One of the Asia-Pacific region’s largest beverage distributors, Coca-Cola Amatil (CCA), maintains an extensive fleet of distributors and manages a complex supply chain that stretches from one side of the country to the other.
The CCA IT team has long used an array of IT monitoring tools to keep track of a diverse environment including SAP, DB2 databases and other applications on a range of server operating systems. But this technically focused monitoring had gradually come to be considered inadequate as business leaders pressed for more relevant metrics and service guarantees. They needed a way to translate enterprise server metrics into business data that could help fix the situation.
This was the challenge facing Steve de Souza and his IT team as they laid out plans for a massive overhaul of their enterprise monitoring and reporting systems. Their goal: To lay down a framework that could collect real-time performance metrics and tie them to actual business processes — thereby providing management with better information about the parts of the business running efficiently and those that were not.
Service guarantees, after all, are worth nothing if business customers can’t translate them into terms that are relevant to their requirements. To this end, CCA is now working through a four-phase project that will eventually elevate its service management and monitoring into the hands of the business leaders crying out for better information.
The first phase involved nearly 18 months of dealing with infrastructure that has now been tied into a reporting and alerting framework that uses IBM Tivoli Composite Application Manager and related components to link network performance metrics to the less black-and-white key performance indicators (KPIs) driving the business.
This was easier said than done, says de Souza: “We’re still working to discover exactly what our business wants monitored,” he told an IBM user group recently, “and whether the agents we have are sending enough information.” Relevant metrics include response time, application details, transaction response times, and Java-based agents for looking at the NetWeaver component of SAP. The third phase will see efforts to tie metrics and frameworks into an overall business service management (BSM) structure that will also involve revisiting data warehouse, user training, failover processes, and more.
“We expect service monitoring will take the most time to implement, just because of the complexity and size of our business processes,” de Souza explains.
“It’s not until we actually go for the business case that we’re going to know exactly what we need to monitor. Getting BSM running is like sticking a spoon into soup: You don’t know what’s in it.”
Progress is being made, however, and by the time the project moves into the fourth phase — a more hardware and network intensive phase — de Souza is confident CCA’s business and IT will be on better speaking terms. BSM will inform the company’s day-today operations with trend analysis, improved reporting, visibility of the warehouse management system and other infrastructure, as well as operational-related metrics such as transaction response times.