Still Hammering Away at that Exchange Data
Today, in addition to a few meetings, I've been hard at work trying to work all the messages from one exchange data feed into my new ticker plant. The previous version of this project had over 300 messages - one per exchange message. I'm going for a far more minimalist design: if it's not a price or a trade, it's in a free-form variant message, and that's going to really help me in minimizing the number of messages people have to understand and deal with.
Yet there are always twists.
Today I realized that there are messages from the exchange that don't update their sequence number. OK, maybe they aren't critical to the function of the ticker plant, but I didn't want to throw them out at the UDP receiver. So I had to adjust the methods on my data source to allow for ignoring the sequence number checks. It made things a little more complex, but in the end, it'll be a better system.
That's where I'm at these days - taking what I think is a good idea and applying a real problem to it and seeing where it needs to stretch and fit. I'm hoping that I can get this all done in a few more days and then get to the real-world tests and verify that things are working as desired.
I've got my fingers crossed...