Addressing system stability

[Keith Smith]

Now that I'm back from the 8 day power outage and a 6 day trip, I'm ready to tackle the uncharacteristically high number of system outages of late.

We had provisioned a virtual server in a hosted facility in advance of the arrival of Hurricane Sandy. This happened in a relative hurry and didn't permit much in the way of shopping around for different options, or quite honestly, much discussion among the Operations team regarding the most appropriate hosting solution.

In the two weeks that have passed since cutting over to the new hosing provider, it's clear that not all is well with this box. Whether it's a lack of memory, or simply the fact that VM's are not ideal for I/O intensive applications such as PilotEdge, I've seen enough to know that I don't want to see any more or wait any longer.

I considered moving back to the original configuration (running out of our office), but that is not without occasional issues as well.

To that end, we have provisioned a dedicated box at a known hosting provider. This will be ready for our use within 12-24 hours from now. We will migrate the voice and data services to that machine as soon as possible. The transition will happen outside of operating hours, and no action will be required on behalf of the pilots. Controllers will be briefed internally about the process of connecting to the new service.

I anticipate this will all take place within the next 2 days. I am positive it will fix SOME of the issues that I've been saying, and my fingers are crossed that it will resolve the FSX issue that has cropped up of late, although I'm still not positive about that.