Saturday, September 16, 2017

Phazr.IO, Information Dispersal Algorithm for masses

Phazr.IO (www.phazr.io), develops an Information Dispersal Algorithm (IDA) technology for storage as an advanced method to protect and secure data.

Founded in 2016, the company is based in Los Angeles, CA, and was created by four co-founders, Gary Jin, Dr. Donald Chang, Steve K Chen and Jim Cheung. It has currently 6 people, but only Gary Jin is listed on LinkedIn. The company is currently in a pre-Series A phase and plans to announce the product around Q1 or Q2 2018.

The web site is pretty limited but the team has chosen to demonstrate their capability in IDA with a cloud storage product named PhazrBits, currently in 1.1 release, available on the AppleStore for iOS devices. It works as a multi-cloud aggregation gateway that splits, encodes and distributes data such photos or documents among multiple public cloud storage targets. The product is free and supported cloud storage back-ends are Google Drive, Microsoft OneDrive, Dropbox, Box and Apple iCloud.

They also started to collaborate with the OpenStack Swift community and pointers exist on openstack.org. Following this phase, the future product will be implemented probably within network adapters or controller boards to provide offload process capabilities. So OEM should be a good go-to-market model.

The IDA approach is a non-systematic erasure code form like the Mojette Transform promoted by Rozo Systems. You can find some entries for the libphazr on GitHub especially with PyEClib or liberasurecode. Phazr.IO is not open source at all, Rozo offers a standard version on GitHub. Tests made by the Phazr.IO team have shown 3x to 10x better than Jerasure and 2x to 3x faster than Intel ISA-L. So we expect a pretty innovative approach. The first use case targets bandwidth environment with video activity, perfect candidate with large files. And we expect to see more iteration soon.
Share:

0 commentaires: