Three years into CIO Peter Mahler's six-year IT transformation strategy at Coles Myer, the retail giant has implemented an end-user and systems availability solution to closely monitor its applications and infrastructure.
The Coles Myer umbrella covers 16 separate brands which, according to senior IT executive Peter Tonkin, each owned its own IT shop, were all operating independently, and suffered from a lot of integration and standardization issues.
"When CIO Peter Mahler joined in 2003 he [wanted to] create a single IT shop operating as a shared service," Tonkin said. "Our strategy is to buy not build, and we want technology stack standardization, and application consolidation, for example, a single SAP."
Now Coles Myer is in the midst of "lots of concurrent, large-scale projects" delivering applications that are fit-for-purpose. Tonkin said the applications, which are performance sensitive, need to be robust and, with the organization moving to an integrated architecture, they all have to talk to each other. Application scalability and complexity are also challenges Coles Myer is dealing with during its transformation phase.
With those requirements in place, Tonkin said Coles Myer went shopping for an application performance management solution that simulates and monitors the availability of business-critical, end-user transactions.
"We wanted something to assist in identifying those bottlenecks in the applications [and which] links end users and infrastructure with alerting and diagnostics," he said, adding that the solution had to be repeatable.
After defining a set of application performance monitoring requirements, and assembling an evaluation team, Coles Myer issued a tender and conducted a "bake-off" between vendors as a proof-of-concept.
"Mercury was the best fit [as] it supports the ongoing application performance management lifecycle," Tonkin said. "You can take it through from testing to production. The technology fits to our requirements."
Coles Myer, in conjunction with Mercury, implemented Business Availability Centre (BAC) for end-user management, system availability management, and a dashboard display.
The solution is now monitoring 10 key business applications, including merchandising, supply chain systems, and other core transaction systems. Also being watched is Coles Myer's online activities like its core portal, online retail sites, and the new B2B gateway.
Overall, there are about 350 end-user transactions and 1000 servers being monitored by 2000 Mercury SiteScope monitors running at 70 monitors per minute.
The BAC infrastructure is spread across two data centres - development and production. There are two production BAC servers, two for testing, and two VMWare virtual machines for development. There are business process monitors in the company's 21 core offices and distribution centres. Oracle is being used for the back end.
"It provides support staff and management visibility into application performance [and] a single view for both end-user and infrastructure performance," Tonkin said, adding the organization really sees this project as a foundation for future IT service management.
"We designed for the project, but also thought about the larger-scale implementations. Don't try to integrate everything if it doesn't make sense."
Tonkin, a 20-year veteran of the company, with more than 10 years in systems management, recommends organizations don't try and do everything at once because "Rome wasn't built in a day".
"Start small, learn, improve, and extend," he said. "When designing, think about the future [and] whatever you build should be reusable. Simple is good. Fit it in, don't rip apart. Don't rip out other infrastructure monitoring tools. You don't want to change all your infrastructure tools. Show success and it will breed more success."
Coles Myer's next direction will be to grow its monitoring deployments, and upgrade to the next release of BAC to improve the diagnostics and monitoring capability.