HMRC uses Hadoop to tackle corporate tax avoidance

HM Revenue & Customs (HMRC) is driving the use of open source technology with a Hadoop NoSQL big data engine to analyse corporate tax.

Government austerity measures have driven HRMC costs down by 20% over the last four years. The organisation is committed to reduce costs by another 22% over the next four years.

Addressing delegates at the Open Source Open Standards 2014 conference in London, Mark Dearnley chief digital officer of the HMRC, said open source software was a great way to change the dynamics of how software is developed.

According to Dearnley, analytics offered among the biggest opportunities for the use of open source software at the HMRC. He said: "Analytics is the first area where open source software has led the thinking."

Working with its system integrators, HMRC has developed a macro enterprise data hub, built on Hadoop. Dearnley said: Open source software is more cost-effective. It drives the commoditisation of infrastructure and use of software and drives a different delivery model, which is massively more cost-effective.

Corporation tax compliance is another example of Hadoop at HMRC. In the UK, companies need to submit tax returns electronically in the iXBRL format specified by HMRC.

Dearnley said it took two and a half months to develop a complete Hadoop stack and load in all the corporation data, allowing tax officers to start analysing company tax returns. He said the users were impressed by how fast IT delivered and the speed with which they could get value.

While using Hadoop for analytics has proved the value of open source software at HMRC, he said his ambition was to create a level playing field for open source software: "At the moment the pendulum is a bit too far, the other way."

HMRC runs 5,000 servers but only 3% run Linux. A quarter of its systems are virtualised, mainly on VMware, and it runs 3% of its system in the cloud, he said implying a substantial opportunity to deploy open source technologies in HMRC's infrastructure. Of the 500 enterprise applications at HMRC, Dearnley said 95% were based on proprietary platforms.

He admitted the penetration of open source software at HMRC was low: "We have some way to go. Our future will be a combination of private and public cloud, commodity compute, some of our databases are rather large and don't run in virtualised environments, so we will optimise our database cloud."

See the article here:
HMRC uses Hadoop to tackle corporate tax avoidance