zx File is fast file transfer software — the file-movement engine inside the Zettar zx data mover. In 2018, under a hardware-imposed 80 Gbps cap, it moved 1 PB in 29 hours over a 5,000-mile loop at 96% utilization — encrypted and checksummed the whole way. Move bulk file and file↔object data at roughly 10× typical movers, and reach a verified match before you cut over.
One engine, tuned for bulk file and file↔object workloads. It runs on standard servers at line rate and scales out with no software ceiling — about 10× the throughput of typical commercial or freeware movers. Your migrations and syncs finish in hours and days, not weeks.
Massive parallelism drives every transfer concurrently across nodes and streams — line rate, scaling out with no software ceiling instead of leaving it idle.
One engine moves data file-to-file and file-to-object — NAS, parallel file systems, and cloud object stores — so you don't stitch together separate tools for each path.
Repeated incremental passes close the gap between source and target. Cut over only when the two are provably identical — no guesswork, no missed files.
Unconditional checksums plus TLS protect every transfer. What lands on the target is verifiably bit-for-bit identical to what left the source.
"Zettar moved an actual petabyte over a 5,000-mile network loop in 29 hours — with encryption and checksumming — at 96% bandwidth utilization."
The file-movement engine inside the unified Zettar zx data mover — bulk file and file-to-object transfer at petabyte scale, roughly 10x faster than typical commercial or freeware movers, encrypted and checksummed end to end.
It runs at line rate and scales out with no software ceiling. The 1 PB in 29 hours record — at 96% utilization, line rate, with SLAC and ESnet — ran on zx File.
rsync and scp are single-stream, and none of the three was built for petascale. zx File drives massive parallelism across nodes and streams for about 10x the throughput, and reaches a verified match before you cut over.
Yes — file-to-file and file-to-object across NAS, parallel file systems, and S3-compatible object stores, in native formats with no wrappers and no lock-in.
Yes. Massive parallelism keeps huge file counts moving instead of stalling on small files, and every transfer is protected by unconditional end-to-end checksums and TLS.
See how zx File moves bulk file and file↔object data at record-setting speed — with massive parallelism, incremental sync to convergence, and end-to-end integrity.