Register
Login
Resources
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Product
Data Engine LLMs Platform Enterprise
Pricing Explore
Connect to our Discord channel

TODO 2.0 KB

You have to be logged in to leave a comment. Sign In
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
  1. FIX Engine
  2. * Actually manage FIX sessions properly
  3. * Parsing FIX messages is the most CPU-intensive activity
  4. in the whole system. Probably need to rethink the current
  5. approach (separate out the FIX engine entirely?)
  6. * What is the best buffer size for reading from a socket?
  7. Small so that overall latency is low? High so as to minimize
  8. number of context switches?
  9. * Add support for all message types
  10. * Add backing store to keep a log of sent/received messages
  11. * Add support for gap fill
  12. * Support different versions of FIX?
  13. System-Wide Improvements
  14. * Actually check return codes for errors on all of the
  15. system-level calls (pthread, sockets, etc.)
  16. * Need to run more performance tests on newer hardware
  17. * Profile threading performance. The trading engine creates
  18. a lot of threads, depending on how many books are open and
  19. how many clients are connected. Are those threads contending
  20. poorly for access to shared resources? How does the kernel
  21. scheduler cope with many threads under high load?
  22. * Add unit tests
  23. Order Matching/Book
  24. * Add a transaction logging system to keep a record of
  25. filled orders
  26. * Hook into the FIX engine to send order acknowledgment and
  27. order fill reports
  28. * Add support for multicast and TCP market data feeds
  29. * Matching algorithm based on binary heaps works ok, but doesn't
  30. seem optimal. Is there any way to exploit the fact that
  31. bids and asks tend to cluster around a security's current
  32. price? Binary heaps also don't seem to work well for canceling
  33. or otherwise changing orders - direct access into the heap is
  34. slow because the heap is ordered according to the heap property.
  35. Need to augment the book with some kind of table of all open
  36. orders that would facilitate quicker direct access?
  37. * Add support for other order types
  38. Research Topics
  39. * How to price a security?
  40. * How to route orders to other exchanges?
  41. * How to fill order on other exchanges who may have a better price?
Tip!

Press p or to see the previous file or, n or to see the next file

Comments

Loading...