Reliability And Availability As The Funadmental Conceps Of Network Dependability

Reliability: There are several other properties in fault tolerance in addition to dependability. One of them is reliability, and unlike dependability, which is the property of the system that we can trust it to perform its function. Reliability is something we can measure. To measure reliability, we consider the system to always be in one of these two states. The normal state is the service accomplishment state, so the system is providing the service that we want it to provide.

The other state is the service interruption. So, this is, currently, the service is not provided, it’s not accomplishing the service that we expect from it. So, Reliability can now be defined by measuring continuous service accomplishment. A typical measure for reliability is the main time to failure or MTTF. And that basically is how long do we have service accomplishment before we get into the next service interruption. So, it is not enough, for example, that over the past two years the system has been in service accomplishing for one year. If we have a system that accomplishes service for a month, then is interrupted for a month, then accomplishes service again because it has been fixed, then it’s again interrupted for a month, in that mean time to failure will be one month. That is how long you have of continuous service accomplishment, on average, between periods of service interruption. Another popular metric related to this is the availability.

Availability: Availability measures service accomplishment as a fraction of overall time. So, reliability measures how long do we get service accomplishment until the next failure. Availability measures what percentage of time were we in the service accomplishment state. So, if the system on average was half in this state and half in the other state, then availability will be 50%. 50% of the time, the system was accomplishing service. But reliability will be on average how long each of these accomplishment states lasted until we saw the interruption. So, for example, if we have one year of service followed by one year of service interruption, we get 50% here, you get one year here. If you get one month at a time for two years you get 50% here, but you get one month here instead of one year.

So, for availability, in addition to the mean time to failure, we need to know the mean time to repair, or MTTR, so this is once the system has service interruption how long does it last until it goes back to the service accomplishment state. And then our availability is the mean time to failure. How long did it function properly at the time, divided by the overall time, which was the mean time to failure plus the mean time to repair? After that, we are back in the accomplishment state.

So, this is how long we had both service accomplishment and service interruption, and this is just the service accomplishment. Virtualization is rapidly gaining attention of computational sustainability by professionals of information resources. Virtualization is improving the productivity, accuracy, live migration, availability, agile database development, flexible and cost effective, and so on. Many studies show that virtualization can improve the system availability with the innovated software techniques. However, there are some exceptions. If the application did not execute properly the plan, it can also lead to a significant issue.