NAS catches on
- By Edmund X. DeJesus
- Aug 01, 2002
The IBM NAS 300 is a dual-processor system with several levels of RAID and up to 6.61T of capacity. It's priced at $115,000.
Network-attached storage meets your needs quickly, simply and inexpensively
BlueArc Corp.'s Si7500 holds up to 250T of data and has several levels of RAID. Its price starts to $60,000.
Data accumulates; it's practically a law of nature. And as an agency's data grows, so do the need to store it and the complexity of accessing it.
Network-attached storage offers distinct advantages in getting a handle on all that data, including straightforward access, easy implementation and relatively low cost.
The idea behind NAS is simple. Just as you can connect main servers and desktop PCs to a LAN, you can also connect storage servers: devices whose sole purpose is to store large amounts of data in a small space and provide that data over the network as needed.
This arrangement has many pluses. Client computers can access storage without making demands on the main server, freeing the server's processing power for other purposes.
You also can move shared files from desktop PCs to the NAS, simplifying and speeding access for other users. In fact, users can get at data even if the main server is down.
Because users can access storage directly, without using the main server, network traffic decreases. And because a NAS device runs a simpler operating system than does a main server, it can often deliver data faster.
NAS devices typically use a type of RAID for protection. It provides redundancy if a single disk goes bad.Skinning a cat
You might also be able to reduce certain file-transfer operations with NAS. For instance, if you're e-mailing a multimegabyte file to many users on the network, you can store one copy of the file on the NAS and provide a simple link to it for the recipients.
NAS is particularly useful in environments with diverse operating systems. This is common in government agencies that need special computers for particular purposes or ones that have merged with other networks. NAS devices typically work with multiple platforms, usually including Windows, Macintosh, Unix and Linux.
In fact, for agencies stuck with a hodgepodge of systems, the ability to bridge OSes could be the most important aspect of NAS.
NAS devices can save money. They build on and leverage existing network infrastructure and administrator skills. They don't require special protocols, and typically run over standard TCP/IP networks.
NAS devices basically perform only one task, so they are reliable, reducing problems for administrators and users. Any failures that occur can be fixed rapidly and easily, often by hot-swapping disks. Their simple design also means low up-front cost.
Adding storage is usually quick and simple, with no need to take down the server and inconvenience users, or require staff to work off-hours to make changes.
Almost any networked agency can benefit from NAS, but storage-intensive applications clearly stand to gain the most. For example, apps that use many images'computer-aided design or scanned paper forms, for instance'would gain performance by placing those images on NAS servers.
Similarly, multimedia and document management applications also would profit from NAS. Agencies with classified or confidential data could use NAS to deposit information in a secure single location, with centrally managed security and permissions. The high reliability of NAS is also attractive to agencies that must provide always-on services.
If your agency is considering NAS, there are a few features to keep in mind, depending on your priorities. One is support for the OSes on your network. Another is ease of use for administrators: It should be simple to employ with no learning curve. Many NAS devices can be remotely maintained via the Web, for example.
The NAS device also should be simple to install: plug in, attach and go. And capacity is important: Can it support your users and applications now, as well as those anticipated for the future? And is it easily scalable?
Agencies might also have requirements to meet specific goals. For example, you might need a certain level of performance to satisfy users or project objectives. Or, your security requirements could require definite features.
Many agencies can take advantage of NAS to improve their data storage, but there are some situations where NAS is not the best idea.
For instance, the word 'network' in NAS is especially pertinent. If your agency has several networks with cross-network communication, NAS probably won't help much and might even hurt. The traffic created by accessing NAS devices across different networks will reduce performance.
If you're dealing with special platforms, such as mainframes, that either don't participate with the network or that the NAS can't handle, you're also stuck. A common problem is if a server needs to access a megadatabase. In such a case, going through the network creates a bottleneck.
For agencies with large storage demands, NAS might not suffice. You can add NAS capacity only up to a point before traffic throughput suffers. When that happens, it's time to consider a storage area network.
NAS does have disadvantages. For example, many NAS devices offer various levels of RAID, but low-priced units could lack fault-tolerance. And the limits on NAS capacity could be quickly tested by a large organization. Naturally, the speed of data transmission is limited by the LAN itself.Edmund X. DeJesus of Norwood, Mass., writes about IT.