In the previous chapter, we talked about how the amalgamation of voice and data into a single circuit is a goal for many organizations' internetworks. As more and more data and traffic cross the network, there will be a larger demand for storage. Unfortunately, conventional means of data storage-disk arrays and tape backups connected to the server-just aren't enough to keep up with the task.
Furthermore, with the moniker "Information Age" comes the burden of storing data in all sorts of sizes and compositions. It wouldn't be so bad if the term "information" were limited to pages of the printed word, punctuated with the occasional photograph. In recent years, however, "information" has evolved to mean rich multimedia content, mixing both graphics and sound. Those graphic and sound files are becoming lengthier and with greater detail, which means larger-sized files. Organizations aren't just storing the data for archival use, either. This information is routinely accessed either internally or served up to Web site visitors. For example, just look at the proliferation of television programs, movies, and music that can be bought online and downloaded to a consumer's computer or media device. The good news is that there's more multimedia on the horizon. The bad news is that organizations dealing with this content need to find a way to store it and serve it as efficiently as possible.
As wonderful as multimedia is, the storage requirements of these files take their toll. Conventional network storage gear is finding itself incapable of maintaining all this information. If that weren't enough, transferring large amounts of data creates its own problems as network bandwidth is continually challenged with other loads.
Having endured for two decades, the parallel Small Computer System Interface (SCSI) bus that has facilitated server-storage connectivity for LAN servers imposes limits on network storage. While SCSI connections are fine for desktop computers and LANs, they are not robust enough for large-scale storage needs. Adding to the limitations of SCSI is the traditional use of LAN connections for server storage backup, which takes away from usable client bandwidth.
Organizations have a couple of popular options when it comes mass storage, including:
Direct Attached Storage (DAS) With this option, an organization's active data is maintained on disk drives that are connected to computer servers. This becomes wasteful when a company has several servers running multiple applications, however, as DAS becomes inefficient: some of the drives run out of space, while others have lots of unused space.
Network-Attached Storage (NAS) With this option, organizations remove storage systems from servers to make better use of storage space and to make networks more modular, making repairs and upgrades easier. For instance, one NAS solution involves a disk drive with hardware that makes the file system available across an IP network. The data is pooled in one location. This makes backup, archiving, and retrieval faster and easier. This is ideal for some scenarios, but NAS performance is governed by the network speed and specific protocols. We'll examine some NAS solutions at the end of this chapter.
Storage Area Network (SAN) This is a more complex and powerful storage option. SANs connect servers and storage on their own networks.
In this chapter, we'll first talk about the generalities of SANs and how you can design a SAN for your own internetworking needs. Then we'll talk about Cisco's SAN solution.