Our press office announced our web app’s existence today. Since we’ve already got a couple thousand users registered, the announcement’s a bit of an anti-climax. Regardless, we’re planning to celebrate. To our amazement, and to our dismay, we’ve been working on this for two full years. It’s about time we close this project down.
Unfortunately, we continue to have performance problems. While I was tracking down the design flaw I discussed a few days back, one of our network contractors was looking for delays and latencies. He was able to identify one actionable bottleneck–we have a single-threaded component where we need the app to do concurrent processing. It turns out that some transactions block the channel for over a minute, which tends to clog things up. This change was already on our enhancement list, but the need wasn’t as well documented. We’ve now moved this to the top of the list, of course, and are awaiting a plan and a cost estimate from our vendor. Steve is working on it.
This won’t fix the problem, but may reduce it to something we can live with. We’ll see how it goes.