During Phase 15 deployment verification, we ran benchmarks at 128/256/512 MBps throughput capacity with a 202-byte JSON manifest:
| Throughput |
P50 @ conc=1 |
P50 @ conc=25 |
P50 @ conc=50 |
| 256 MBps |
56.9 ms |
60.3 ms |
257.9 ms |
| 512 MBps |
59.8 ms |
59.9 ms |
246.1 ms |
Conclusion: For metadata-heavy workloads (JSON manifests, small config files, document headers), throughput capacity increase has zero effect on latency. The bottleneck is connection overhead (TLS + S3 AP routing), not bandwidth. Save costs by staying at 128 MBps for these workloads.
Sizing reference from a specific test environment, not a service limit.
Documentation: 8 Languages ×ばつ 28 Patterns
Every pattern includes documentation in:
🇯🇵 Japanese (primary) · 🇺🇸 English · 🇰🇷 Korean · 🇨🇳 Chinese (Simplified) · 🇹🇼 Chinese (Traditional) · 🇫🇷 French · 🇩🇪 German · 🇪🇸 Spanish
Each language includes:
-
README.md — Overview, deployment, success metrics
-
docs/architecture.md — Mermaid data flow diagram
-
docs/demo-guide.md — Step-by-step demo with verification checklist
Each UC README includes Success Metrics with Business Outcome, Technical KPI, Quality KPI, Cost KPI, and Go/No-Go criteria. This article summarizes the portfolio; detailed success criteria live with each pattern.
What Changed Since Phase 14
| Metric |
Phase 14 |
Phase 15 |
Delta |
| Use cases |
17 |
28 |
+11 |
| Total patterns |
24 |
35 |
+11 |
| Test count |
~800 |
1,499+ |
+699 |
| Industries covered |
14/22 |
19/22 |
+5 |
| Languages |
8 |
8 |
— |
| Shared modules |
8 |
11 |
+3 |
| Documentation files |
~400 |
~700 |
+300 |
Who Should Use Each New Pattern?
Recommended Starting Patterns
| Start here if... |
Pattern |
Why |
| You want document intelligence |
UC20 or UC26 |
Multilingual extraction + property/lease analysis |
| You want log analytics |
UC18 |
CDR/syslog anomaly detection with baseline |
| You need PII-safe document triage |
UC27 |
Protected characteristic exclusion built-in |
| You need inspection workflows |
UC22 or UC25 |
Safety-critical escalation + tri-modal |
| You want ESG extraction |
UC23 |
Multi-framework mapping (GRI/TCFD/ISSB) |
Full Pattern List
| If you are... |
Start with... |
Why |
| Telecom operator with CDR data |
UC18 |
Anomaly detection across network logs |
| Ad agency managing creative assets |
UC19 |
Automated brand compliance scoring |
| Hotel chain with inspection photos |
UC20 |
Facility condition monitoring at scale |
| Agricultural cooperative |
UC21 |
Crop health + traceability in one workflow |
| Railway/transit operator |
UC22 |
Safety-critical deterioration detection |
| ESG reporting team |
UC23 |
Multi-framework metric extraction |
| Grant-making foundation |
UC24 |
Application processing + outcome matching |
| Power utility with drone programs |
UC25 |
Tri-modal inspection (visual + SCADA + thermal) |
| Real estate portfolio manager |
UC26 |
Property analysis + lease extraction |
| Recruiting team (APAC/EMEA) |
UC27 |
PII-compliant recruiting document triage |
| Chemical manufacturer |
UC28 |
SDS compliance + lab notebook digitization |
What's Next
-
VPC-internal Lambda benchmark — True VPC path performance (eliminates Internet latency)
-
FPolicy TCP-level Replay Storm — Real ONTAP event replay (requires ECS rebuild)
-
Cross-repository integration — Link patterns to fsxn-lakehouse-integrations for analytics pipelines
-
Glue Data Catalog integration — Schema versioning and data quality checks for output datasets
-
Community contributions — Pattern template for community-submitted industry use cases
Resolved from Phase 14: FlexCache ×ばつ S3 AP integration confirmed as not currently supported by AWS — tracked in Field Feedback Log. FC1 Recovery Metrics depend on this feature. Both remain pending AWS feature availability.
Ownership Model
| Layer |
Recommended Owner |
Shared modules (shared/) |
Platform / DevOps team |
UC business logic (functions/) |
Application / data team |
| FSx ONTAP and S3 AP infrastructure |
Storage / platform team |
| IAM, data classification, encryption |
Security team |
| Success metrics and Go/No-Go |
Business owner |
| Regulatory compliance mapping |
GRC / legal team |
Compliance Positioning
These templates do not certify compliance with any specific regulation. They provide implementation hooks for audit logging, retention, classification, and human review that customers can map to their regulatory controls. Each organization must independently validate compliance with applicable regulations (FISC, HIPAA, GDPR, NARA, local labor law, etc.).
NetApp / ONTAP Operational Notes
For production deployments on FSx for ONTAP, review the ONTAP-specific guidance in docs/ontap-integration-notes.md, including:
- SVM / volume / protocol scope assumptions
- NFS/SMB visibility of S3 AP-generated outputs (file ownership = AP file system identity)
- IAM + S3 AP policy + ONTAP file identity behavior, separate from NFS export policy evaluation
- Snapshot / SnapMirror / retention impact on output artifacts
- Scheduler vs FPolicy trigger mode selection
- FlexCache / FlexClone combination patterns per UC
- NetApp support diagnostic bundle
- OT/manufacturing safety caveat
FlexCache/FlexClone note: UC ×ばつ FC combination patterns describe adjacent architecture patterns. Validate current AWS/FSx feature support before assuming direct S3 AP access to cached or cloned paths.
Benchmark scope: Results are from Single-AZ, First-generation FSx ONTAP. Validate separately for Multi-AZ or newer generation file systems.
Regulated research workflows (UC7, UC28, FC5): Capture input dataset version, model/prompt version, reviewer action, and output checksum as lineage metadata. See shared/lineage.py v2 fields.
Stats
-
New patterns: 11 (UC18-UC28)
-
New Lambda functions: 44 (4 per pattern average)
-
New tests: 699
-
New documentation files: ~300 (across 8 languages)
-
New shared modules:
data_classification.py, human_review.py, schemas/events.py
-
Deployment verified: All 28 UCs achieved SUCCEEDED status in ap-northeast-1
-
Benchmark runs: 2 additional (256/512 MBps small-file comparison)
-
Cost: ~10ドル total for deployment verification (Lambda + Step Functions + Bedrock Nova Lite)
Try It Today
git clone https://github.com/Yoshiki0705/FSx-for-ONTAP-S3AccessPoints-Serverless-Patterns.git
cd FSx-for-ONTAP-S3AccessPoints-Serverless-Patterns
# Quick test (no AWS account needed)
make test-quick
# Deploy any pattern with DemoMode (no FSx ONTAP needed)
cd telecom-network-analytics
cp samconfig.toml.example samconfig.toml
sam build && sam deploy --guided
Repository: github.com/Yoshiki0705/FSx-for-ONTAP-S3AccessPoints-Serverless-Patterns
Full series: FSx for ONTAP S3 Access Points on DEV.to