https://blog.turboawesome.win/series/fx-systems/ Recent content in Fx-Systems on Bits, Trades & Systems Hugo en-us Wed, 06 Jan 2016 10:31:00 +0000 https://blog.turboawesome.win/2016/01/when-the-scale-changes-moving-into-institutional-finance/ Wed, 06 Jan 2016 10:31:00 +0000 https://blog.turboawesome.win/2016/01/when-the-scale-changes-moving-into-institutional-finance/ Four years in a latency-obsessed trading environment, then a move to a large financial institution where the problems — and the constraints — were completely different. https://blog.turboawesome.win/2015/07/risk-aggregation-in-real-time-design-constraints-from-the-dealing-desk/ Wed, 08 Jul 2015 13:20:00 +0000 https://blog.turboawesome.win/2015/07/risk-aggregation-in-real-time-design-constraints-from-the-dealing-desk/ Aggregating live risk across thousands of positions in real time is a different problem than batch risk. The constraints from the dealing desk — sub-second freshness, continuous availability, correct arithmetic — shaped everything. https://blog.turboawesome.win/2015/03/building-a-trade-blotter-that-doesnt-lie-under-load/ Wed, 04 Mar 2015 10:12:00 +0000 https://blog.turboawesome.win/2015/03/building-a-trade-blotter-that-doesnt-lie-under-load/ A trade blotter shows traders their current positions and recent executions. When the system is under load, naive implementations show stale, inconsistent, or missing data. Here's the design that kept our blotter honest. https://blog.turboawesome.win/2014/12/end-of-year-architecture-review-what-held-what-failed-what-changed/ Wed, 10 Dec 2014 14:33:00 +0000 https://blog.turboawesome.win/2014/12/end-of-year-architecture-review-what-held-what-failed-what-changed/ A look back at three years of building HFT infrastructure: the components that held up under production load, the ones that were redesigned, and the architectural decisions I'd make differently now. https://blog.turboawesome.win/2014/06/slippage-spread-and-rejection-engineering-around-market-microstructure/ Wed, 25 Jun 2014 14:08:00 +0000 https://blog.turboawesome.win/2014/06/slippage-spread-and-rejection-engineering-around-market-microstructure/ The difference between a quoted price and the price you actually trade at is not a rounding error — it's a measurable, systematic cost with engineering-level causes and engineering-level mitigations. https://blog.turboawesome.win/2014/02/fx-pricing-engine-architecture-from-feeds-to-executable-quotes/ Wed, 19 Feb 2014 11:08:00 +0000 https://blog.turboawesome.win/2014/02/fx-pricing-engine-architecture-from-feeds-to-executable-quotes/ End-to-end architecture of an FX pricing engine: LP feed ingestion, aggregation, spread logic, and quote distribution. The design choices that kept us under 500µs tick-to-quote. https://blog.turboawesome.win/2013/10/order-book-implementation-data-structures-for-price-level-aggregation/ Wed, 16 Oct 2013 14:22:00 +0000 https://blog.turboawesome.win/2013/10/order-book-implementation-data-structures-for-price-level-aggregation/ An order book aggregates bids and offers at each price level. The implementation choices — sorted structures, cache behaviour, update semantics — have significant performance implications at market data rates. https://blog.turboawesome.win/2013/06/market-connectivity-building-a-low-latency-feed-handler/ Tue, 11 Jun 2013 11:02:00 +0000 https://blog.turboawesome.win/2013/06/market-connectivity-building-a-low-latency-feed-handler/ A feed handler is the entry point for all external market data. The design decisions at this boundary — NIO vs threads, kernel bypass, CPU affinity — determine the latency floor for everything downstream. https://blog.turboawesome.win/2012/08/fix-protocol-101-what-every-finance-engineer-must-know/ Tue, 07 Aug 2012 09:44:00 +0000 https://blog.turboawesome.win/2012/08/fix-protocol-101-what-every-finance-engineer-must-know/ FIX is ugly, verbose, and everywhere in electronic trading. Understanding its structure, quirks, and failure modes is unavoidable if you work in financial software. https://blog.turboawesome.win/2012/05/building-a-price-feed-aggregator-in-java-first-attempt/ Wed, 09 May 2012 14:22:00 +0000 https://blog.turboawesome.win/2012/05/building-a-price-feed-aggregator-in-java-first-attempt/ My first serious Java performance project: aggregating price feeds from multiple venues into a single best-bid-offer view. What I built, what was wrong with it, and what I learned rewriting it. https://blog.turboawesome.win/2012/03/understanding-the-fx-spot-market-from-an-engineers-chair/ Wed, 21 Mar 2012 10:44:00 +0000 https://blog.turboawesome.win/2012/03/understanding-the-fx-spot-market-from-an-engineers-chair/ What actually happens when a currency is traded: the market structure, the participants, and the data flows that an engineer building FX systems needs to understand before writing a line of code. https://blog.turboawesome.win/2012/02/from-zero-to-production-my-first-month-in-electronic-trading/ Wed, 08 Feb 2012 09:17:00 +0000 https://blog.turboawesome.win/2012/02/from-zero-to-production-my-first-month-in-electronic-trading/ What nobody tells you when you join a trading technology firm fresh out of university — the systems, the culture, and the first time you push something live.