What this test does
An order, its acknowledgement, one client clock.
The common HP transport campaign measures scheduled NewOrderSingle → ExecutionReport response latency. The server derives the reply from the received order; scheduler delay, message construction, both endpoints and the network path are included.
Swipe the diagram horizontally, or open it at full size.
See diagrams for all four session scenarios → · See the codec operation diagrams →
View captured FIX messages for this workflow
Captured sample FIX messages
One complete, correlated chain from the synthetic QuickFIX/C++ functional audit. These are actual captured bytes illustrating the current canonical wire shape, not messages from the older transport campaign or measured latency samples.
The acknowledgement echoes order 11=1, symbol LCOM1, buy side, quantity 1000 and price 10000.00000. ExecType 150=0 and OrdStatus 39=0 mean New: CumQty is 0, LeavesQty is 1000 and AvgPx is 0. No LastQty/LastPx fill fields are present.
1. CLIENT → SERVER · NewOrderSingle 35=D
8=FIX.4.4|9=132|35=D|34=2|49=CLIENT|52=20261003-00:42:20.433|56=SERVER|
11=1|21=1|38=1000|40=2|44=10000.00000|54=1|55=LCOM1|60=20260703-12:00:00.000|
10=201| 2. SERVER → CLIENT · ExecutionReport · New acknowledgement 35=8
8=FIX.4.4|9=182|35=8|34=2|49=SERVER|52=20261003-00:42:20.433|56=CLIENT|
6=0.00000|11=1|14=0|17=TCP-EXEC|37=TCP-ORDER|38=1000|39=0|44=10000.00000|54=1|55=LCOM1|60=20260703-12:00:00.000|150=0|151=1000|
10=060| | represents SOH (ASCII 0x01). Remove display line breaks and restore SOH to recover the exact message, including valid BodyLength (9) and CheckSum (10). FIX SendingTime (52) is a wall-clock field, not the monotonic measurement clock: do not subtract these timestamps to estimate latency. Tag 60 retains the fixed July fixture time.
Download the original synthetic wire capture → · Download all checked examples and source hashes →
Explore all four workflows, measurement points and sample messages →
Qualified transport reference
Retained HP open-loop campaign, 20 September 2026; 50,000 requests/s. Newer full-session tests have not supplied a replacement passing distribution. Rows below are not a current four-flow competitor ranking.
| Transport | p50 (µs) | p99 (µs) | p99.9 (µs) | Maximum (µs) |
|---|---|---|---|---|
| Kernel TCP | 13.577 | 17.935 | 23.523 | 109.617 |
| Solarflare Onload | 5.560 | 6.044 | 6.593 | 19.537 |
| Mellanox VMA | 6.142 | 6.701 | 7.183 | 20.516 |
| Solarflare TCPDirect | 4.702 | 4.976 | 5.363 | 13.581 |
| Mellanox SocketXtreme | 5.788 | 6.620 | 7.702 | 22.113 |
Retained passing rate sweep
SCAN kernel TCP, one session, 29 September 2026. Three five-second measurement windows per target. This remains diagnostic: BrokerForge housekeeping processes ran on other cores. Failed targets have no metrics in this table; these results are not mixed into the transport CDF.
| Target /s | Achieved /s | p50 (µs) | p99 (µs) | p99.9 (µs) |
|---|---|---|---|---|
| 50,000.000 | 50,000.000 | 27.923 | 32.890 | 40.081 |
| 75,000.000 | 75,000.000 | 22.593 | 27.311 | 35.050 |
Latest codec measurements
Qualified 25-adapter converted-values-v2 campaign, 3 October 2026. The native-backed flavors have their own measured interop rows. Historical codec tables and unsupported llfix session figures are not projected here.
Current full-session admission
Review all four scenario gates and the complete C++ and C roster →