Previous Chapter
Chapter Home
Storage Networking Primer Home

But the information technology professionals in most organizations are responsible for keeping their enterprises' data secure and accessible while they await the brave new world of information processing. Even from this short-term vantage point, SANs have tremendous potential. Used optimally, SANs can improve applications' access to data, dramatically reduce unnecessary redundancy, eliminate bandwidth contention between client messages and data access, and off-load resource-intensive tasks like backup to intelligent storage devices on I/O-optimized interconnects, Information technology professionals cannot afford to ignore the near-term benefits of SAN technology.

A Couple of Closing Clarifications

This section touches upon a couple of aspects of terminology that in the authors' experience tend to be sources of confusion rather than clarification in discussions of storage networking. We hope it dispels some of the very basic uncertainty around SAN terminology for readers and lets them get on to the serious stuff.

Is a SAN Storage or a Network?

Strictly speaking, a SAN is a storage area network. Much energy is expended uselessly on silly arguments about whether the term SAN should include storage and computers or only the interconnects and other components that link the storage and computers together. Since it's difficult to imagine talking about a SAN without talking about the storage and computers it interconnects, we adopt the more inclusive usage in this book:

A storage area network (SAN) is a network of interconnected computers and data storage devices. As we use it in this book, the term SAN specifically includes both the interconnection infrastructure and the computers and storage that it links together.

Of course, we frequently have occasion to refer specifically to the storage devices attached to a SAN, as distinguished from the more inclusive concept. We use the terms SAN-attached storage and SAN-attached storage device for this purpose.

What Is This Area Thing, Anyway?

One might wonder what the area in a storage area network is, and rightly so. From what we've described so far, the term storage network would seem to apply equally well if not better. What is this area thing, anyway?

There's a strong relationship between the SAN concept and the popular I/O interconnect technology called Fibre Channel. In fact, it's fair to say that Fibre Channel technology is what got the whole SAN movement started.

In the mid-1990s, the companies committed to developing Fibre Channel were in the doldrums. Five years or more had been spent on development, and the products and market weren't emerging as fast as investors would have preferred. Something was needed to generate excitement among consumers and jump-start growth.

Members of the Fibre Channel Loop Community and the Fibre Channel Association, the two trade groups promoting the technology at that time, hit upon the idea of capitalizing on the similarities between Fibre Channel and local area networking:

Universal connectivity. With Fibre Channel, lots of storage devices can be connected directly to lots of computers.

Campuswide separation. With Fibre Channel, storage devices don't have to be in the computer room. They can be anywhere on a fairly sizable campus.

Bandwidth multiplication. With switches (Chapter 7 describes switch capabilities), many high-performance links can be aggregated into a fabric with even higher aggregate performance.

Dynamic reconfiguration. Fibre Channel allows devices and computers to be connected to and disconnected from the network while it is operating.

Enabling of new storage usage styles. Just as local area networks enabled client/server computing, Fibre Channel connectivity enables new ways of using storage to achieve business goals. Sharing access to disk and tape subsystems and direct copying of data between devices are two examples.

Noticing the networking industry's success in creating a strong identity for itself through the use of the LAN acronym, the Fibre Channel community set out to achieve something similar by introducing the term storage area network, or SAN, to describe networks of intelligent storage devices and servers. White papers were written, lectures were given, and articles were published, all pointing out similarities between the local area network that connects all the computers in a building or campus and a storage area network, with the potential to likewise connect all the storage within a similar area. The SAN acronym, with its implicit analogy to LANs, was repeated over and over again.

Although the SAN concept was largely a so-called "marketecture" when first introduced, the aforementioned promises have essentially been delivered upon in the intervening years. The perception created by early Fibre Channel industry marketers has become reality. Products and capabilities are being delivered and the term SAN has become thoroughly embedded in the public consciousness.

Besides, storage networking makes a lousy acronym.

Is "NAS" Just "SAN" Spelled Backward?

One often hears discussions of network-attached storage, or NAS, usually in contexts that leave unclear whether NAS is a kind of SAN, a competitor to SAN technology, or something else entirely.

The term network-attached storage (NAS) is used pretty consistently throughout the industry to denote intelligent storage devices that connect to networks and provide file access to clients, which may range from desktop computers to enterprise-class application and database setters. From the time of their introduction, NAS devices have almost universally used Ethernet interconnects and TCP/IP-based protocols to connect to and communicate with their clients.

As we mentioned earlier, the term SAN is typically understood as including the storage attached to the network. What we didn't say earlier is that today the term SAN is also closely identified with block-access storage devices that is, disk drives and devices such as RAID subsystems that behave as disk drives from the client's viewpoint.

Thus, as the terms are commonly used today:

A NAS device provides file access to clients to which it connects using file access protocols (primarily CIFS and NFS) transported on Ethernet and TCP/IP.

A SAN device (or, as we promised to call it, a SAN-attached storage device) is a block-access (i.e., it is a disk or it emulates one or more disks) that connects to its clients using Fibre Channel and a block data access protocol such as SCSI.

Figure 1.8 illustrates these two storage paradigms.



Figure 1.8 SAN and NAS storage devices

A glance at Figure 1.8 clarifies the major architectural difference between NAS and SAN storage as the terms are commonly used today. In the NAS paradigm, the file system that organizes blocks of storage into objects that are convenient for applications to deal with resides in the storage device. The NAS storage device is responsible for allocating storage space and for keeping clients from stepping on each others' toes as they make (possibly conflicting) file access requests. On the host side of the interconnect, a file access client translates applications' file I/O requests into network messages and sends them to the NAS device for execution.

By contrast, in today's SAN paradigm, the file system is on the computer side of the interconnect. Systemwide storage capacity management and conflicts among client data access requests are resolved by cooperation among the SAN-attached servers. This makes host-side software much more complex than with NAS devices.

By absorbing the file system (and hence the on-disk data format) into the storage device, the NAS model makes concurrent data access by different types of computers easy. In fact, today, NAS is the only widely available way to make the same data accessible to computers of different types.

That doesn't quite tell the whole story. NAS file access protocols are very general and functionally rich. Moreover, they usually connect to TCP/IP-based networks, which are designed to support very general interconnection topologies. Because of their functional richness and generality, these protocols are predominantly implemented in soft ware, which executes slowly compared to the device-level firmware and hardware typically used to implement SAN protocols. Raw data access performance of NAS devices, therefore, tends to be lower than that of otherwise comparable SAN devices, and both client and server processor utilization for accessing data tends to be higher. In simple terms, the trade-off today is, therefore, as follows:

• Choose NAS for simplicity of data sharing, particularly among computers and operating systems of different types.
• Choose SAN for the highest raw I/O performance between data client and data server. Be prepared to do some additional design and operational management to make servers cooperate (or at least not interfere) with each other.

Vendors and users of NAS devices often assert that the networks connecting their devices to host computers are SANs, and they are right. According to the SNIA, a storage area network is any network that is predominantly used to transfer data between storage devices and computers or other storage devices. NAS devices are certainly storage devices and the mix of traffic on the networks that connect them to their clients is certainly dominated by storage I/O in any reasonably busy information processing operation. We mention this to only nip in the bud any possible confusion around this use of the term SAN. One can only hope that over time the vendor and user communities will evolve to a broader and more descriptive common understanding of the term SAN.

Of course, it's reasonable to ask, What about the other possibilities—block-access devices that use Ethernet to attach to clients and file servers that use Fibre Channel? Both of these are possible and, in fact, development is occurring within the storage industry that may result in both types of devices being brought to market. Today, however, almost all NAS devices are Ethernet-attached file servers and almost all SAN devices are Fibre Channel-attached block-access devices. Table 1.1 summarizes the state of affairs in SAN and NAS devices today.

Summary

• A storage area network (SAN) is any network whose primary purpose is to enable storage devices to communicate with computer systems and with each other.
• The key feature that defines a SAN is any-to-any connectivity of computers and storage devices. SANs reduce or eliminate the incidence of unconnected islands of information.
• SANs can reduce the cost of providing information services. In some cases, they can enable new services that it was not previously practical to provide.
• At the current state of this new industry segment, SANs' potential has barely been touched. Ultimately, SANs may lead to heterogeneous distributed systems in which applications running on different operating systems and hardware platforms can meaningfully access the same data.
• Already today, SANs are enabling new ways of providing information services, including an entire new industry segment called storage solution providers, who make quality of service-based access to storage available to their clients as a ser­vice.
• Strictly speaking, NAS devices connect to SANs, although the two acronyms are most often used with mutually exclusive meanings. The key unique feature of NAS devices is that they place the file system on the storage device side of tile network. Today, the only widespread incidence of heterogeneous data sharing is with NAS devices.

Previous Chapter
Chapter Home
Storage Networking Primer Home