A NASty Job, Part 2

by Wayne M. Krakau - Chicago Computer Guide, March, 2001
Last month, I detailed some possible candidates for independent NAS (Network Attached Storage) devices. Now, I'll tell you what they are. This is a description of a generic NAS device, so there will be exceptions in some of the details.

In its most basic form, a NAS device is a disk drive in a little black box with a network socket, bundled with intelligence. Traditionally, it has just enough intelligence to control its own disk drive and not much else. It is dependent upon the existence of a traditional file server on the same LAN. Without that conventional file server with its conventional NOS (Network Operating System), the NAS simply wouldn't function. The file server does most of the management of data and security, while the NAS's own brain just concentrates on shoving data down the pipe (that being 10/100 ethernet connection).

In the background, it would really have more intelligence than would be readily apparent, in that it would pretend to be whatever brand of file server it found on the network. It would also have to act cooperatively with the server to allow for security. NAS devices can mimic several different NOSs. Some can even have multiple identities at once, depending upon what type of client is making the request for data. All of this does take a certain amount of computing horsepower.

While there is no law requiring it, almost all NAS devices use a modified Linux kernel to harness that computing horsepower. The kernel is the base, underlying core of an operating system. It's the raw, stripped down operating system without any extras. Each manufacturer modifies the Linux kernel to suit their particular needs. Even different models within a single manufacturer's product line may have different versions of the Linux kernel.

For those of you who have been lost in the Arctic for the last couple of years (remember, penguins are in the Antarctic - and a few other places, like the Hilton in Honolulu - but not the Arctic), Linux is a modern derivative of UNIX (itself a 30-year-old operating system) that was put into the public domain (with certain limitations) by its developer. Since it is effectively free (if you follow the rules), manufacturers love to use it as the embedded operating system for things like NAS devices. Just look for the penguins at any trade show to get more details. (Disney has its mouse and Linux has its penguin.)

To make life easier, and to avoiding requiring customers to have Linux experts on staff, a browser-based (Internet Explorer, Netscape Navigator, etc.) menu system is used to manage the NAS device. None of the systems that I've seen could really be called easy to use (like their advertising says), but they are at least useable. It's not hard to tell that they were written by ex-Unix, now Linux geeks.

These programs generally remind me of the directions to an electronic appliance that was designed overseas, where the directions have been translated into English by someone who learned English as a second language without ever visiting an English-speaking country. Nobody bothers hiring a native-born English speaker to proofread these directions, and nobody at the NAS manufacturer hires an interface design specialist to proofread their management program interface.

At least the interfaces are usually simple enough that the common day-to-day tasks can be taught to an interested nonprofessional (in computing). In many cases, a talented amateur can even get the system initially configured, depending upon how close the customer's needs are to the system defaults.

NAS devices have different models offering different levels of speed and safety, just as file servers do. The simplest and least expensive models have a single disk drive. Next up the scale are models using two mirrored drives. At the top of the scale are models offering a full RAID (Redundant Array of Independent Disks) system, usually RAID Level 5. Within a given model, you can usually configure the disks in whatever way you want, assuming you have enough disks, a minimum of two for mirroring and three for RAID 5. You can trade the extra speed and safety for additional storage.

Finally, after selling dependent NAS devices for a while, manufacturers realized that by simply reprogramming them, using the flexibility of the Linux kernel, they could make them independent of traditional file servers. That was the big breakthrough.

Since they were reprogramming them, anyway, some manufacturers added extra features. Some added SCSI ports to attach shared devices such as CD towers. Others added special functions that turned their products into Web servers, e-mail servers, asynchronous communications servers (modem sharing), and combination router/firewalls. All of this is in addition to their new roles as stand-alone file servers. In addition, they never lost the ability to work cooperatively with conventional file servers.

Next month I'll cover the benefits and limitations of using independent NAS devices. Now, I'm going to see if I can find even more obscure sets of technically-oriented acronyms to further confuse the issue.

©2001, Wayne M. Krakau