Gateways and Clusters: The SAN/NAS Handshake

Share it on Twitter  
Share it on Facebook  
Share it on Linked in  

Whether you call it unified storage or bridging the SAN/NAS divide, there is clearly a strong demand for getting the various storage environments that exist in virtually any enterprise to work together.


But what's the best way of going about it? In fact, is there a single "best way" at all? Probably not, judging by the variety of solutions out there.


One of the newest approaches is the NAS gateway, which backers like EquaLogic and Tek-Tools Software claim offers a single point of management for NAS and SAN to allow greater use of low-cost heterogeneous SAN arrays within the NAS infrastructure, according to Enterprise Storage Forum. They also allow both storage networks to work with both block and file-based traffic, improving data availability throughout the enterprise.


Integrated storage management is a key component of unified storage, although there is some disagreement as to what that actually means. In this eWEEK piece, CommVault claims its Simpana 7.0 suite offers a single, Web-based means of protection, replication and archiving across multiple tiers of storage. While most systems use multiple controllers for SAN and NAS environments, Simpana delivers a common management architecture that tackles different environments through plug-in options, letting a single operator manage the entire storage network.


NAS clustering is also emerging as a way to simplify the SAN/NAS relationship. Sanbolic Inc., of Watertown, Mass., says its Melio cluster file system allows multiple NAS appliances to be clustered into a common file system on external SAN storage, giving all storage servers concurrent read and write access. This arrangement is said to be ideal for write-intensive applications that require substantial network file resources.


Clustered storage also offers better scalability than traditional storage networks. I/O bandwidth, for one, can be scaled up or down more easily, and the clustered approach is more modular, so you can add what you need when you need it, even while the network still considers it a single storage system, rather than a multitude of arrays.


That's probably part of the reason clustered storage is drawing most of the attention in high-performance computing environments.