Start-up Caringo this week is enhancing its content-addressable storage system by adding local- and wide-area replication.

Caringo’s CAStor 2.0 software runs on commodity x86 servers that have a pair of Serial ATA drives in a cluster configuration. Each node in the cluster contains a USB memory key, which contains the CAStor software and enables the server to run as a content-addressable storage device. A CAStor cluster can consist of hundreds of nodes.

All files stored on a CAStor cluster are replicated to other nodes for disaster recovery. If a disk fails, the data can be reconstructed from the data replicated to other nodes.

The new version of CAStor adds asymmetric local- and wide-area replication to enhance the disaster recovery aspects of the cluster. Redundant data resides on separate sub-clusters within the same site or on a remote site.

Local-area replication lets an IT administrator create sub-clusters in CAStor and split the cluster based on location. Wide-area replication is used for disaster recovery and availability.

CAStor 2.0 also includes a new Web-based management console, from which IT administrators can monitor the health of the cluster and perform administrative tasks. It also supports booting of the servers from the network, centralized configuration files and performance enhancements.

This version of CAStor also introduces the concept of ‘anchor streams.’ An anchor stream is an updateable object with a unique identifier, where a user can store and update information such as database backups. Anchor Streams are useful for backups and system recovery. IT can create an anchor stream and write information to the unique identifier each time they back up CAStor. To recover information, the IT administrator simply accesses the unique identifier.

CAStor was created by Paul Carpentier, a former architect of FilePool, which EMC acquired and turned into Centera.

The product is available now.

Deni Connor is principal analyst for Storage Strategies NOW.