Do you easily recognize the “five 9s” or (99.999%) meaning? Definitely easy to understand that it service availability, most of IT organization identifies its services using a percentage. Unfortunately just percentage without clear understanding of what it means, or how it could be measured.
Easily, if the service availability marked as 95% that means the customer can get the value of this service up to 95% of the total time. The rest time portion (5%) the customer may encounter some difficulties (incidents or problems) with the proposed service.
Conceptually, the availability is measured from the agreed service level against the actual service level. Whatever the measurement tool, it builds the service hierarchy. Generally, the service design itself counts on the customers’ business needs and also the service availability measured and reported based on this design. Some customers consider the “Email Service” consists of components as HW email server(s), DB, network components, messaging applications, middleware… etc. Other customers may mention that the “Email Service” consists of sub services as sending email, receiving email, access shared calendars… etc. So, the first step to design the service availability is to create the “service hierarchy deign” based on the customers’ business needs and which components are in.
Secondly; how we can monitor and measure each service’s component availability and how much the component itself weighs from the service total weight? Referring to the above sample; for the case which considers the “Email service” designed from components perspective, we need to monitor and measure the actual availability for each component (e.g. monitoring tool for HW email server; monitoring tool for DB; etc).
Thirdly; we have to set the exact weight of each monitored components, say (the HW email server weighed by 30% of the service total weight, the DB weighed by 30% of the service total weight, the middleware weighed by 30% of the service total weight and network components weighed by 10%, etc). Technically, this is the core of service availability process design as when we consider the service effected?
The fourth point is keeping the monitoring tool(s) up and running for service status data collections. After that, we will use some tools to measure the whole service availability and generate required reports based on each service component weight.
What is your best advice for the service availability design?
Author: Mohamed Zohair (All Rights Reserved by the author). Source: Original Text (based upon first hand knowledge). Image: © macniak – photodune.net Help us to improve it: how-to, discussion.