Node Requirements
FlameWire targets full‑archive nodes so developers can query any historical state. Recommended specs per chain (2025‑Q2):
Ethereum – Geth --gcmode=archive
16 vCPU
64 GB
8 TB
1 Gbps
Highest disk‑I/O and RAM footprint; needed if you must serve every legacy debug trace.
Ethereum – Erigon archive
8 vCPU (12 vCPU ideal)
32 GB
4 TB
1 Gbps
Table‑oriented storage ≈ 50 % smaller than Geth; good balance of speed and cost.
Sui – full node
16 vCPU
128 GB
4 TB
500 Mbps
Indexer doubles IOPS; NVMe essential for checkpoint queries.
Bittensor – archive node
8 vCPU
32 GB
2 TB
500 Mbps
Runs performance scoring and gossip; modest disk but steady network traffic.
General tips
Prefer bare‑metal or dedicated cloud instances to avoid noisy‑neighbour I/O penalties.
All drives should be NVMe with latency < 100 µs and ≥ 100 k IOPS; SATA SSDs will bottleneck archive reads.
Stripe two NVMe drives (RAID 0) on Ethereum if budget permits—doubles IOPS for big syncs and heavy tracing.
Plan storage head‑room: Ethereum archive grows ~0.5 TB/year; Sui growth is ~0.2 TB/year at current throughput.
1 Gbps provides comfortable headroom; miners in bandwidth‑constrained regions should cap RPC concurrency to avoid latency penalties.
Last updated