With its 4th generation, a new era is coming for WEKA

WEKA, a leader in high end file storage, made a big splash with the 4th generation of WEKAfs unveiled yesterday. Originally available on-premises and then on AWS for a few years, the company now supports AWS, Azure, GCP and OCI making the residence of the operation a useless question. And as the cloud represents most of the time the only elastic option, offering WEKAfs almost everywhere breaks an adoption barrier. Very good.

With a proven model for on-premises with several key systems/servers vendors and on AWS with 230PB moved to the giant, this important step will for sure accelerate WEKA success. The recent quarters have demonstrated a rapid growth with accounts in various domains.

WEKAfs was designed to maximize the parallelism of any operations - data and metadata - with instance of data and metadata servers as containers making the architecture very flat and horizontal to deliver high IOs at very low latency, high throughput for small and large files.

Beyond this cloud agnostic approach, the team adds an incremental and continuously replication mechanism to facilitate file system promotion to active one when users need it. Designed since day 1 as a data tiering solution coupling local NVMe and object storage, WEKA also provides now an optimized data reduction technology based on variable chunks to reach new levels of economics.

WEKAfs exposed also NFS v4.1, SMB with a WEKA flavor and a native S3 interface in addition to GPUDirect making WEKA a perfect U3 data platform. I introduced the U3 concept several years ago as a combination of Universal, Unified and Ubiquitous model.

The product delivers a very rapid object retrieval from low cost storage when cold data are stored. The release also introduces a new UI that once again simplifies the management of clusters at scale. 

Definitely a key announcement that shakes the file storage landscape that confirmed once again that WEKA is a really a different animal.