The Communication Runtime for Distributed Systems
Decouple service logic from system operation. Control distributed execution at runtime — not through infrastructure.
A unified runtime that replaces your service mesh, API gateway, RPC framework, and most your YAML files.
Core ideas:
-
Routed RPC: Services don't call each other. The runtime routes.
-
Admission Control: Concurrency in slots. Your CPU metrics are lying to you.
Reference: Ubuntu 24.04, 4 vCPU/8GB = 1,500 slots/node.
2019 baseline: 2 vCPU/4GB = ~500 slots at 20k ops/s.
Tune based on your API Gateway Router actual throughput.
-
Transport Agnostic: REST, WebSocket, SSE, MCP - same execution model.
-
Runtime-Aware Edge: Cache invalidation by business logic, not TTL guesswork.
Not for you if: You've never run distributed systems in production. VIEApps NGX assumes you understand why "microservices" became "microservice hell".
For you if: You've paid the K8s tax. You've been bitten by service mesh latency. You've watched a single slow service take down your entire cluster. You know infrastructure shouldn't be your product.
By moving control from service code and infrastructure layers into a unified runtime, VIEApps NGX enables scalable, resilient, and loosely coupled system architectures.
This is not middleware. This is your system's central nervous system.
May 7, 2026 - National traffic spike:
- 3,278 concurrent requests
- 1.15M requests in 60 minutes
- 0 rejections, 0 nodes scaled
- 36% capacity used
The runtime applied backpressure when Redis choked. No circuit breakers fired.
We don't do circuit breakers. We do admission control.
The code is free. The expertise is hard-earned.
VIEApps NGX is open-source. You can clone it right now.
But if you don't understand:
-
Why ReactiveX for execution pipelines
-
Why WAMP Routed RPC instead of gRPC
-
Why slots instead of CPU thresholds
-
Why admission control beats auto-scaling
...then you'll deploy it wrong. And blame the tool.
We've been running this in production since 2019. We learned these lessons so you don't have to.
Code: https://github.com/vieapps
Docs: https://vieapps.com
Rollback plan: You write it.
On-call: You are.
Commercial Support: For when you realize free code doesn't mean free operations
Apache 2.0. Use it. Fork it. But if you go to production with it, talk to someone who's been there.
If K8s was your first distributed system, VIEApps NGX is your last.
Warranty: None.
If it breaks production, you get to keep both pieces.
This is not a toy. This is a weapon. Don't point it at your foot unless you're ready to pull the trigger.
