A 150,000-barrel WTI physical crude purchase opportunity has just appeared at Cushing: the counterparty is offering at a $0.47/bbl discount to the current ICE benchmark. The WTI desk has approximately 30 seconds before the spread closes or the counterparty moves on. Before the trader can execute, the pre-trade system must simultaneously check open position headroom, VaR contribution, credit exposure for this counterparty, and logistics availability at Cushing. Without Redis, that check takes 1.4 seconds across 6 systems. The window is often gone.

RDI continuously syncs open positions from the ETRM, risk limit state from the risk engine, and credit exposure from the credit system — YAML pipelines, sub-second CDC lag, zero custom ETL. Kafka streams live market price ticks and position change events. The current Vitol credit exposure and Cushing logistics availability are already in Redis when this trade opportunity fires.

Redis RAM holds the live trading context — current market price, open position counters, active limit state, in-flight deal checks. Redis Flex surfaces the counterparty's full transaction history and P&L attribution. Redis Search checks the regulatory compliance context for this trade structure. All five systems' data is assembled before the first byte of the Bloomberg terminal has refreshed.

Redis RAM serves the live position counters and limit state. Redis Flex surfaces the counterparty exposure history. Redis Search checks regulatory compliance context. The pre-trade engine checks all five risk dimensions simultaneously — the same data the risk system would validate sequentially, served from Redis in a single pass.

The pre-trade engine evaluates three execution paths — execute at full size, resize to a smaller position, or pass. Position, VaR, credit, logistics, and spread-to-break-even all clear at 150,000 barrels. The spread of $0.47/bbl is 2.6x above the VaR-adjusted break-even. All five pre-trade checks return green simultaneously. EXECUTE at full size.

Without Redis, pre-trade checks across 5 systems take 1.4 seconds sequentially. At that latency, the $0.47 spread that triggered the opportunity has often moved to $0.31 or lower — or the counterparty has closed the offer. The WTI desk estimates 23% of actionable spread opportunities are missed because the pre-trade window outlasts the market window. Redis eliminates that gap.

The distributed path takes 1.4 seconds to query 5 systems sequentially — by then the spread has moved from $0.47 to $0.31, and the desk captures only $46,500 of the $70,500 opportunity. The Redis path stages all five context sources and clears all five checks in 6.8ms — the desk executes at $0.47 and captures the full spread.

The architecture is designed to operate continuously across every trading desk and every pre-trade check. RDI keeps positions, limits, credit state, and logistics availability always fresh — the pre-trade engine never makes a synchronous call to a downstream system. Redis RAM holds the hot trading context. Redis Flex holds the full deal history and counterparty exposure archive at tiered cost. The pre-trade check that used to take 1.4 seconds now takes 6.8 milliseconds — and the desk acts in the window.