Since middle east banking giant Emirates NBD completed its datacentre consolidation project in January 2010, it has reduced operational spending on its datacentres by 40 per cent while improving uptime statistics.
Speaking at analyst firm Gartner's Data Centre and IT Operations Summit 2011 this week, Sajid Lokhandwala, vice president, IT infrastructure and Operations Group, Emirates NBD, described the benefits of the project.
"The consolidation of network services reduced overhead cost, which in turn reduced Opex spending by 40 per cent. We saw further savings from the standardisation of network links, and reductions in staff requirements and power consumption."
He added that there has to date been no unplanned outage since the project was completed, and the 98 per cent uptime service level agreement (SLA) has been met.
The consolidation project, which reduced five datacentres down to two, began in early 2007 after Emirates bank and the National Bank of Dubai merged.
The merger brought the datacentres together under one corporate umbrella, with different disaster recovery mechanisms, competing protocols and standards, and new departments with similar objectives.
"There were too many connections, too much complexity, and a lack of standards as the two organisations came together," said Lokhandwala.
"The driver behind the project was to capture the synergies from merger. We needed to reduce the administration and management overhead, consolidate and rationalise the application landscape, as well as improve availability."
This by itself may sound challenging enough, but many of the involved systems were business critical, with no tolerable downtime. The business also decided against a technology freeze, essentially giving the project a moving target as the IT landscape was still dynamic.
Lokhandwala explained that as with so many projects, its success was all in the planning.
"We spent six months analysing what could go wrong to miminise surprises later on.
"One problem was that we were not able to establish the connectivity requirements of every application because they had different standards, with two companies coming together. But we were aware of the problem at least."
With so much complexity involved in the project and so much at stake in its success, the decision was made to hire external consultants to help with programme management.
"We had skilled manpower but people were extremely busy carrying out other projects at the same time. We decided to engage a partner to help with programme management and act as an arbritrator."
Information infrastructure firm EMC was brought in as consultants, and its Smarts application discovery management tool was used to help find out exactly what applications were running, and what needed to be migrated.
"Once we had this information, we designed the application migration programme," said Lokhandwala.
The actual migrations were bundled together by criticality of funtion, measured by the business-critical nature of the system, the SLA in place and the operational risk of the migration.
Tier one bundles comprised the most critical systems that were able to tolerate less than an hour's downtime. By contrast, tier four systems were able to be offline for 48 hours.
The project was successfully delivered on time and on budget, and Lokhandwala puts that partly down to effective project governance keeping the project teams on target.
"On the project governance side, we had a small team dedicated entirely to the project. We also had steering committee of senior management overseeing it from a higher level, and smaller workgroups across key application sets."
He finished with the news that he is about to do it all over again, thanks to another acquisition.
"Two months ago we acquired another bank, and in January we're starting another datacentre consolidation project," he concluded.