Skip to content
Deterministic execution lanes

Scale the work. Preserve the record.

Velocity organizes signed transactions into parallel execution lanes with receipts, state commitments, and a shared commitment hierarchy.

VelocityIllustrated study
SHARED RULES. SEPARATE ACTORS.
Shared rules. Separate actors.

Illustrated coordination. No live network activity.

Velocity / An illustrated overview
01 / Built for a purpose

Parallelism with something to verify.

More workers are useful only if their output can be accounted for. Velocity gives each execution lane a defined input and a committed result, then connects those results through a wider hierarchy.

01

Signed inputs, assigned lanes

Transactions carry lane and slot context. Execution starts from explicit signed work rather than a throughput counter disconnected from user actions.

02

Receipts for completed work

A lane records transaction and receipt roots along with its state delta and post-state root. Those commitments identify what was executed and what changed.

03

A chain of lane history

Lane headers reference their prior lane state and slot boundaries. Later work can be related to the history it extends.

04

A hierarchy of commitments

The architecture aggregates lane results into a broader network record. Scaling execution and confirming aggregate state remain distinct responsibilities.

05

Evidence-based performance

Historical large deployments demonstrate a measured configuration. New environments need fresh proof using the same committed-work definitions.

02 / From intent to outcome

A clear path through
velocity.

  1. 01

    Assign signed work

    Place the supported transactions into their configured lane and slot context.

  2. 02

    Execute and commit locally

    Produce the lane’s receipts, transaction root, and resulting state commitments.

  3. 03

    Aggregate the record

    Connect the lane commitments through the deployment’s shared hierarchy and inspect the supporting evidence.

The details matter

Good questions.
Clear answers.

Does Velocity remove the cost of conflicting state access?

No. Parallel execution still needs a defined access and conflict model. Workload structure is part of any performance claim.

Can I compare lane TPS directly with another chain’s headline?

Only when the workloads, signature checks, commitment definitions, hardware, topology, and measurement windows are comparable. A headline alone does not establish that comparison.

Keep exploring

One capability.
A connected platform.

01Benchmark methodology02MACA03State storage