Friday, July 8, 2011

Hidden Issues Behind 99.9% Service Level

Your IT Director is surely proud to constantly deliver 99.9 and more percentage of availability of systems and network. Though it is probably the most commonly used metrics, it is not necessarily the best metrics to use. It does not, for instance, tell you anything about the impact of a downtime or outage. What we usually would be presented are great uptime graphs like this one:


What does this tell us really?

Imagine this, due to a network or system issue, check-in is not possible for an hour. The check-in agents will have to fall back to manual check-in. In addition to cost of additional staff, this may result in flight delays. Again, what is your plan to address this? Buy more technology? What is the ROI?

What if you knew the delay cost of the affected flights and could relate this to the responsible IT department? You then could trade an investment off against the cost of downtimes.

Tuesday, July 5, 2011

We Need To Know When Something Does Not Happen

An event is anything that happens. A business event is an event that has a meaning to conduct commercial activities. Examples at the airport include a passenger checking in, a runway closure, a change in the ATC acceptance rate, an aircraft touchdown, aircraft doors closing, and a failure in the baggage sorting system. Events small and large take place all day, every day in every corner of a company and its environment.

A fundamental characteristics of events is that they cannot be entirley foreseen. Some events require actions as a direct response, for instance when the runway visibilty drops under a specific threshold. Other events are opportunities that can be exploited. For example, when a flight is cancelled, the stranded passengers could be directed to a desk which requires walking through a shopping area.



Now, an event is "something that happens" which is fine but raises the question of what to do when something which is expected to happen does not happen. For example, aircraft boarding should start 35 minutes prior to off-blocks. For this we need to know the expected boarding time based on standards, we need the latest departure time (planned or estimated) and we need to know the current status of the flight. We need to monitor this all the time. And, we must make sure that we are not getting an alert every second in the event, that boarding has not started on time. Why do we need to know this information at all? Because a late boarding is a very good indicator for a late departure.

This is just one example of time-based or timed-out events. You may want to build this and other features on top of legacy operational systems or you may want to use our brilliant and proven real-time processing system to handle this for you for all systems, every day, all the time.