30,000 latency samples
10,000 closed-loop round trips across three runs—not a single lucky trace.
The evidence-led FIX engine
LibHFT
The shortest, most defensible path from FIX message to market.
LibHFT lets teams choose native C++, independent Java/.NET engines, or native-backed providers without surrendering the FIX contract: measured tails, protocol-correct work, and repeatable cross-engine evidence.
Language surfaces
A single FIX-critical path across native C++, Java, .NET and a benchmarked Rust adapter.
Tail evidence
Peer results and percentile markers stay tied to the stated cross-machine measurement contract.
Visible operations
The read-only Workbench makes fleet health, session state and conformance evidence inspectable locally.
$ bench_roundtrip --scenario nos_er --transport onload
NOS → ER p508.85 µscross-machine Onload
p9910.16 µstight 1.31 µs envelope
p99.911.0 µs30,000 round-trip latency samples
Fresh native-lane result: a closed-loop NewOrderSingle → ExecutionReport round trip, with fixed client/server cores. Read the benchmark record A separate open-load contract measures a 50,000-message/s aggregate; its pacing and latency evidence are reported independently.
Performance you can interrogate
Fast at the median. Controlled at the tail.
Latency is only valuable when the measurement contract is credible. LibHFT records the message path, transport, cores, samples, percentile tails and comparable work—not just a convenient headline.
CDF-STYLE · RECONSTRUCTEDNOS → ER, native Onload
lower is better
1.65× faster
Than the measured NexusFIX NOS→ER p50 in the same fresh native Onload lane.
30% lower p50
With Onload than LibHFT’s equivalent plain-TCP path on the measured contract.
THE MEASURED NATIVE LANE
Fresh cross-machine Onload NOS → ER. Same fixed-core contract; p50/p99/p99.9 in µs, lower is better.
| Engine | p50 | p99 | p99.9 | LibHFT advantage |
|---|---|---|---|---|
| LibHFT | 8.85 | 10.16 | 11.0 | Reference |
| NexusFIX | 14.61 | 15.41 | 16.28 | 1.65× |
| Fix8 | 30.4 | 31.9 | — | 3.4× |
| QuickFIX/C++ | 121.8 | 123.9 | 124.9 | 13.8× |
Architecture freedom without a FIX rewrite
Five runtime surfaces. One source-checked session contract.
Choose the implementation ownership that fits the desk: native C++, independent Java or .NET engines, or Java/.NET APIs over the C++ engine. The matrix is generated from source evidence, so the capability claim stays tied to what each surface actually implements.
5deployable surfaces
27/27session behaviours across every surface
17/17tracked typed FIX message types
| Surface | Implementation ownership | Tracked API capability | Session protocol |
|---|---|---|---|
| C++ | Native C++ reference | 73/7311 deliberate n/a | 27/27 |
| Java pure | Independent Java session and codec engine | 84/84 | 27/27 |
| Java/JNI | Java API over the C++ engine | 83/831 deliberate n/a | 27/2713 inherited from engine |
| .NET managed | Independent C# session and codec engine | 81/813 deliberate n/a | 27/27 |
| .NET native | C# API over hftnet / C++ engine | 81/813 deliberate n/a | 27/2713 inherited from engine |
Modernise without a capability downgrade
Every portable C++ capability tracked by the inventory is present in Java pure, .NET managed and .NET native—so a runtime decision does not force a thinner FIX contract.
Make ownership explicit
Independent managed engines and C++-backed providers are labelled separately. Inherited protocol behaviour is visible, never implied by a generic “supported” badge.
Build for the failure path too
The 27-behaviour protocol ledger covers logon, heartbeat, resend, gap fill, rejects, reconnect, durable sequence state and controlled recovery—not only happy-path sends.
The capability inventory is generated and checked against source in conformance runs. “n/a” is an intentionally recorded scope decision, not a hidden gap. Inspect the binding coverage matrix
Protocol breadth, without a reinvention project
Bring the QuickFIX dictionary. Choose the better engine.
LibHFT loads QuickFIX-format XML dictionaries at runtime, so the protocol asset your counterparties already speak can move with you. Standard versions are bundled, generated-table paths stay distinct from runtime loading, and the proof is kept in the source tree.
STANDARD FIX DICTIONARIESFIX 4.0 → FIX 5.0 SP2
Bundled QuickFIX-format inputs for FIX 4.0, 4.1, 4.2, 4.3, 4.4, 5.0, 5.0 SP1, 5.0 SP2 and FIXT 1.1—runtime-loadable on every LibHFT surface.
9bundled standard dictionary inputs
5runtime surfaces that load them
Generated where the codec is owned.
Template-specific FAST code generation is available in C++, pure Java and managed .NET. Java/JNI and .NET native expose the underlying FAST primitives; neither is presented as a whole-message codec or venue certification.
Codec, bus and venue evidence—not one vague label.
Generated SBE codecs, framing and UDP/multicast/mapped-memory paths are available in C++, Java and .NET. The native venue lane capture-replays CME MDP 3.0 market data; CME iLink 3/FIXP and B3 Entry Point remain explicitly scoped framing or schema slices, not completed session runners.
Fail over without resetting the conversation.
Two instances arbitrate one durable store, the standby retries with backoff, promotion restores sequence state, and the counterparty sees continuity. These active/passive building blocks span all five runtime surfaces—never dressed up as replicated or hot/hot HA.
EXCHANGE-PROTOCOL EVIDENCE
Named feeds. Independently sourced packets. Exact boundaries.
LibHFT’s native venue library does not stop at generic binary framing: its decoder and framing slices are replayed against independently sourced packet captures, so a self-consistent parser cannot certify itself.
Capture replay exercises actual packet structure alongside the implementation and marks the exact protocol/message boundary. That is stronger evidence than a hand-built fixture, while still remaining distinct from exchange certification.
SELECTED CAPTURE-REPLAYED PROTOCOL FAMILIES
Evidence boundary: the native venue inventory includes decoder/framing slices at protocol-specific scope. “Capture-replayed” is not a claim of a production session runner, counterparty onboarding, certification or venue support for every message in a family.
QUICKFIX MIGRATIONInventory. Reuse the dictionary. Verify the behaviour.
Keep protocol definitions familiar while moving the hot path, runtime ownership and latency evidence forward.
Inspect the migration assessmentProtocol coverage is generated from checked-in artefacts and conformance-checked. FAST, SBE and venue scopes are deliberately qualified above; limited capture evidence is not venue certification. Inspect the protocol breadth matrix Inspect venue evidence Inspect HA evidence
Proof, not performance theatre
Every claimed win has to do the work.
LibHFT’s benchmark harness compares codec behaviour across 19 engines in five languages. Mutation tests deliberately change business fields while preserving FIX framing, so an adapter cannot look fast by silently skipping extraction.
19engine adapters
5runtime families
55negative parity probes
0tolerance for skipped field work
The engineering advantage
Choose the engine that makes the critical path visible.
For firms where a few microseconds, a recovery edge case or an unmeasured tail can change an outcome, LibHFT turns the FIX stack into an engineering advantage. Start with the source, inspect the evidence, then design the deployment around your own hardware and market workflow.
Plan a LibHFT architecture session