IdeaWell

Project ideas from Hacker News discussions.

What if Jev spoke Arrow?

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1. The current performance bottleneck is round‑trip latency, not JSON deserialization
- “While Jev is fast, using it is hardly bottlenecked by deserializing from JSON.” – orf
- “With the API as it is today, the bottleneck is the thousands of round trips (one per state).” – ianmcook

2. Debate over the best serialization/transport approach (Arrow IPC, Arrow Flight, gRPC)
- “grpc is an RPC framework and Arrow IPC is a serialization format. You can use Arrow Flight which uses grpc to send Arrow IPC.” – mihevc
- “Using arrow is cool, but grpc seems like a more natural fit than arrow IPC and would give you the same benefits.” – orf
- “It would be straightforward to wrap this as an Arrow Flight (Arrow over gRPC) endpoint.” – ianmcook

3. Practical considerations and trade‑offs (HTTP/2 usage, weight, browser compatibility)
- “If I read the example correctly they are using long living http/2 connection to send arrow IPC. Which seems pretty pragmatic.” – mihevc
- “But it would make it heavier and less browser‑compatible.” – ianmcook

🚀 Project Ideas

Arrow Flight Batch Wrapper for TypeSafe

Summary

  • A lightweight middleware that buffers individual TypeSafe state requests and sends them in a single Arrow Flight (gRPC + Arrow IPC) batch, eliminating per‑request round trips.
  • Core value proposition: dramatically reduces latency and CPU overhead caused by thousands of JSON deserialization steps.

Details

Key Value
Target Audience Developers using TypeSafe (or similar Scala/JS backends) who suffer from high round‑trip costs
Core Feature Transparent batching of incoming HTTP/JSON calls into Arrow Flight messages sent to the backend
Tech Stack Scala (or Node.js), gRPC‑Java/Node, Arrow Flight library, Netty or Akka HTTP for front‑end
Difficulty Medium
Monetization Hobby

Notes

  • HN commenters noted: “The bottleneck is the thousands of round trips (one per state)” – this wrapper directly addresses that pain point.
  • Enables discussion on optimal batch sizes, back‑pressure handling, and compatibility with existing TypeSafe clients.

Arrow IPC over gRPC Adapter (Arrow Flight Lite)

Summary

  • A drop‑in adapter that exposes existing REST/JSON endpoints as Arrow Flight services, letting clients switch to binary Arrow IPC with minimal code changes.
  • Core value proposition: gives teams the performance benefits of Arrow Flight without rewriting their APIs.

Details

Key Value
Target Audience API owners who want to experiment with Arrow performance but cannot overhaul their service contracts
Core Feature Auto‑generates Arrow Flight service definitions from OpenAPI/JSON Schema and translates between JSON and Arrow records
Tech Stack Python (FastAPI + grpcio + pyarrow) or Go ( Gin + go-arrow + grpc-go ), Protobuf for service definition
Difficulty Low
Monetization Hobby

Notes

  • Commenters said “grpc seems like a more natural fit than arrow IPC” – this adapter lets you have both, using gRPC as the transport for Arrow data.
  • Sparks practical utility discussions about zero‑copy serialization, schema evolution, and browser‑compatible fallbacks (via gRPC‑Web).

TypeSafe Arrow Flight Benchmark Suite

Summary

  • An open‑source benchmarking harness that measures request latency, CPU usage, and bandwidth for TypeSafe APIs under JSON vs Arrow Flight (batched and unbatched) workloads.
  • Core value proposition: provides concrete data to decide whether investing in Arrow Flight/batching is worthwhile for a given workload.

Details

Key Value
Target Audience Performance engineers, backend teams, and open‑source maintainers evaluating TypeSafe optimizations
Core Feature Configurable load generator (hey/wrk2) with pluggable encoders (JSON, Arrow Flight) and detailed reporting (latency percentiles, throughput, serialization cost)
Tech Stack Go (hey) + Go Arrow flight client, optional Python scripts for plotting with matplotlib/pandas
Difficulty Low
Monetization Hobby

Notes

  • HN thread highlighted: “serialization would start to matter, and that's where Arrow would come in” – this tool lets you measure exactly that.
  • Encourages community discussion on trade‑offs, helps users justify adoption of Arrow Flight, and can be extended to other RPC frameworks.

Read Later