Design resource loader and ROS-like pub/sub
Overview
This question evaluates system design and distributed-systems competencies including resource management, async concurrency, caching and eviction, streaming and partial loads, dependency resolution, failure isolation, observability, and pub/sub concerns such as topic/schema/versioning, discovery, QoS, ordering, replay, partitioning, latency, clock semantics, and security. Commonly asked to assess architectural reasoning about scalability, low-latency trade-offs, fault tolerance, and deployment/topology choices, it is categorized under System Design and tests both high-level conceptual understanding and practical application-level design decisions; English.
System Design: Resource Loader and ROS-like Publish/Subscribe
Context
You are designing two foundational systems for a large-scale, real-time interactive platform with both client and server components. The platform must run across heterogeneous devices and networks, support low-latency experiences, and scale to millions of users and thousands of services.
Part A — Resource Loader
Design a Resource Loader that supports multiple resource types (e.g., images, configs, ML models, audio). Provide:
- Public APIs and data model
- How new resource types are plugged in (extensibility)
- Async/concurrent loading and cancellation
- Prioritization and dynamic reprioritization
- Batching and coalescing
- Caching: memory and disk; eviction policies; cache keys; validation
- Deduplication of in-flight requests
- Retries, timeouts, and hedged requests
- Dependency resolution (manifests, topological loading)
- Streaming/partial loads (e.g., progressive textures, range requests)
- Failure isolation and fault containment (e.g., decoder crashes)
- Thread-safety and correctness under concurrency
- Observability (metrics, logs, tracing)
- Deployment, scaling strategies (client/server, edge), and trade-offs
Part B — ROS-like Publish/Subscribe System
Design a ROS-like pub/sub system for distributed nodes (clients, services, simulators, and ML components). Define:
- Topics, message schema/serialization, and versioning
- Discovery mechanisms (LAN/WAN, NAT) and routing
- QoS and reliability options (best-effort vs reliable), delivery semantics
- Ordering guarantees (per-publisher, per-key)
- Replay and durability options
- Backpressure/flow control and handling slow/disconnected subscribers
- Partitioning and scaling across nodes and regions
- Latency/real-time considerations and zero-copy paths
- Clock synchronization and time semantics
- Fault tolerance, high availability, and failover
- Security (authentication, authorization, encryption)
- Comparison and justification of brokered vs brokerless and federated topologies
Be explicit about key trade-offs and justify your choices with concise reasoning.
Solution
Login required