Cloud-Based App Development Tools

Gluster Wednesday materialized a Connector for OpenStack that will hitch the GlusterFS file system to the OpenStack Compute block storage controller and resolve a built-in limitation in the young open source cloud platform.

See, OpenStack can store data as objects, complements of its embedded OpenStack Object Storage (a k a Swift), or it can store data as files, complements of CEPH, which has just been contributed to the effort. It can't do both at the same time.

Gluster says it can and is offering its widgetry to OpenStack as a core piece of the stack.

The contribution is supposed to give OpenStack highly scalable, highly available VM storage functionality and let users add scale-out integrated file and object storage to any deployment.

What that means is that users can scale-out the number of VMs deployed in their cloud environment and the widgetry will support the virtual motion of the VMs in the OpenStack compute environment.

Gluster says these two capabilities together will enable OpenStack users to centralize on one storage solution for VMs, object and file data and simplify their storage environment.

So much for the Connector, now about GlusterFS, which the Connector is supposed to, well, connect to OpenStack's Xen and KVM hypervisors, as you might assume, and deliver massive scalability, high availability and replication.

The company says the widgetry is designed for the most demanding workloads.

Currently in beta, GlusterFS 3.3 provides a software-only system for data storage that lets users access the same data as an object and a file, which is supposed to simplify management and control storage costs.

It's advertised as the "first true unified file and object data storage." It's also supposed to accelerate enterprise cloud adoption.

Gluster CTO AB Periasamy says, "Traditional object storage technologies have not been widely deployed because their approach has not integrated well with legacy applications - leaving 95% of those applications unable to migrate easily to the cloud. Additionally, NAS has been unable to scale across the Internet to the public cloud."

So GlusterFS 3.3 should ease the migration of legacy file-based applications to object storage for use in the cloud.

By integrating file and object storage in the file system GlusterFS 3.3 is supposed to solve both the NAS scaling and legacy application integration with object storage problems.

Users can access data as objects from an Amazon S3-compatible interface - and of course OpenStack is an Amazon clone - and access files from a NAS interface like NFS or CIFS.

For infrastructure-as-a-service offerings, the company says GlusterFS 3.3 will let organizations build their own Amazon-like storage.

It's also supposed to deliver massive scalability like petabytes of capacity and GB/s of throughput on commodity physical and cloud hardware.

The company says Gluster can scale-out availability on-demand, enabling both Synchronous N-way replication in the data center and across high-speed links and Asynchronous Geo-Replication between data centers across WANs.

GlusterFS is packaged in an Amazon Machine Image (AMI) for deployment in the AWS cloud.

For private clouds, Gluster Virtual Storage Appliances integrate GlusterFS into a virtual machine for deployment on hypervisor environments such as Red Hat KVM, VMware ESX and Citrix Xen and run on the enterprise's choice of certified hardware platforms.

The GlusterFS 3.3 beta can be downloaded at http://bit.ly/33beta. The Gluster Connector for OpenStack should be available by now.