Initial Exchange Tests are Wonderfully Successful
This afternoon I was able to finally complete the initial exchange data feed tests on my new ticker plant infrastructure. The data was from a test file the exchange provides, but sent to my UDP receiver via a little test app that we have to simulate the exchange sending the data. I was not surprised to see a few mistakes in the logic - most notably in the parsing, but that's to be expected. Once I got those pointed out, the mistakes were obvious, and the fixes were trivial.
Then it just worked. Fantastic!
The speed was nice, but when you're testing on a single box, it's not really a fair test. I'll have to wait for the two test boxes that are supposed to be provided to me, and then I'll be able to make a much more real-world test. That will be interesting.
What I need to do next is to consolidate the codebase a little - the exchanges send the same data (so they say) on two independent UDP multicast channels, and we've created a dual-headed UDP receiver. We also have a single-headed one. There's too much that's similar, and so I want to consolidate them into a multi-headed UDP receiver and have it just work for from one to, say ten, inputs. The upper limit is really arbitrary, but I can't see them doing more than this anytime soon.
Anyway, it's back to the code, and then the new hardware for the full-up speed tests.