems (PoSH, LIDS and all CLASS / ECF versions), cease to be operational and will become historic data repositories. This ensures no data is lost and removes the risk of a mass migration at cut over."

For now, the near-term priority is, as Evans puts it, "we must make sure that all migrated systems and applications are stable, we've returned to business as usual process flows with those services, and are ensuring maximum availability of those services. So we're not taking our eye off that and not switching focus. Priority number one is run the day-to-day really well; make sure the market gets the best service that they can possibly get and we are still tuning things a little bit. But we also have hundreds of people working on Blueprint 2, developing those applications and meeting the major milestones in the market. We completed [the first two sequences of that process] in March, our teams are fully focused on sequence three and sequence four; working very closely with the JV, very closely with Lloyds and and the wider market on on that is an equal priority for us."

Follow The Stack on LinkedIn