End-to-End Service Monitoring

March 12, 2009 by James Smith

We are faced with the problem of knowing if the ICT architecture is delivering the services a business demands. So how can we know?
The following outlines the approach we are taking, we aim to start with two services: mobile e-mail and remote access.

An organisation provides services to customers; it must clearly state what these are so as customers can browse, match their needs, and select those that fit. When setting business strategy we focus on improvements to existing services, new services and the removal of those no longer providing benefit.

Understanding what services the business requires from ICT – presented in concepts like the ITIL service catalogue – and where these contribute to the organisational service model is essential if we are to ensure they are available, at the right quality and delivering the required function. This presents a number of challenges for ICT:

  • Listing the services required by the business
  • Defining the level of service required
  • Measuring service delivery
  • Relaying the information to the right roles

Warwickshire have adopted the ITIL best practice which uses the Service Catalogue as a method of listing what the business requires, backed up with service level agreements (SLAs) that define the quality requirements. This leaves ICT with the issues of ensuring it can measure the performance, tie it back to services and surface the information to the right people.

To tackle the issue of measuring services we take each one and identify the architecture components that contribute and linking them – using metadata in the configuration management database (CMDB) – to the service. During the process we clearly state what information must be gathered about the component in order to satisfy the business process (in this case the SLA process). Allowing us to report on the component – tell me which services you link to – or by reporting on the components required to provide the service – tell me which components make up your service.

We now have a service defined from end-to-end and understand the information that must be collected.

Now we have to report on the availability of the service. We can determine this by aggregating the status of all the components, and presenting the overall status. So if a single component is unavailable (RED) or is performing below the required quality threshold (AMBER) then the overall service status is reflected as RED or AMBER.

Furthermore we can apply this approach to other business processes and we have a much better understanding of our infrastructure and how it performs.

%d bloggers like this: