|
Key Features
A key design goal was to build a very rich functionality superset into the feed handlers themselves. Although the broader Wombat architecture has components supporting value added servers, re-publishers, internal publishers, etc, the underlying feed handler design is intended to roll as many critical functionality sets as possible into a single process (often execution thread).
These include:
- Connectivity management
- Data parsing and normalization
- Symbology mapping
- Value added data
- Full order books
- Corporate actions
- Caching
- Subscription management
- Fault tolerance
- Initial values
- Snapshot and streaming data
This approach has effectively compressed a broad cross-section of market data system functionality into a single architectural level.
A more detailed discussion of the Wombat feed handler architecture is available in the "Wombat Universal Feed Handler Suite" white paper downloadable from Technical / Whitepapers.
It's also worth reading the on-line article: Two-Level Market Data Architectures: Are They Really Necessary?
|