SwiftStack Briefing Note
IT knows about object storage for its ability to store trillions of files, or objects, scale to almost limitless capacities and do it inexpensively. Cloud storage providers are accepting the technology with open arms, but enterprises are more cautious. They are choosing to stretch their network attached storage (NAS) investment just a little bit further. While the capabilities of object storage appeals to enterprises, they have existing architectures that need integration. They need more than just an object storage system, they need a system that can support all of their secondary storage needs.
SwiftStack is a software defined-storage (SDS) solution based on OpenStack Swift. It is a turnkey, software defined, object storage solution designed to help all organizations, not just OpenStack organizations, manage their unstructured data. SwiftStack is announcing version 4.0. This version of the solution adds many features to address the needs of enterprise customers.
Desktop & laptop users cannot easily interact with object storage, because the built-in clients speak file protocols. SwiftStack Drive is an optional agent for Mac or Windows that enables users to mount object storage containers to write to a SwiftStack object store directly. Applications that are native object can read and modify those same files. For example, a device could dump its data to the SwiftStack Cluster via SwiftStack Drive, and then a Hadoop Analytic engine can read that data.
SwiftStack always had the ability to create metadata about objects as they are written, where that metadata could be searched. But the application creating the data had to perform that search. If another application wanted to search that application’s data it would need to index all the object metadata. In the era of petabyte sized system that re-index is very time consuming. SwiftStack 4.0 works with Elasticsearch to provide a more global query of stored data. As the SwiftStack cluster stores data, it feeds the data into Elasticsearch. The result is any application or user can leverage the Elasticsearch to find data they are looking for. The combination also works with Kibana to provide a visualization of the data stored.
Built-in Load Balancer
The value of a scale-out storage architecture is that as IT adds each node, it scales performance and capacity automatically. The problem is that as the number of nodes grows it becomes more difficult to ensure it leverages each node’s capabilities uniformly. To overcome this, organizations often have to purchase a hardware load balancer, which are expensive and make scaling more difficult since they don’t scale in unison with the rest of the cluster. A load balancer software comes with SwiftStack 4.0 which resides on every new node, scales with the cluster and provides built-in high-availability. The result is a simpler way of making sure that there are no performance hot spots that need additional hardware. It also means that the storage team no longer has to count on the network team to manage load balancing, now they can do it themselves.
We’ve been tracking SwiftStack for a few years now and it hits all the marks IT professionals look for in object storage; it is open, affordable and scalable. Now in 4.0, the team delivers key features that enterprises need while maintaining its object roots.