Network Storage

• Identify the main characteristics of network attached storage (NAS).

Data storage and retrieval are fundamental services that networks are required to provide. A network can provide this functionality many different ways; the one you choose will depend on many factors, including the organization's budget and data storage needs. This section explores some of the different data storage choices, including traditional file server storage, NAS, and storage area networks (SANs).

Traditional File Server Storage

Perhaps the most widely used type of network data storage involves using an existing server to store data. This approach is commonly used by small organizations that can't afford a more costly solution or that don't have a volume of data large enough to warrant a more advanced and costly solution.

A traditional file server typically uses a full-featured network operating system and may or may not be dedicated as a file server. In many network environments, the server provides numerous network services, including print services, authentication, and even proxy or firewall service.

You will probably be working with traditional file servers, as these are prevalent in almost every company. Figure 10.8 shows a traditional file server solution.

Network Attached Storage (NAS)

As the storage needs of organizations of all sizes increase, new technologies must be developed to accommodate those needs. One strategy for increased storage needs is NAS. Essentially, NAS is a specialized file server that is designed specifically and used exclusively as a file server.

NAS is a box that typically has no mouse, keyboard, or monitor, and that has a very streamlined operating system. Several hard disks and perhaps a tape drive hold the network files. Additional hard drives can be added to the NAS to increase storage capacity. Clients typically access a NAS over an Ethernet connection; each NAS box is seen as a node on the network and requires an IP address.

Because NAS systems are not attached to a server, they must be able to communicate on the network by using an application protocol designed for file access. Most commonly, the Network File System (NFS) protocol or Server Message Block (SMB) protocol is supported by NAS devices for this purpose.

Because NAS is dedicated to data storage, it's a more costly solution than providing data storage from an existing server. However, the advantages of NAS over the traditional file server are compelling. First, NAS can provide higher data availability; as a dedicated device, the NAS is less likely than a traditional file server to be brought down or to crash, disabling file access. Second, a NAS system is more secure and less susceptible to security attacks than a traditional file server. Third, and somewhat subjectively, a NAS system offers easy administration. Figure 10.9 shows an example of a NAS.

SANs

A SAN is a redundant network that utilizes servers and various storage devices, allowing all data to be stored and accessed very quickly. However, whenever the word redundant is used with networks, you know you are facing a costly solution.

Users typically access a SAN by using a fast interconnection technology called fiber channel. Thanks to fiber channel, a SAN can support transfer rates of up to 1Gbps. A SAN consists of servers that share and store data on RAID arrays or tape drives. All the devices are connected via fiber channel and offer very high-speed data transfer.

Creating a SAN can be as simple as adding a few new servers and storage devices to an existing network, using fiber channel adapters. As the SAN grows, you can add fiber channel hubs and switches for more efficient utilization. Because all the servers and storage devices have alternate paths to each other, a server failure does not make data unavailable. The data is distributed across all the servers and is not server-centric.

SANs are relatively new; as of this writing, many aspects of the technology do not have standards. SANs are an excellent way to handle large amounts of data, but buyers should beware. It is generally accepted that the best way to implement a SAN is to hire a SAN vendor and purchase a turnkey solution. Interoperability problems exist between SANs, and each has unique peculiarities.

Currently, the single biggest disadvantage of SANs is currently their cost. Successful SAN implementation tends to be expensive and involves many people as well.

As the amount and utilization of stored data increases, other technologies must assuredly surface to allow us to continue to maintain and protect data. SANs are likely to be an influence in major storage situations, especially as standards begin to be developed for this technology.