SAN storage systems are the most popular method for networking storage. SAN hardware outsells NAS products by more than $10 billion annually. This is truly an amazing feat, since SANs are more expensive, more complex and less intelligent. SANs have no clue about application data. The advantage of a SAN is its ability to work with any application with optimal performance. Additionally, SANs are time tested and are currently used for a large portion of mission-critical application storage.
NAS, on the other hand, is more intelligent than SAN storage and understands files and file attributes. A fundamental difference between SAN and NAS is the way in which data is accessed. In order to access data on a SAN, users must go through an application server, which provides the intelligence to gain access to files and database information. With NAS systems, users can access files over the network directly without having to go through an application server. This comes in handy when users want to share files as part of a workflow process. However, NAS uses higher layer protocols NFS and CIFS to communicate, adding latency to the process of reading and writing data.
Another type of storage system is content aware storage (CAS). The EMC Centera is a CAS product that uses proprietary APIs to communicate with applications. The software vendor must integrate with the Centera's APIs to store data and take advantage of its retention policies. Alternately, customers can use the Centera Universal Access (CUA) device to store data using standard NFS or CIFS protocols. CAS creates a unique address of each object and its metadata and manages the storage and data retrieval in a very large and flat address space. One of the most popular uses of CAS is setting policies ensuring that files cannot be altered or deleted for specific periods of time to meet with compliance regulations.
CAS is a new approach to storage but much, if not all, of what it does can be accomplished with NAS-based storage systems. While traditional NAS products don't match the scalability of leading CAS products, next generation NAS systems do. In fact, many of the products that compete directly with Centera are NAS-based products with scalable file systems and clustered network architectures. The thing that makes Centera stand out, in addition to its architecture and scalability, is the number of applications that it is integrated with. There are over 200 applications that have been integrated to work with the EMC Centera APIs, and that number is growing.
To make things even more complicated, IBM and SGI have SAN file systems. They actually install a network file system that communicates over FC. These SAN file systems provide the file-level intelligence found with NAS for SAN-based storage.
Having knowledge of files, file attributes and metadata changes the game for managing, protecting and recovering data through its lifecycle. The next article will cover benefits of file-system-based storage now and in the future, and why SAN storage will endure regardless of its lack of file-level awareness.
About the author:
Tony Asaro is a senior analyst for the Enterprise Strategy Group.