serverless-data-mesh 1.2.0__tar.gz → 1.3.0__tar.gz

{serverless_data_mesh-1.2.0 → serverless_data_mesh-1.3.0}/CHANGELOG.md

@@ -4,6 +4,20 @@ All notable changes to this project will be documented in this file.
 
 ## [Unreleased]
 
+## [1.3.0] - 2026-09-19
+
+### Added
+
+- **Lambda Managed Instances (LMI) opt-in**: segments up to **90 minutes** for async/ESM via `enable_lambda_managed_instances` + capacity provider ARN
+- **`execution_timeouts` module** with unit tests validating on-demand (≤15 min) vs LMI (≤90 min), SFN sync cap, IceGuard rollback scaling, and durable budget coherence
+- **Docs**: `docs/lambda-managed-instances.md` — industry-standard 15–90 min segments with IceGuard/VRP Iceberg safety
+
+### Changed
+
+- Terraform lambda module / prod / medallion / multi-account: timeout ceiling **900 or 5400** by capacity mode; SFN invoke wait capped at sync 15 min
+- README / architecture / terraform guides: dual clocks + LMI; IceGuard prevents corrupt Iceberg publication
+- PyPI download badges: use live **pypistats** monthly/weekly counts (shields `pypi/dm` and pe.py static badges were unreliable)
+
 ## [1.2.0] - 2026-07-21
 
 ### Added

serverless_data_mesh-1.3.0/NOTICE

@@ -0,0 +1,14 @@
+Vaquar Pattern (PVDM)
+Copyright © 2024–2026 Vaquar Khan. All rights reserved.
+
+PVDM (Physical · Verify · Durable · Metadata) and the Vaquar Pattern are
+proprietary architectural methods of Vaquar Khan. The method name, operational
+acronym, and publication invariant (commit_metadata ⟹ VRP = PASS) may not be
+rebranded or claimed as an independent invention without attribution to the
+inventor.
+
+This repository's reference implementation (serverless-data-mesh) is licensed
+under the Apache License 2.0. Use of the open-source code does not transfer
+ownership of the proprietary method.
+
+See docs/vaquar-pattern.md for the canonical method specification.

{serverless_data_mesh-1.2.0 → serverless_data_mesh-1.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.4
+Metadata-Version: 2.5
 Name: serverless-data-mesh
-Version: 1.2.0
+Version: 1.3.0
 Summary: Federated data mesh framework for proof-gated, exactly-once Iceberg data product publication on AWS Lambda.
 Project-URL: Homepage, https://github.com/vaquarkhan/aws-serverless-datamesh-framework
 Project-URL: Documentation, https://github.com/vaquarkhan/aws-serverless-datamesh-framework#readme
@@ -10,6 +10,7 @@ Project-URL: PyPI, https://pypi.org/project/serverless-data-mesh/
 Author: Serverless Data Mesh Contributors
 License-Expression: Apache-2.0
 License-File: LICENSE
+License-File: NOTICE
 Keywords: aws,bronze-silver-gold,data-as-a-product,data-mesh,domain-oriented,durable-execution,etl,federated-data-mesh,federated-governance,iceberg,iceguard,lakehouse,lambda,medallion,metadata-driven,pyiceberg,rule-engine,serverless,sparkrules,vaquar-pattern,veridata-recon
 Classifier: Development Status :: 5 - Production/Stable
 Classifier: Intended Audience :: Developers
@@ -62,6 +63,8 @@ An open Python framework for **domain-oriented**, **federated data mesh** teams
 
 Introduces the **[Vaquar Pattern](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/vaquar-pattern.md)** (PVDM): Physical → Verify → Durable → Metadata. Invariant: `commit_metadata ⟹ VRP = PASS`.
 
+**Paper:** [arXiv:2608.14643](https://arxiv.org/abs/2608.14643) · **Reference gate:** [Proof-gated-publication-PVDM](https://github.com/vaquarkhan/Proof-gated-publication-PVDM)
+
 ## Install
 
 ```bash
@@ -83,7 +86,7 @@ pip install "serverless-data-mesh[dev]"     # pytest, ruff, pre-commit
 |---------|--------|
 | 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 Execution + Step Functions (90+ min) |
+| 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 |
@@ -136,6 +139,8 @@ outcome = coordinator.run_workload(workload)
 ## Documentation
 
 - [GitHub README](https://github.com/vaquarkhan/aws-serverless-datamesh-framework#readme)
+- [PVDM paper (arXiv:2608.14643)](https://arxiv.org/abs/2608.14643)
+- [PVDM reference gate](https://github.com/vaquarkhan/Proof-gated-publication-PVDM)
 - [Vaquar Pattern](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/vaquar-pattern.md)
 - [Getting started](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/getting-started.md)
 - [PyPI guide](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/pypi.md)

{serverless_data_mesh-1.2.0 → serverless_data_mesh-1.3.0}/PYPI.md

@@ -10,6 +10,8 @@ An open Python framework for **domain-oriented**, **federated data mesh** teams
 
 Introduces the **[Vaquar Pattern](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/vaquar-pattern.md)** (PVDM): Physical → Verify → Durable → Metadata. Invariant: `commit_metadata ⟹ VRP = PASS`.
 
+**Paper:** [arXiv:2608.14643](https://arxiv.org/abs/2608.14643) · **Reference gate:** [Proof-gated-publication-PVDM](https://github.com/vaquarkhan/Proof-gated-publication-PVDM)
+
 ## Install
 
 ```bash
@@ -31,7 +33,7 @@ pip install "serverless-data-mesh[dev]"     # pytest, ruff, pre-commit
 |---------|--------|
 | 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 Execution + Step Functions (90+ min) |
+| 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 |
@@ -84,6 +86,8 @@ outcome = coordinator.run_workload(workload)
 ## Documentation
 
 - [GitHub README](https://github.com/vaquarkhan/aws-serverless-datamesh-framework#readme)
+- [PVDM paper (arXiv:2608.14643)](https://arxiv.org/abs/2608.14643)
+- [PVDM reference gate](https://github.com/vaquarkhan/Proof-gated-publication-PVDM)
 - [Vaquar Pattern](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/vaquar-pattern.md)
 - [Getting started](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/getting-started.md)
 - [PyPI guide](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/blob/main/docs/pypi.md)

{serverless_data_mesh-1.2.0 → serverless_data_mesh-1.3.0}/README.md

@@ -1,11 +1,14 @@
-<div align="center">
+<div align="center">
 
 # Serverless Data Mesh
 
 **Governed, exactly-once lakehouse writes on AWS Lambda - with cryptographic proof, not just green job logs.**
 
 [![Python 3.12+](https://img.shields.io/badge/python-3.12%2B-blue.svg)](https://www.python.org/downloads/)
-[![PyPI](https://img.shields.io/pypi/v/serverless-data-mesh.svg)](https://pypi.org/project/serverless-data-mesh/)
+[![PyPI version](https://img.shields.io/pypi/v/serverless-data-mesh.svg)](https://pypi.org/project/serverless-data-mesh/)
+[![PyPI downloads/month](https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fpypistats.org%2Fapi%2Fpackages%2Fserverless-data-mesh%2Frecent&query=%24.data.last_month&label=downloads%2Fmonth&color=blue)](https://pypistats.org/packages/serverless-data-mesh)
+[![PyPI downloads/week](https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fpypistats.org%2Fapi%2Fpackages%2Fserverless-data-mesh%2Frecent&query=%24.data.last_week&label=downloads%2Fweek&color=brightgreen)](https://pypistats.org/packages/serverless-data-mesh)
+[![Docker](https://img.shields.io/badge/ghcr.io-serverless--data--mesh-blue?logo=docker)](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/pkgs/container/serverless-data-mesh)
 [![License](https://img.shields.io/badge/license-Apache--2.0-green.svg)](LICENSE)
 [![AWS Lambda](https://img.shields.io/badge/AWS-Lambda%20%2B%20Durable%20Execution-orange.svg)](https://docs.aws.amazon.com/lambda/)
 [![Iceberg](https://img.shields.io/badge/Apache-Iceberg-00A4EF.svg)](https://iceberg.apache.org/)
@@ -16,17 +19,17 @@
 
 An open Python framework for **federated data mesh** lakehouse publication on AWS:<br/>
 **domain-oriented ownership**, **data as a product**, and **self-serve write infrastructure** for cross-domain teams.<br/>
-**Producer** domains publish governed **data products** · **Steward** notaries enforce **federated computational governance** · **Publisher** zones expose consumer-ready **Iceberg data products** to the mesh.
+**Producer** domains publish governed **data products** - **Steward** notaries enforce **federated computational governance** - **Publisher** zones expose consumer-ready **Iceberg data products** to the mesh.
 
-[**PyPI**](https://pypi.org/project/serverless-data-mesh/) · [**Walkthrough (in README)**](#how-to-start--walkthrough-gifs--benefits) · [**GIF gallery**](docs/visual-tutorial.md) · [**Durable compute**](examples/durable-compute/) · [**Vaquar Pattern**](docs/vaquar-pattern.md) · [**Deploy**](infrastructure/terraform/README.md)
+[**PyPI**](https://pypi.org/project/serverless-data-mesh/) · [**Docker (GHCR)**](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/pkgs/container/serverless-data-mesh) · [**Walkthrough**](#how-to-start---walkthrough-gifs--benefits) · [**GIF gallery**](docs/visual-tutorial.md) · [**Vaquar Pattern**](docs/vaquar-pattern.md) · [**Paper (arXiv)**](https://arxiv.org/abs/2608.14643) · [**Deploy**](docs/aws-production-deploy.md)
 
 </div>
 
 ---
 
-## How to start — walkthrough (GIFs + benefits)
+## How to start - walkthrough (GIFs + benefits)
 
-Read this section in the README — no extra docs required.  
+Read this section in the README - no extra docs required.  
 **Interactive auto-play demo:** after `ui --open`, open [http://127.0.0.1:8765/walkthrough](http://127.0.0.1:8765/walkthrough) · or open [docs/demo-walkthrough.html](docs/demo-walkthrough.html)
 
 **Also useful:**
@@ -44,7 +47,7 @@ Read this section in the README — no extra docs required.
 ```bash
 pip install serverless-data-mesh
 
-# Optional: 60s proof (no AWS) — clean commit, corrupt blocked
+# Optional: 60s proof (no AWS)  -  clean commit, corrupt blocked
 serverless-data-mesh demo
 
 # Generate sample mesh + open control center
@@ -70,7 +73,7 @@ serverless-data-mesh ui --path examples/medallion-e2e/generated --open
   <img src="docs/images/tutorial/tutorial-overview.gif" alt="Step-by-step demo overview GIF" width="720" />
 </p>
 
-### Step 1 — Install & prove the gate
+### Step 1  -  Install & prove the gate
 
 <p align="center">
   <img src="docs/images/tutorial/step-01-install-demo.gif" alt="Step 1: install and demo" width="720" />
@@ -84,10 +87,10 @@ serverless-data-mesh demo
 | | |
 |--|--|
 | **You do** | Install, then run `demo` once |
-| **You get** | Clean write **commits**; corrupt write is **blocked** — in &lt;60s, **no AWS** |
+| **You get** | Clean write **commits**; corrupt write is **blocked**  -  in &lt;60s, **no AWS** |
 | **Benefit** | Proof that “pipeline succeeded” ≠ trustworthy data |
 
-### Step 2 — Create mesh YAML
+### Step 2  -  Create mesh YAML
 
 <p align="center">
   <img src="docs/images/tutorial/step-02-new-mesh.gif" alt="Step 2: create mesh YAML" width="720" />
@@ -102,9 +105,9 @@ serverless-data-mesh new --template medallion --output my-mesh
 |--|--|
 | **You do** | Scaffold (or edit) a medallion / northstar contract |
 | **You get** | A domain-owned `mesh.yaml` |
-| **Benefit** | Domain teams own the contract — not a central Glue ticket queue |
+| **Benefit** | Domain teams own the contract  -  not a central Glue ticket queue |
 
-### Step 3 — Apply (compile pipelines)
+### Step 3  -  Apply (compile pipelines)
 
 <p align="center">
   <img src="docs/images/tutorial/step-03-apply.gif" alt="Step 3: apply and compile" width="720" />
@@ -122,7 +125,7 @@ serverless-data-mesh apply \
 | **You get** | Handlers, Step Functions ASL, VRP config, layer Lambda manifest |
 | **Benefit** | One YAML → many proof-gated pipelines (bronze / silver / gold) |
 
-### Step 4 — Start the control center UI
+### Step 4  -  Start the control center UI
 
 <p align="center">
   <img src="docs/images/tutorial/step-04-ui.gif" alt="Step 4: open control UI" width="720" />
@@ -140,7 +143,7 @@ serverless-data-mesh ui --path my-mesh/generated --open
 | **You get** | KPIs, pipelines, trust, PVDM, durable clocks, Tutorial tab |
 | **Benefit** | Review readiness **before** you spend AWS money |
 
-### Step 5 — Deploy Durable Lambda (AWS)
+### Step 5  -  Deploy Durable Lambda (AWS)
 
 <p align="center">
   <img src="docs/images/tutorial/step-05-deploy.gif" alt="Step 5: package and deploy" width="720" />
@@ -149,17 +152,17 @@ serverless-data-mesh ui --path my-mesh/generated --open
 ```bash
 ./infrastructure/terraform/scripts/package_lambda.sh
 cd infrastructure/terraform/environments/prod
-# dual clocks — see examples/durable-compute/terraform.tfvars.example
+# dual clocks  -  see examples/durable-compute/terraform.tfvars.example
 terraform apply
 ```
 
 | | |
 |--|--|
 | **You do** | Package the Lambda zip + `terraform apply` ([durable-compute example](examples/durable-compute/)) |
-| **You get** | Durable Lambda on Firecracker, on-demand scale-to-zero, dual clocks (e.g. 900s × 5400s) |
-| **Benefit** | 90+ minute backfills without idle EMR/Glue clusters |
+| **You get** | Durable Lambda on Firecracker, on-demand scale-to-zero, dual clocks (segment + workload) |
+| **Benefit** | Backfills longer than 15 minutes without idle EMR/Glue clusters |
 
-### Step 6 — Observe & attest
+### Step 6  -  Observe & attest
 
 <p align="center">
   <img src="docs/images/tutorial/step-06-observe.gif" alt="Step 6: observe and attest" width="720" />
@@ -174,7 +177,7 @@ serverless-data-mesh dashboard --open
 |--|--|
 | **You do** | Run attestation demo and/or trust dashboard |
 | **You get** | VRP proofs + PVDM-A decision records + CloudWatch trust metrics |
-| **Benefit** | Auditable publication — consumers only see snapshots after **VRP PASS** |
+| **Benefit** | Auditable publication  -  consumers only see snapshots after **VRP PASS** |
 
 ### Quick links (same as the walkthrough footer)
 
@@ -182,20 +185,32 @@ serverless-data-mesh dashboard --open
 - **Control center:** `serverless-data-mesh ui --path examples/medallion-e2e/generated --open`
 - **Durable Lambda clocks:** [examples/durable-compute/](examples/durable-compute/)
 
-### Docker (control center image)
+### Docker (GHCR - published image)
+
+Official image (release `v1.3.0` publishes via GitHub Actions to GHCR):
 
 ```bash
-# Pull (after publish)
-docker pull ghcr.io/vaquarkhan/serverless-data-mesh:1.2.0
-docker run --rm -p 8765:8765 ghcr.io/vaquarkhan/serverless-data-mesh:1.2.0
-# → http://127.0.0.1:8765/
+docker pull ghcr.io/vaquarkhan/serverless-data-mesh:1.3.0
+# or: ghcr.io/vaquarkhan/serverless-data-mesh:latest
+docker run --rm -p 8765:8765 ghcr.io/vaquarkhan/serverless-data-mesh:1.3.0
+# open http://127.0.0.1:8765/
+```
+
+Package page: [ghcr.io/vaquarkhan/serverless-data-mesh](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/pkgs/container/serverless-data-mesh)
 
-# Or build locally
+```bash
+# Build locally instead
 docker build -t serverless-data-mesh:local .
 docker run --rm -p 8765:8765 serverless-data-mesh:local
 ```
 
-**PyPI:** `pip install serverless-data-mesh==1.2.0`
+### Install & download stats
+
+| Channel | Command / link |
+|---------|----------------|
+| **PyPI** | `pip install serverless-data-mesh==1.3.0` · [project](https://pypi.org/project/serverless-data-mesh/) · [monthly stats](https://pypistats.org/packages/serverless-data-mesh) · [total downloads](https://pepy.tech/projects/serverless-data-mesh) |
+| **Docker** | `docker pull ghcr.io/vaquarkhan/serverless-data-mesh:latest` |
+| **GitHub Release** | [v1.3.0](https://github.com/vaquarkhan/aws-serverless-datamesh-framework/releases/tag/v1.3.0) |
 
 ---
 
@@ -234,7 +249,7 @@ Most data mesh programs reorganize teams but leave the **write path** centralize
 |------|-------------------|-------------------------------|
 | **Silent data loss** | Partition row counts drift; discovered days later | VRP `FAIL` blocks Iceberg snapshot; consumers never see bad data |
 | **"Job succeeded" ≠ correct** | Glue exit code 0 with 6 missing rows | Multiset cryptographic proof per chunk |
-| **Lambda 15-min limit** | "Use EMR/Glue for real backfills" | Durable Execution + Step Functions resume → **90+ minutes** |
+| **Lambda 15-min limit** | "Use EMR/Glue for real backfills" | Durable + IceGuard (any total runtime); optional LMI segments up to **90 min** async |
 | **Retry duplicates data** | Re-invoke creates duplicate Parquet | IceGuard rollback + `workload_id` checkpoints |
 | **Platform bottleneck** | Every domain waits on central ETL | Each domain owns a Lambda writer + declared contract |
 | **No audit evidence** | Sample rows and debate | Immutable VRP proofs in Steward S3; offline `verify_proof` |
@@ -247,12 +262,28 @@ Most data mesh programs reorganize teams but leave the **write path** centralize
 
 ## The solution: Vaquar Pattern
 
-This framework is the **reference implementation** of the **[Vaquar Pattern](docs/vaquar-pattern.md)** — a proprietary architectural method invented by **Vaquar Khan** for proof-gated serverless lakehouse publication.
+This framework is the **reference implementation** of the **[Vaquar Pattern](docs/vaquar-pattern.md)**  -  a proprietary architectural method invented by **Vaquar Khan** for proof-gated serverless lakehouse publication.
+
+<p align="center">
+  <img src="docs/images/pvdm-lifecycle-readme.png" alt="PVDM lifecycle: Physical, Verify, Durable, Metadata with fail-closed VRP FAIL branch" width="920" />
+</p>
+
+<p align="center">
+  <img src="docs/images/pvdm-fail-closed-vs-traditional.png" alt="Traditional exit-code publish vs PVDM fail-closed gate: corrupt data never reaches consumers" width="920" />
+</p>
 
 <p align="center">
-  <img src="docs/images/vaquar-pattern-proprietary-pvdm.png" alt="The Vaquar Pattern PVDM: proprietary method by Vaquar Khan — Physical Verify Durable Metadata" width="920" />
+  <img src="docs/images/vaquar-pattern-proprietary-pvdm.png" alt="The Vaquar Pattern PVDM: proprietary method by Vaquar Khan  -  Physical Verify Durable Metadata" width="920" />
 </p>
 
+**Read the paper & reference gate**
+
+| | |
+|--|--|
+| **arXiv preprint** | [arXiv:2608.14643](https://arxiv.org/abs/2608.14643) — *Proof-Gated Publication: Verify-Before-Commit Content Integrity for Serverless Data-Mesh Lakehouses* |
+| **Reference gate + adversarial suite** | [github.com/vaquarkhan/Proof-gated-publication-PVDM](https://github.com/vaquarkhan/Proof-gated-publication-PVDM) (stdlib gate, 30/30 suite, Spark/Iceberg benchmarks) |
+| **This repo** | Production AWS mapping (IceGuard · veridata-recon · Durable SDK · Glue/Iceberg) |
+
 > **Proof-Gated Serverless Lakehouse Publication (PVDM)**  
 > Physical → Verify → Durable → Metadata  
 > Invariant: `commit_metadata ⟹ VRP = PASS`
@@ -262,31 +293,38 @@ This framework is the **reference implementation** of the **[Vaquar Pattern](doc
 | **Method name** | Vaquar Pattern |
 | **Operational acronym** | **PVDM** (Physical · Verify · Durable · Metadata) |
 | **Inventor** | Vaquar Khan |
+| **Copyright** | © 2024–2026 Vaquar Khan  -  proprietary method (name + invariants) |
 | **Status** | Proprietary method · open reference implementation (Apache-2.0) |
-| **Cite** | [docs/vaquar-pattern.md](docs/vaquar-pattern.md) |
+| **Cite** | [arXiv:2608.14643](https://arxiv.org/abs/2608.14643) · [docs/vaquar-pattern.md](docs/vaquar-pattern.md) · [NOTICE](NOTICE) |
 
 <p align="center">
   <img src="docs/images/why-sdm-four-phase-connectivity.png" alt="Vaquar Pattern four phases: Physical, Verify, Durable, Metadata with cross-account flows" width="920" />
 </p>
 
+<p align="center">
+  <img src="docs/images/vaquar-pattern-invariant.png" alt="Vaquar Pattern invariant: metadata commit only when VRP equals PASS" width="920" />
+</p>
+
 | Phase | Component | Outcome | Implemented? |
 |-------|-----------|---------|--------------|
-| **Physical** | IceGuard SafeWriter | Parquet in Publisher S3; checkpoints in Steward S3 | **Yes** — `IceGuardDurableCoordinator` + IceGuard watchdog |
-| **Verify** | veridata-recon | VRP proof stored; `validate_then_commit` gate | **Yes** — PASS blocks FAIL before metadata |
-| **Durable** | AWS Durable SDK + Step Functions | 15-min segments → 90+ min workloads | **Yes** — `@durable_execution` + SFN resume |
-| **Metadata** | GlueCatalogConnector | Iceberg snapshot **only after** VRP PASS | **Yes** — proof-gated commit |
+| **Physical** | IceGuard SafeWriter | Parquet in Publisher S3; checkpoints in Steward S3 | **Yes**  -  `IceGuardDurableCoordinator` + IceGuard watchdog |
+| **Verify** | veridata-recon + keyed Steward binding | VRP proof stored; `validate_then_commit` + Metadata TOCTOU gate | **Yes**  -  PASS blocks FAIL before metadata |
+| **Durable** | AWS Durable SDK + Step Functions | 15-min segments chained to any configured total | **Yes**  -  `@durable_execution` + SFN resume |
+| **Metadata** | GlueCatalogConnector | Iceberg snapshot **only after** VRP PASS | **Yes**  -  proof-gated commit |
 
 Optional (not a fifth PVDM phase): SparkRules DRL / rules gate before Physical; PVDM-A decision attestation after Verify.
 
 What makes this new vs Outbox, Saga, Medallion, and Glue bookmarks: **Iceberg publication is gated on cryptographic multiset proof**, not executor success.
 
+**See it live:** [How to start (GIF walkthrough)](#how-to-start---walkthrough-gifs--benefits) · `serverless-data-mesh demo` · UI `/walkthrough`  
+**Paper:** [arXiv:2608.14643](https://arxiv.org/abs/2608.14643) · **Reference gate:** [Proof-gated-publication-PVDM](https://github.com/vaquarkhan/Proof-gated-publication-PVDM)  
 **Pattern spec:** [docs/vaquar-pattern.md](docs/vaquar-pattern.md) · **Full blog:** [docs/blog-the-vaquar-pattern.md](docs/blog-the-vaquar-pattern.md) · **Compute model:** [examples/durable-compute/](examples/durable-compute/)
 
 ---
 
 ## Durable Lambda compute model
 
-**No idle EMR/Glue clusters. No EC2 on-demand fleet.** Domain writers run on **AWS Lambda** with **Durable Execution**, Firecracker isolation, and **fully configurable dual clocks**.
+**No idle EMR/Glue clusters.** Domain writers run on **AWS Lambda** with **Durable Execution**, Firecracker isolation, **15–90 minute segments** (on-demand or [Managed Instances](https://aws.amazon.com/blogs/compute/announcing-90-minute-function-timeout-on-aws-lambda-managed-instances/)), and IceGuard so incomplete writes never corrupt Iceberg.
 
 <p align="center">
   <img src="docs/images/durable-lambda-compute-model.png" alt="Durable Lambda compute: Firecracker microVM, on-demand Lambda, dual clocks for per-invoke and durable budget" width="920" />
@@ -296,36 +334,42 @@ What makes this new vs Outbox, Saga, Medallion, and Glue bookmarks: **Iceberg pu
 |------------|-------------|-------------------------|
 | **Durable Lambda** | `@durable_execution` + Terraform `durable_config` | Glue ETL as the write path |
 | **MicroVM isolation** | AWS Lambda on **Firecracker** microVMs (AWS-managed) | Custom Firecracker / self-managed microVMs |
-| **On-demand compute** | **On-demand Lambda** scaling (scale to zero) | EC2 on-demand / ASG fleets; provisioned concurrency (optional later) |
-| **Configurable time** | Per-invoke timeout **and** durable total budget | Fixed 15-min wall clock for the whole backfill |
+| **On-demand compute** | **On-demand Lambda** (default, scale to zero) | Unmanaged EC2 / ASG fleets |
+| **Managed Instances (opt-in)** | LMI capacity provider + timeout up to **90 min** async | Claiming sync invokes exceed 15 min |
+| **Iceberg safety** | IceGuard rollback + VRP before metadata | “Job SUCCEEDED” as proof of integrity |
 
 ### Dual clocks (Terraform-tunable)
 
 | Knob | tfvars / variable | Meaning | Limits |
 |------|-------------------|---------|--------|
-| **Container clock** | `lambda_timeout_seconds` | One Lambda invocation | 1–**900** s (AWS hard max) |
-| **Workload clock** | `durable_execution_timeout_seconds` | Total durable budget across replays | Default **5400** s (90 min), up to ~1 year |
+| **Container clock** | `lambda_timeout_seconds` | One Lambda invocation | **15 min** on-demand · **up to 90 min** with LMI (async/ESM) |
+| **Capacity mode** | `enable_lambda_managed_instances` | Attach LMI capacity provider | Opt-in; needs capacity provider ARN |
+| **Workload clock** | `durable_execution_timeout_seconds` | Total durable budget across segments | **Any duration** you set (AWS up to ~1 year) |
 | **IceGuard lead time** | `iceguard_rollback_threshold_ms` | Rollback before hard kill | Auto or explicit ms |
-| **SFN resume** | `max_resume_attempts`, `resume_wait_seconds` | Re-invoke after `rolled_back` | Tunable |
+| **SFN resume** | `max_resume_attempts`, `resume_wait_seconds` | Re-invoke after `rolled_back` | Sync SFN invoke still ≤15 min |
 
 ```hcl
 # infrastructure/terraform/environments/prod/terraform.tfvars (example)
 enable_durable_execution           = true
-lambda_timeout_seconds             = 900    # 15 min per segment
-durable_execution_timeout_seconds  = 5400   # 90 min total durable budget
+lambda_timeout_seconds             = 900      # 15 min on-demand (default)
+# Optional industry-standard longer segments (async/ESM on LMI):
+# enable_lambda_managed_instances = true
+# lambda_managed_instances_capacity_provider_arn = "arn:aws:lambda:..."
+# lambda_timeout_seconds          = 5400      # up to 90 min
+durable_execution_timeout_seconds  = 10800    # set to your backfill wall-clock
 lambda_memory_mb                   = 4096
-iceguard_rollback_threshold_ms     = 30000  # yield 30s before timeout
+iceguard_rollback_threshold_ms     = 30000    # yield before hard timeout → no corrupt Iceberg
 ```
 
-Long backfills = **many durable segments** (each ≤ `lambda_timeout_seconds`), not one forever-running process. IceGuard rolls back incomplete chunks; Durable SDK replays completed steps; Step Functions resumes on `rolled_back`.
+Long backfills = **many durable segments** (each ≤ `lambda_timeout_seconds`), not one forever-running process. IceGuard rolls back incomplete chunks; Durable SDK replays completed steps; Step Functions resumes on `rolled_back`; metadata commits only after VRP PASS.
 
-→ [Durable compute example](examples/durable-compute/README.md) · [Architecture](docs/architecture.md#durable-lambda-compute-model)
+→ [Durable compute example](examples/durable-compute/README.md) · [LMI 15–90 min guide](docs/lambda-managed-instances.md) · [Architecture](docs/architecture.md#durable-lambda-compute-model)
 
 ---
 
 ## Verification & business rules
 
-The Vaquar Pattern’s **Verify** phase is powered by [veridata-recon](https://pypi.org/project/veridata-recon/) on PyPI. Optional [SparkRules](https://pypi.org/project/sparkrules/) DRL runs **before** physical writes and VRP — on Lambda, not Glue ETL.
+The Vaquar Pattern’s **Verify** phase is powered by [veridata-recon](https://pypi.org/project/veridata-recon/) on PyPI. Optional [SparkRules](https://pypi.org/project/sparkrules/) DRL runs **before** physical writes and VRP  -  on Lambda, not Glue ETL.
 
 ### veridata-recon: VRP per chunk
 
@@ -348,7 +392,7 @@ proof = VRPProofGenerator().generate(source_rows, sink_rows, identity_fields=["p
 validate_then_commit(proof)  # raises if verdict != PASS
 ```
 
-**Attacks blocked:** drop, duplicate, mutation, schema drift — see the [consumer safety benchmark](eval/validate_then_commit_benchmark.py) (`make benchmark`).
+**Attacks blocked:** drop, duplicate, mutation, schema drift  -  see the [consumer safety benchmark](eval/validate_then_commit_benchmark.py) (`make benchmark`).
 
 → [Vaquar Pattern invariant (image)](docs/blog-the-vaquar-pattern.md#4-the-vaquar-invariant) · [Observability: VRP metrics & S3 proofs](docs/observability-production.md)
 
@@ -362,12 +406,12 @@ Domain teams declare rules in YAML (`spark_rules_enabled: true`) or load Steward
 
 | Mode | Install | When to use |
 |------|---------|-------------|
-| **Pure Python** | `pip install "serverless-data-mesh[rules]"` | Default — sub-ms per fact, small Lambda zip |
+| **Pure Python** | `pip install "serverless-data-mesh[rules]"` | Default  -  sub-ms per fact, small Lambda zip |
 | **PySpark on Lambda** | `pip install "serverless-data-mesh[spark]"` | Large partitions, same DRL at scale |
 | **Steward governance** | `SPARKRULES_DRL_S3_URI=s3://steward-rules/...` | Central rule packs + `RuleFireSummary` audit lineage |
 
 ```yaml
-# mesh.yaml excerpt — compiler emits SparkRules hook in readers.py
+# mesh.yaml excerpt  -  compiler emits SparkRules hook in readers.py
 runtime:
   engine: pyarrow
   package_extras: rules
@@ -409,7 +453,7 @@ enriched, audit = connector.apply_chunk(source_records)
 <td width="50%" valign="top">
 
 ### Serverless at scale
-- 15-min Lambda segments → 90+ min backfills
+- 15-min Lambda segments chained to a configurable total runtime (overcomes the AWS per-invoke limit)
 - IceGuard watchdog rollback before hard timeout
 - Durable step replay: no duplicate committed chunks
 - Terraform-tunable `lambda_timeout_seconds`
@@ -448,7 +492,7 @@ enriched, audit = connector.apply_chunk(source_records)
 
 ## Create pipelines from YAML
 
-**Metadata-driven pipeline creation** — define your mesh in YAML; the compiler generates proof-gated Lambda pipelines, Step Functions orchestrators, VRP config, consumer SLAs, and Terraform manifests. Domain teams only implement `readers.py` (source/sink I/O).
+**Metadata-driven pipeline creation**  -  define your mesh in YAML; the compiler generates proof-gated Lambda pipelines, Step Functions orchestrators, VRP config, consumer SLAs, and Terraform manifests. Domain teams only implement `readers.py` (source/sink I/O).
 
 <p align="center">
   <img src="docs/images/pipeline-creation-flow.png" alt="Metadata-driven pipeline creation: Write YAML, compile with serverless-data-mesh apply, deploy to AWS" width="920" />
@@ -509,7 +553,7 @@ serverless-data-mesh compile \
   --output domains/
 ```
 
-→ [Metadata-driven pipeline guide](docs/metadata-driven-pipeline.md) · [5-min AWS deploy](docs/first-mesh-on-aws.md) · [Medallion E2E example](examples/medallion-e2e/README.md)
+→ [Metadata-driven pipeline guide](docs/metadata-driven-pipeline.md) · [5-min AWS deploy](docs/first-mesh-on-aws.md) · [Medallion example](examples/medallion-e2e/README.md)
 
 ---
 
@@ -556,7 +600,7 @@ serverless-data-mesh compile \
 Event / Step Functions
         │
         ▼
-Lambda :live  (15-min segments, up to 90+ min total)
+Lambda :live  (15-min segments; total budget = durable_execution_timeout_seconds)
         │
         ├── Read source        → domain-specific reader
         ├── SparkRules         → optional DRL filter
@@ -579,7 +623,7 @@ Lambda :live  (15-min segments, up to 90+ min total)
 
 ### Try in 60 seconds (no AWS)
 
-Works on **Windows, Mac, and Linux** — pure-Python verifier fallback when Rust wheels unavailable.
+Works on **Windows, Mac, and Linux**  -  pure-Python verifier fallback when Rust wheels unavailable.
 
 ```bash
 pip install serverless-data-mesh
@@ -588,12 +632,12 @@ serverless-data-mesh demo
 
 ### Install from PyPI
 
-**Package:** [`serverless-data-mesh`](https://pypi.org/project/serverless-data-mesh/) · **Requires Python 3.12+** · **Latest:** [1.2.0](https://pypi.org/project/serverless-data-mesh/1.2.0/)
+**Package:** [`serverless-data-mesh`](https://pypi.org/project/serverless-data-mesh/) · **Requires Python 3.12+** · **Latest:** [1.3.0](https://pypi.org/project/serverless-data-mesh/1.3.0/)
 
 | | Link |
 |---|------|
 | **PyPI project** | https://pypi.org/project/serverless-data-mesh/ |
-| **Download files (1.2.0)** | https://pypi.org/project/serverless-data-mesh/1.2.0/#files |
+| **Download files (1.3.0)** | https://pypi.org/project/serverless-data-mesh/1.3.0/#files |
 | **5-min AWS guide** | [docs/first-mesh-on-aws.md](docs/first-mesh-on-aws.md) |
 
 **Recommended install (pip):**
@@ -603,7 +647,7 @@ serverless-data-mesh demo
 pip install serverless-data-mesh
 
 # Pin a version (recommended for production)
-pip install serverless-data-mesh==1.2.0
+pip install serverless-data-mesh==1.3.0
 
 # Upgrade
 pip install -U serverless-data-mesh
@@ -732,18 +776,24 @@ outcome = coordinator.run_workload(workload)
 
 ## Deploy to AWS
 
+**Prod-grade checklist (SNS confirm + alarms + smoke):** [docs/aws-production-deploy.md](docs/aws-production-deploy.md)
+
 ```bash
 ./infrastructure/terraform/scripts/package_lambda.sh
 cd infrastructure/terraform/environments/prod
+cp terraform.tfvars.example terraform.tfvars
+# set unique buckets + ops_alert_emails, then:
 terraform init && terraform apply
+# Confirm the SNS subscription email before go-live
 ```
 
 | Path | Use when |
 |------|----------|
-| **[Terraform (production)](infrastructure/terraform/README.md)** | Step Functions, DLQ, monitoring, configurable timeouts |
+| **[AWS production deploy](docs/aws-production-deploy.md)** | SNS, SFN alarms, DLQ IAM, smoke scripts, runbook |
+| **[Terraform (production)](infrastructure/terraform/README.md)** | Step Functions, DLQ, SNS, monitoring, dual clocks |
 | [Multi-account mesh](infrastructure/terraform/environments/multi-account/README.md) | Producer / Steward / Publisher across AWS accounts |
 | [SAM alternative](infrastructure/sam/README.md) | SAM-native teams |
-| [Deployment guide](docs/deployment.md) | Manual Lambda / IAM setup |
+| [Deployment guide](docs/deployment.md) | Env vars / IAM summary |
 
 → [Terraform step-by-step](docs/terraform-guide.md)
 
@@ -754,19 +804,22 @@ terraform init && terraform apply
 | Document | What you will learn |
 |----------|---------------------|
 | **[Metadata-driven pipelines](docs/metadata-driven-pipeline.md)** | **Complete guide: YAML schema, bronze/silver/gold, compile, deploy** |
+| **[Vaquar Pattern (proprietary method)](docs/vaquar-pattern.md)** | Formal PVDM spec  -  cite this; inventor attribution |
+| **[PVDM paper (arXiv:2608.14643)](https://arxiv.org/abs/2608.14643)** | Proof-gated publication preprint (methods + evaluation) |
+| **[PVDM reference gate](https://github.com/vaquarkhan/Proof-gated-publication-PVDM)** | Stdlib gate + 30-case adversarial suite + Spark benchmarks |
 | **[Visual tutorial (GIFs)](docs/visual-tutorial.md)** | Step-by-step animated walkthrough + control UI Tutorial tab |
-| **[Vaquar Pattern (proprietary method)](docs/vaquar-pattern.md)** | Formal PVDM spec — cite this; inventor attribution |
-| **[Durable compute example](examples/durable-compute/README.md)** | Lambda dual clocks, Firecracker, on-demand, tfvars |
+| **[Durable compute example](examples/durable-compute/README.md)** | 15–90 min segments, Durable, IceGuard, tfvars |
+| **[Lambda Managed Instances](docs/lambda-managed-instances.md)** | Opt-in LMI up to 90 min async; Iceberg safety |
 | **[Observability (production)](docs/observability-production.md)** | Structured logs, VRP S3 proofs, CloudWatch dashboard, DLQ smoke tests |
-| **[Medallion E2E example](examples/medallion-e2e/README.md)** | One YAML → 6 pipelines + orchestrators |
+| **[Medallion example](examples/medallion-e2e/README.md)** | One YAML → 6 pipelines + orchestrators |
 | **[Retail flat ETL example](examples/retail-mesh/README.md)** | 5 domain pipelines, PySpark on Lambda |
-| **[Vaquar Pattern blog](docs/blog-the-vaquar-pattern.md)** | **Full article: images, E2E journey, adoption playbook** |
+| **[Vaquar Pattern blog](docs/blog-the-vaquar-pattern.md)** | Full article: images, journey, adoption playbook |
 | **[Vaquar Pattern spec](docs/vaquar-pattern.md)** | Formal spec - cite this in architecture docs |
 | **[Why Serverless Data Mesh](docs/why-serverless-data-mesh.md)** | Blog: industry problem, connectivity, portfolio stack |
 | **[Data mesh patterns](docs/data-mesh-patterns.md)** | 13 named patterns + concept coverage matrix |
-| **[End-to-end guide](docs/data-mesh-end-to-end.md)** | Three-account journey, IAM, deploy order |
+| **[Multi-account deploy guide](docs/data-mesh-end-to-end.md)** | Three-account journey, IAM, deploy order |
 | [Getting started](docs/getting-started.md) | Hands-on tutorial for domain engineers |
-| [Architecture](docs/architecture.md) | Components, failure modes, 90-min execution |
+| [Architecture](docs/architecture.md) | Components, failure modes, configurable durable execution |
 | **[Due diligence: claims vs repo](docs/due-diligence-architecture-claims.md)** | What is shipped vs overstated vs roadmap (PVDM-A / MCP) |
 | [PVDM-A / Agentic roadmap](docs/roadmap-pvdm-a-agentic.md) | Decision attestation + MCP plan (not shipped yet) |
 | [Glue connector](docs/glue-connector.md) | Lambda + Spark vs Glue ETL |
@@ -782,10 +835,11 @@ terraform init && terraform apply
 serverless-data-mesh/
 ├── docs/
 │   ├── metadata-driven-pipeline.md # Complete YAML → pipeline guide (medallion)
-│   ├── blog-the-vaquar-pattern.md  # Full Vaquar Pattern blog (images, E2E)
+│   ├── blog-the-vaquar-pattern.md  # Full Vaquar Pattern blog (images)
 │   ├── vaquar-pattern.md           # Formal pattern spec (cite this)
 │   ├── why-serverless-data-mesh.md # Blog article with diagrams
 │   ├── data-mesh-end-to-end.md     # Three-account deploy guide
+│   ├── aws-production-deploy.md    # SNS, alarms, smoke, runbook
 │   ├── data-mesh-patterns.md       # Pattern catalog + coverage matrix
 │   ├── observability-production.md # VRP logs, metrics, S3 proofs, DLQ
 │   └── images/                     # Architecture and product diagrams
@@ -838,16 +892,20 @@ make pre-commit     # Local quality gates
 
 ## License
 
-Apache-2.0. See [LICENSE](LICENSE).
+Apache-2.0 for the reference implementation code. See [LICENSE](LICENSE).
+
+**PVDM / Vaquar Pattern** is a **proprietary method** of **Vaquar Khan** (© 2024–2026). See [NOTICE](NOTICE) and [docs/vaquar-pattern.md](docs/vaquar-pattern.md).
+
+**Paper:** [arXiv:2608.14643](https://arxiv.org/abs/2608.14643) · **Reference gate + suite:** [Proof-gated-publication-PVDM](https://github.com/vaquarkhan/Proof-gated-publication-PVDM)
 
 ---
 
 <div align="center">
 
-**Serverless Data Mesh** · [PyPI](https://pypi.org/project/serverless-data-mesh/) · [Vaquar Pattern blog](docs/blog-the-vaquar-pattern.md) · [GitHub](https://github.com/vaquarkhan/aws-serverless-datamesh-framework)
+**Serverless Data Mesh** · [PyPI](https://pypi.org/project/serverless-data-mesh/) · [Paper](https://arxiv.org/abs/2608.14643) · [Vaquar Pattern blog](docs/blog-the-vaquar-pattern.md) · [GitHub](https://github.com/vaquarkhan/aws-serverless-datamesh-framework)
 
 *Domain teams own the write path. The mesh proves correctness before consumers see a snapshot.*
 
-*Vaquar Pattern (PVDM) — proprietary method by Vaquar Khan · reference implementation Apache-2.0*
+*Vaquar Pattern (PVDM)  -  © 2024–2026 Vaquar Khan · proprietary method · reference implementation Apache-2.0 · see [NOTICE](NOTICE)*
 
 </div>

serverless_data_mesh-1.3.0/VERSION

@@ -0,0 +1 @@
+1.3.0