Subscribe
  • Home
  • /
  • Storage
  • /
  • InfiniBox unified storage: NAS capabilities

InfiniBox unified storage: NAS capabilities


Johannesburg, 05 Nov 2015
Whitepaper: InfiniBox unified storage: NAS capabilities
Whitepaper: InfiniBox unified storage: NAS capabilities

As network attached storage (NAS) continues to outpace growth expectations, storage vendors often respond to this demand by creating solutions from limited, inefficient architectures. For example, implementing NAS over SAN (which creates performance bottlenecks), providing NAS-only solutions (which creates storage silos), or providing scale-out solutions (which lack the ability to automatically distribute workloads across all nodes).

Unlike these solutions, InfiniBox was designed from day one to be a unified storage solution. Its core architecture is based on a data layout and metadata structure that caters to the needs of each protocol type (block and file). This core includes all the data services (snapshots, clones, dual-parity data protection, common data pooling, etc) that both SAN and NAS need, and use. This allows INFINIDAT to offer a similar feature set on both SAN and NAS, simplifying storage administration and reducing footprint, cost and complexity. Our core design and all of our development are based on three priorities - reliability, performance, and ease of use.

Designing a NAS solution from the ground up to scale to over 2PB requires a lot of attention to performance optimisation. With an extremely efficient Btree-based metadata, InfiniBox NAS leapfrogs the large number of indirect metadata reads required to perform a read from disk in large NAS environments. It always points directly to the iNode, reducing read latency to the absolute minimum. Combined with a thick SSD layer to cache data and metadata, InfiniBox NAS offers an ideal combination of cost and performance.

Share