Volnux: Technical FAQ - nshaibu/volnux GitHub Wiki

1. How does Volnux handle "Task Explosion" (e.g., 1 Million+ concurrent events)?

Volnux uses Atomic Batching and a Distributed Instruction Pointer. Instead of loading a million tasks into a local event loop, the ControlFlowEvent partitions data into manageable ResultSets.

• The Mesh Advantage: Our RemoteExecutor uses gRPC streaming to load-balance these tasks across the cluster.
• Backpressure: Because the mesh is "Nervous-System Integrated," if Redis latency spikes, the Gossip Protocol throttles task submission at the source, preventing a system-wide "OOM" (Out of Memory) crash.

2. How do I debug a Pointy-Lang script if it's "Data, not Code"?

Debugging Volnux is actually easier than debugging an opaque Python async loop.

• AST Inspection: Since the script is parsed into an Abstract Syntax Tree (AST), you can visualize the entire execution path before it runs. This can be done by the Volnux visualisation layer.
• The Live Spine: You can query the Redis Context Spine in real-time. By looking at a specific Context ID, you can see the exact "Snapshot" of every variable at any point in the tree's history without stopping the engine.

3. What is the "Total Cost of Ownership" (TCO) compared to Temporal?

Temporal requires a heavy persistence layer (Cassandra/Postgres/ElasticSearch) and a complex "Visibility" store.

• Volnux Efficiency: Volnux runs entirely on Redis. By moving from "Event Sourcing" (storing every history step) to "State-Pointer Rehydration" (storing only the current instruction and parents), we reduce storage overhead by up to 80%.
• Worker Density: Because Volnux workers are stateless and use an iterative engine, you can run significantly more "Logical Workflows" per GB of RAM than a thread-heavy Temporal worker.

4. Is Volnux "Replay Safe"?

Yes, but by design, not by restriction. Unlike Temporal, which forces you to avoid non-deterministic Python (like datetime.now() or random numbers), Volnux is Resumption-Based.

• The Logic: We don't re-run your code to find the state. We Rehydrate the state and jump to the next instruction. This means you can use any standard Python library or side effect without breaking the engine's ability to recover from a crash.

5. How does the "Doubly-Linked Tree" handle Data Privacy (GDPR/HIPAA)?

Most orchestrators have a "Flat" namespace (Global variables). Volnux uses Lexical Scoping.

• The "Pruning" Feature: A sub-workflow (Child Context) can only access data from its direct parent. You can explicitly "Prune" or "Shadow" sensitive data (like Patient IDs) so that downstream analytics tasks physically cannot see the memory address of the sensitive data, even if they are in the same workflow.