CrackingWalnuts offers 24 free interactive system design tools for engineers. Use these calculators and visualizers to build intuition for distributed systems concepts, prepare for system design interviews, and estimate capacity requirements.
Estimate QPS, storage, bandwidth from DAU and request patterns.
Convert nines to downtime. Compound SLA for dependent services.
Transfer time from 1 MB to 1 PB across common link speeds.
Interactive latency reference chart from L1 cache to cross-continent network calls.
Answer workload questions, get a database recommendation.
Convert Mbps, MB/s, QPS, storage units, and time durations.
Hash ring with key-to-node mapping. Add/remove nodes live.
Token Bucket vs Fixed Window vs Sliding Window side-by-side.
Membership testing with false positive demonstration.
Step through operations, watch evictions, compare hit rates.
Configure N, W, R. Simulate failures, test consistency.
Bully vs Ring vs Raft. Trigger elections, simulate failures.
Closed-Open-Half-Open state machine. Send requests, watch state transitions.
Hash vs Range partitioning. Detect hotspots, analyze distribution.
Pub/Sub vs Point-to-Point vs Consumer Groups. Simulate message delivery.
Round Robin vs Weighted vs Least Connections vs IP Hash. Kill servers mid-stream.
Partition the network, toggle CP vs AP, watch consistency and availability trade off.
Watch information spread through a cluster tick by tick. Adjust fanout, kill nodes.
Insert, search, split, compact. Compare read-optimized vs write-optimized structures.
Two-Phase Commit vs Saga. Inject failures, watch abort and compensate flows.
Exponential backoff with jitter. Thundering herd demo shows why jitter matters.
Edit data blocks, watch hashes propagate. Compare trees with log(n) verification.
Borrow, use, return connections. Leak them and watch the pool drain.
Project storage, QPS, bandwidth over time. Know when to shard or go multi-region.
Calculate LLM inference cost by task type, model tier, and volume. Compare API vs self-hosted.
Size a self-hosted GPU fleet from model size, quantization, and QPS targets.
Calculate vector database storage, HNSW overhead, and embedding costs for RAG.