What Kind of NAS or SAN Solution Should You Use?

If you choose to use a NAS solution, you should map out the data requirements of your network. As mentioned earlier in this chapter, make sure you locate resources close to clients . Global data can be replicated to multiple sites, although this is more easily done using a SAN.

Most NAS devices are simple to install and manage. For all practical matters, to an experienced LAN technologist, it's just about plug-and-play. A few screens of installation and configuration information are all that are needed to get a simple NAS appliance up and running. If you are choosing between a NAS and a SAN solution, use NAS when your data resources can be compartmentalized for the most part, and use a SAN solution when more than one file server may need to access the same data at the same time. And if you need to keep up-to-date copies of data at different locations, SAN equipment performs this function faster and more reliably than NAS. In general, NAS is the choice for a small number of clients, whereas SAN is the choice for larger environments that use a much larger quantity of data, and have a requirement of almost 0% downtime.

Don't forget the budget when considering your need to add serverless -based storage to a network. It may be that NAS is a less expensive solution for now, although you project a data requirement available only by a SAN in the future. If your NAS solution involves SCSI devices, you can always use a SCSI-Fibre-Channel bridge to make use of these SCSI devices if you later install a SAN.

At this time it would be hard to justify committing to a SAN on a small scale, based on the prospect that later you will expand the SAN to encompass larger data requirements (or other features of a SAN such as RAID or serverless backups ). This is because of the current environment that hasn't yet decided on precise standards that allow equipment from various vendors to work seamlessly together. NAS devices are similar to file servers and use protocols that are well defined.

Because many functions are implemented in hardware (the HBAs), SAN hardware devices may suffer from a lack of interoperability with hardware from other manufacturers. This situation is changing, and it will be some time before the adoption of the standards process is complete. The Bluefin initiative being advocated by the Storage Networking Industry Association (http://snia.org) is taking a good step in that direction, by providing a basis for creating a single management application that can be used with hardware from various manufacturers.

Lastly, if your need is not for expanded data storage, but instead for providing data availability for disaster recovery purposes, then a SAN might justify its expense in even the smallest of environments. As mentioned earlier in this chapter, the basic Fibre Channel SAN can cover several kilometers to connect two or more SANs. In a metropolitan geographical environment this may be sufficient. All you have to do is connect together several offices you have in a single local environment and create automatic data replication between each site.

For a business that is more geographically diverse, tunneling Fibre Channel frames through a long-distance protocol may be a better idea. By separating copies of your data by great distances, you can prevent a single local disaster from impacting all copies of your data.

In Figure 11.1 you saw that the topologies of NAS and SANs are much different. NAS devices operate similarly to a LAN client or server. As you can also see in Figure 11.1, a SAN uses its own network that is separate from the LAN used by clients of the servers that make use of the SAN.

In Figure 11.2 you saw that an Arbitrated Loop can be connected by a simple hub to centralize wiring. If all of your equipment is from the same vendor, an Arbitrated Loop is a great inexpensive introduction into a SANs environment. Switches, however, are more standardized and provide for a greater capability of exchanging data between SANs and management applications. NAS appliances also offer management applications, but incorporating the two technologies may offer an additional expense.

The main considerations for choosing between a NAS and a SAN solution mainly pertain to cost and data availability. Here are some other considerations:

• The size of your network and the amount of data that needs to be available to file servers or even high-end workstation clients. If the data requirements, as well as bandwidth requirements, are high, then the SAN is probably the best solution.

• If localized access for data is required (such as in a departmental LAN situation), and the data is generally localized to clients on the same LAN, then NAS may be a good solution.

• For data that changes frequently and is global to an enterprise, except for the smallest network, a SAN is a good choice. This provides multiple servers the capability to access the same data, and provide clients with a fast access to that data.

Tip

Both NAS and SAN devices offer the capability to map different file systems to selected devices. This is more prevalent in SAN equipment than in NAS devices, which generally offer one or more file systems, and usually one of those for each device attached to the different NAS devices on the LAN. SAN solutions, however, usually allow you to partition RAID sets into separate volumes , each of which can host a different file system.

Thus, using a SAN, you can more easily allow many different hardware/operating-system computers to access data on the SAN. And because many non-Windows operating systems (such as Unix/Linux) support reading FAT and FAT32 file systems used by older Windows systems, data interchange can be a simple affair. With the use of the Common Internet File System (CIFS), non-Windows clients, such as Unix/Linux platforms ( especially those using SAMBA ”www.samba.org), can be incorporated easily into a Windows environment.