Serverless Data Mesh
Federated data mesh lakehouse publication on AWS Lambda - with cryptographic proof, not just green job logs.
An open Python framework for domain-oriented, federated data mesh teams on AWS. Implements data as a product, self-serve write infrastructure, and federated computational governance at the Iceberg lakehouse layer.
- Producer domains publish governed data products
- Steward notaries enforce mesh contracts with VRP proofs
- Publisher zones expose consumer-ready Iceberg data products
Introduces the Vaquar Pattern (PVDM): Physical → Verify → Durable → Metadata. Invariant: commit_metadata ⟹ VRP = PASS.
Paper: arXiv:2608.14643 · Reference gate: Proof-gated-publication-PVDM
Install
pip install serverless-data-mesh
serverless-data-mesh demo # <60s local PVDM demo, no AWS
serverless-data-mesh new --template medallion --output my-mesh
serverless-data-mesh apply --contract my-mesh/mesh.yaml --output my-mesh/generated
pip install "serverless-data-mesh[rules]" # + SparkRules on Lambda
pip install "serverless-data-mesh[spark]" # + PySpark + SparkRules
pip install "serverless-data-mesh[all]" # rules + spark
pip install "serverless-data-mesh[dev]" # pytest, ruff, pre-commit
Works on Windows, Mac, and Linux. Uses veridata-recon cryptographic VRP when wheels are available; pure-Python fallback otherwise (same PASS/FAIL gate).
What it solves
| Problem | Answer |
|---|---|
| Silent data loss on backfill | VRP FAIL blocks Iceberg snapshot |
| "Job succeeded" is not proof | Cryptographic multiset proof per chunk |
| Lambda 15-minute limit | Durable + IceGuard (any total runtime); optional LMI segments up to 90 min async |
| Retry duplicates data | IceGuard rollback + workload_id checkpoints |
| Central ETL bottleneck | Per-domain Lambda writer + transaction boundary |
| Federated blast radius | Producer · Steward · Publisher accounts |
Building blocks
| Package | Role |
|---|---|
| iceguard | Physical SafeWriter, timeout rollback, S3 resume |
| veridata-recon | VRP proof generation and validation |
| aws-durable-execution-sdk-python | Cross-invocation step replay |
| pyiceberg | Glue Iceberg REST metadata via GlueCatalogConnector |
| sparkrules | Optional DRL business rules ([rules] extra) |
Quick example
from serverless_data_mesh import (
IceGuardDurableCoordinator,
DomainTransactionBoundary,
VRPProofGenerator,
)
boundary = DomainTransactionBoundary(
domain_id="orders-domain",
source_namespace="raw_orders",
target_table="orders_curated",
partition_spec={"dt": "2026-06-14"},
)
coordinator = IceGuardDurableCoordinator(
durable_context=durable_ctx,
lambda_context=lambda_ctx,
proof_generator=VRPProofGenerator(),
catalog_adapter=glue_adapter,
)
outcome = coordinator.run_workload(workload)
Optional extras
| Extra | Adds |
|---|---|
rules |
SparkRules DRL on Lambda |
spark |
PySpark + SparkRules |
all |
rules + spark |
dev |
pytest, ruff, mypy, pre-commit |
publish |
build, twine |
Documentation
- GitHub README
- PVDM paper (arXiv:2608.14643)
- PVDM reference gate
- Vaquar Pattern
- Getting started
- PyPI guide
- Full documentation
License
Apache-2.0
Release files for serverless-data-mesh 1.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| serverless_data_mesh-1.3.0.tar.gz | 48.1 MB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| serverless_data_mesh-1.3.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size:48.2 MB
Release files / serverless_data_mesh-1.3.0.tar.gz
| Download URL | serverless_data_mesh-1.3.0.tar.gz |
|---|---|
| Size | 48.1 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
7558cbe7fb6cb069d6de50715112ba621fb1fd0a06271f38a52238cbc653c28c
|
|
BLAKE2b-256 checksum How to use checksums |
44f98bb95f11755e41e4aa2cd370bf1f5854a15035b535fe3c58ce313dab4337
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 19, 2026.
Transparency logRelease files / serverless_data_mesh-1.3.0-py3-none-any.whl
| Download URL | serverless_data_mesh-1.3.0-py3-none-any.whl |
|---|---|
| Size | 117.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
1c606505706718c0ed05d4a0b7ecc3ed636d12170214a6b7c94003a6e1a0a243
|
|
BLAKE2b-256 checksum How to use checksums |
06654550412f516828b59e57249a745ade8e2bb3910a33e54a672e8580870867
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 19, 2026.
Transparency log