Plan your Buckit deployment by configuring servers, drives, and erasure code parity to find the right balance between storage efficiency and data resilience.
Buckit uses erasure coding to protect data against drive and server failures.
Each object is split into K data blocks and M parity blocks,
forming an erasure set of size N = K + M (the stripe size).
Storage Efficiency = K / (K + M) = (N - M) / N.
For example, with a stripe size of 16 and parity 4: efficiency = 12/16 = 75%.
Read Quorum requires K blocks (data shards).
Write Quorum requires K + 1 blocks when M = N/2
to prevent split-brain scenarios, otherwise K blocks.
The system can tolerate up to M simultaneous drive failures
while maintaining full read and write capability.
| Parity (M) | Data Blocks (K) | Efficiency | Drive Tolerance |
|---|