Hot vs Cold Storage
Hot storage is fast, higher-cost storage for data you access often. Cold storage is slower, lower-cost storage for data you rarely touch. Most systems use both and move data between them as it ages.
Matching storage to how data is used
The distinction exists because access patterns are uneven. Recent data is queried constantly and needs to be fast. Old data is queried rarely, usually for audits or long-range analysis, and does not justify premium storage prices.
Hot storage, like local SSD or block storage, gives low latency at a higher per-gigabyte cost. Cold storage, like object storage, costs a fraction as much but has higher latency per request. Keeping everything hot is expensive. Moving everything cold makes recent queries slow. The balance is to place each piece of data where its access pattern fits.
Good systems handle this automatically, ageing data from hot to cold while keeping it all queryable.
How Arc handles Hot vs Cold Storage
Arc stores data as Parquet on object storage, the cold and cheap tier, while keeping it queryable, and supports tiering so frequently used data can sit closer to compute. You keep long history without paying hot-storage prices for all of it.
Arc is a high-performance columnar database. Open Parquet on storage you own, single Go binary, production-ready in 30 seconds.