NAS vs. SAN
The debate is over.

Recently, network attached storage (NAS) vs. storage attached networks (SAN) has been portrayed as the major battle in the world of information management. Truth is, however, the so-called battle is largely hype.

NAS and SAN—acronyms for methods of attaching storage (where information lives) and servers (where applications live), IP-respectively—are merely different solutions to different data storage needs. Not only are they complementary, but, in some ways, the two are starting to converge. The point to remember is that it is all about the information and the application—how and for what purpose information needs to be moved, managed, enhanced and presented.

SANs were introduced with the goal of consolidating “islands” of server-bound storage. The idea was to free the servers to do what they do best—process data—while housing information in a central location, where it can be managed, protected and shared. SANs also reduce much of the bulk server-to-server data movement that has been clogging the conventional corporate networks—local area networks, for example—for years.

SANs are optimized for applications, such as databases, data warehousing, transaction processing and technical computing. In each case, a single, tens or hundreds of servers or mainframe hosts require vast amounts of information to perform given applications. Typically, these applications require high volumes of reads and writes, the information needs to be moved quickly between storage and server, and the data needs to be available to only a single or few users or applications at any given moment.

ADVANTAGES OF SAN

The success of SANs is due to a network interconnection technology referred to as Fibre Channel. A combination of industry-standard cabling hardware and software protocols, Fibre Channel provides the capability to connect multiple servers with multiple storage systems.

By implementing a full Fibre Channel-based SAN—consisting of interconnected storage, servers and switches—a robust fabric topology can be implemented, connecting many storage elements to many computing systems, while still using familiar channel I/O-based technology. As a result, storage infrastructures can now be designed to provide access to any storage system from any host or server, greater consolidation of storage, better utilization of storage assets and management by fewer people.

The advantages are immediately apparent: increased connectivity, increased manageability, increased flexibility for change and higher performance for a campus or metropolitan computing environment.

NAS uses the maturity and ubiquity of IP-based networks to provide access to stored information. While SANs essentially provide channel topologies to deliver large, block-level data to servers, NAS enables many users to access individual files at the same time. NAS is ideal for serving Web pages to hundreds or thousands of PCs simultaneously, or enabling multiple designers to work on a shared engineering design. When you access a network drive through your workstation, you are using NAS.

To leverage the existing IP network for this purpose, NAS solutions employ specialized file-serving protocols, such as NFS (for Unix) and CIFS (for Windows NT). These protocols enable the servers to communicate efficiently with a file server—a specialized computer that manages, accesses and distributes information over the network.

ADVANTAGES OF NAS

The ability of the NAS file server to provide a centralized point of management for the file system enables multiple clients to access single files simultaneously. This makes NAS ideal for applications and environments that require the sharing of files among multiple hosts or clients, even if they each have different operating systems.

While both technologies offer advantages and disadvantages, enterprise storage network implementations often require both NAS and SAN. For example, a typical enterprise environment will have network-shared directories and Web-serving (well suited for NAS), and, at the same time, have client/server and database application (well suited for SAN). In optimal cases, the file server used for NAS can be connected on the SAN to its storage.

Further, storage networking solutions that implement multipath file-serving (MPFS) software enhance the performance and flexibility of the storage network by leveraging the best of both SAN and NAS for use by a traditional NAS application. Using MPFS software, large-volume NAS-based data can be sent over a SAN channel to a host, thereby avoiding the performance decrease that may occur in an IP network when large files are transferred.

Increasingly, networked storage is becoming a marriage of enterprise storage, storage area networks and network attached storage. The most practical and productive approach is to consider where each technology can be best applied—separately or in combination—to achieve the information needs of the organization.

Fierro is director of enterprise storage networks at EMC Corp., Hopkinton, MA.