May 10, 2018 Francesca Krihely


Barclays built an Operational Data Platform to offload expensive mainframe operations, and store and process customer transactions for business operations, analysis and reporting. You can read the full story on Barclays journey off the mainframe in Diginomica and Computerworld. I've produced a high-level summary for you below: 

High profile outages as a result of mainframe failure

Barclays new digital banking services were struggling due to heavy reliance on legacy infrastructure and apps. The bank experienced two outages that prevented customers from making payments -- all as a result of mainframe failures.

The continued growth in traffic and launch of new digital services led to increased cost of operations and decreased performance. For their team, the mainframe was single point of failure for many applications. As a result, outages resulted in poor customer service, brand erosion, and regulatory concerns. 

Moving off the mainframe to the Data Lake 

After analyzing their digital channels, the team learned that that 92% of traffic generated by 25 transaction types, 85% of these read-only.

To speed up operations to near real-time, they created operational data lake (ODL) to offload these operations from mainframe MongoDB based ODL to power existing apps, new digital services and other APIs. MongoDB's document structure allowed de-normalized data, reducing complexity while increasing performance.

Five Million hits to one query 

After implementing the ODL, Barclays saw the following results: 

  • Reduced time to market for new digital services, including personalization
  • Simplified regulatory compliance (e.g., PDS2)
  • Stand-in capability to support resiliency during planned and unplanned mainframe outages
  • Reduced number of read only transactions to mainframes (MIPS cost), freeing up resources for additional growth
  • Reduced reliance on availability of specialized hardware and qualified mainframe staff

Barclays plans to expand the ODL and make it a first port of call for customer data. One example where this will help improve the customer experience and cost is for their online and mobile banking pages. Considered one of the most intensive data processes, online and mobile banking landing pages requests hit the mainframe five million times a day for one of the application's heaviest queries. In MongoDB, this query is a single lookup, making it simpler and cheaper for the bank to provide the high availability and performance their customers expect. 


Previous Article
How to Integrate MongoDB Atlas and Segment using MongoDB Stitch
How to Integrate MongoDB Atlas and Segment using MongoDB Stitch

It can be quite difficult tying together multiple systems, APIs, and third-party services. Recently, we fac...

Next Article
Learn more about transactions at MongoDB World
Learn more about transactions at MongoDB World

In February, we announced that MongoDB 4.0 will support multi-document transactions. Curious to know what t...