A relatively unknown player aiming to de-throne Snowflake, Redshift, and BigQuery
I honestly had not heard of Firebolt until last week. I think this is probably the case with most of the tech community. Their performance claims over the data warehousing competition are something this article aims to dive into in a technical way.
What we know about the Technology
“Firebolt has spent several years innovating at the file format, indexing, and query engine levels “ — CEO Eldad Farkash
Materials on their tech is a bit sparse given their startup and stealth nature, but some highlights:
- It appears that Firebolt “is built cloud-native from the ground up in C++” (clarified by FireBolt head of product marketing — thanks!)
- Boasts a high-performance proprietary F3 — Firebolt File Format
- Leverages sparse indexes, natively sorted data on write
- Uses JIT / LLVM code optimizations, and Cost Based Query optimizers
- Optimized fast MERGE functions (update-insert / upserts) with some form of multi-versioning (MVCC)