Execution Routing

The Routing Decision in 6 Microseconds

A venue scoring model trained on queue depth, spread dynamics, and rolling fill statistics picks the optimal execution path for every order. Per-order decision. No cached rankings. Under 6 microseconds.

Execution routing decision flow visualization
Routing Logic

What the Router Evaluates Per Order

Each routing decision is computed independently using current market state. No cached venue rankings. No batch scoring windows.

Queue Depth Analysis

Current bid and ask queue depth at each venue is factored into fill probability estimates. A venue with a thin book may offer a better price but lower fill certainty.

Historical Fill Rates

Rolling per-venue fill statistics at the current price level inform expected fill rate. The model distinguishes intraday regime from long-run baseline.

Spread and Impact Cost

Effective spread and estimated market impact are combined into a total cost model. Venues with wide spreads are penalized even when queue depth looks favorable.

Venue Latency Profile

Gateway round-trip latency is tracked per venue. A slightly suboptimal venue that consistently acknowledges faster than a peer will often score higher on time-sensitive strategies.

Smart Order Flow

Venue Scoring in Action

An incoming order is evaluated against all eligible venues simultaneously. Two are selected. The fill returns confirmation.

Incoming Order 10,000 shares AI Router Scores 5 venues in real time V1 V2 V3 V4 V5 Fill Confirmed ACK
Venue Coverage

Connected Exchanges and ECNs

Direct market access connections to all major US equity venues plus primary European exchanges. FIX and native binary protocols supported per venue.

US Equities

NYSE NASDAQ BATS IEX ARCA EDGX BYX PSX

US Options

CBOE BOX MIAX ISE PHLX NOM

Futures

CME ICE Eurex LSE Derivatives

FX and Fixed Income

EBS Reuters BGC Tradeweb
Performance

Routing Decision Latency Benchmarks

Independent benchmarks measuring time from normalized tick receipt to order submission. Numbers reflect p50 and p99 on a standard co-location setup at NY4 with a 10GbE direct cross-connect.

Benchmark methodology is documented in our Pulse technical blog. Numbers reflect sustained throughput under representative market open conditions, not ideal lab scenarios.

View Latency Details
Latency benchmark chart showing routing decision times
Get Started

See Routing Improvements in Your Own Environment

We run a shadow routing session alongside your current infrastructure. You compare fill statistics before switching any live order flow to EurekaLabs.