Glossary · Technology & Connectivity
Latency
Also: execution latency, round-trip latency
Published
Latency — Latency is the time between an action and its effect in a trading system, for example between sending an order and receiving the execution report, or between a price change on a venue and a market maker's updated quote. It is measured in microseconds to milliseconds and is reduced by co-location, efficient protocols, and fast internal systems.
Components of Latency
Round-trip latency adds the network path to and from the counterparty, the time the counterparty's systems take to process the message, and the time the client's own systems take to act on the response. Co-location addresses the first component; protocol choice and system design address the rest. Jitter, the variation in latency, often matters as much as the average for systems that must react predictably.
Why It Matters
For a market maker, latency is the window during which its resting quotes can be picked off after the market has moved. For a desk hedging a client block across venues, it is the window during which the hedge price can drift. For a client trading a streaming price, it determines how long a quote can remain valid and how tight the provider can afford to make it.
Latency Is Not Everything
Institutional flow is rarely latency-sensitive in the microsecond sense. What matters is that the provider's latency is low enough to keep quotes tight and hedges efficient, and consistent enough that execution quality does not depend on the time of day.