STORAGE: The Network-Attached Server

As

the network administrator for a workgroup LAN, consider this scenario:

once again, one of your network file servers is running out of storage space.

You send out the usual broadcast messages for users to delete old files, but you

know you haven’t eliminated the problem, only postponed solving it. It’s

time to rethink your options for increasing disk storage space.

If your network server has an available drive bay, you might consider adding

a hard disk. This sounds economical. But you’ll need to take the server off

line, and you’ll hear from your users about that. After totaling up the time

and effort it takes to reconfigure the server–and probably sacrificing part of

your weekend–it may not add up to much of a bargain.

You could add more network storage with another PC server. That would put a

bigger dent in your budget, especially after purchasing the usual operating

system software licenses. Plus, it will take a day or more of your time to

configure the new server and integrate it into your existing network.

Or you could consider a third alternative: a new network file-sharing

technology called network-attached storage (NAS). NAS file servers expand

storage capacity independent of the PC server and provide a simple,

cost-effective solution to adding network storage.

Traditional storage: Shortcomings

Traditionally, organizations have addressed the need for more storage by

adding disk drives to an existing server or installing a new general-purpose

server on the network. Both solutions can be time consuming for the IT

professional and disruptive to the workgroup, since the server must be taken off

line. To minimize such disruptions, hardware upgrades are usually scheduled for

off-hours, requiring the IT staff to work evenings or weekends. However, in

environments with demands for high data availability, there are few or no

convenient time periods for users to be without access to their network

resources. In addition to the user inconvenience, any time spent adding storage

is time that the IT professional is not available to deal with other network

problems.

After hardware, software licenses and installation time are considered, a new

server for the workgroup can cost anywhere between $3,000 to $10,000. It may

take a day or more to install. Even after the new server is installed and

tested, problems may occur once the system is on the network and put under load.

A new disk drive may be relatively inexpensive, but installation can take

several hours to half a day. As demands for data storage continue to grow,

traditional methods of adding storage to the workgroup have become costlier.

Next generation storage

The PC has paved the way for network server technology. Over the last two

decades, PCs have evolved into general-purpose network servers responsible for

managing a number of complex functions including storage, application delivery,

printing and gateway functions. All activity is routed through this

general-purpose server, taxing server resources and, often, creating a

bottleneck. Throughput and response time suffer as a result.

In an attempt to alleviate these conditions, system administrators have begun

to modify general-purpose network servers to deliver a limited set of

specialized functions. One such specialized need is file sharing. Many PC

servers are configured with a large amount of storage and dedicated as online

file repositories. Because of the general-purpose nature of PC server design,

its deployment as a network file server carries unnecessary and unused

components, not to mention the IT resources needed to fine-tune a multi-purpose

network operating system for a specific function.

NAS file servers were developed as an alternative to the general-purpose

server. They are single-purpose appliances, dedicated as file servers, which

provide a flexible and inexpensive alternative for adding storage to the

workgroup. NAS servers connect directly to the network, not to the file server,

giving rise to the term "network attached". Workgroup users cannot

distinguish between accessing files on a general-purpose server and on a NAS

file server. The only difference a user might notice is improved network

performance and the ability to access files even if the general-purpose server

is taken off line.

Unlike traditional servers, whose operating systems charge a "per

seat" licensing fee for each user connected to the server, there are no

licensing fees with NAS servers. The system software is embedded, and software

upgrades are usually free via Web download. With a NAS server, you pay for only

those hardware and software components that are needed for network storage.

Choosing a NAS file server

To meet the workgroup’s increasing demand for data storage, NAS file

servers are rapidly growing as a product category. They are easy to install and

maintain, and they are more economical to own and operate than traditional,

general-purpose file servers. Because these products are relatively new, it may

not always be clear how to evaluate them compared to traditional storage

options. What product features are important when choosing a NAS server? What

benefits should you expect? The following criteria will help you evaluate and

select the best product for your workgroup LAN.

Easy installation: Adding a NAS file server to an existing network should be

as simple as "plug it in and turn it on". The default installation

should take no more than 15 minutes. The NAS server should be easy to install

and integrated into an existing network without disrupting other servers, and

there should be no additional server or client software to install. Network

users will have access to all their files and network resources during the

installation process.

Look for a solution that is operational in no more than three steps:

• Plug in the power cord

• Connect the
  
  Ethernet cable

• Turn on the
  
  power.

Compatibility with existing networks: To ensure easy

integration into an existing network, choose a NAS server that is compatible

with your established networking protocols and workstation types. Check to see

if the server can respond to a dynamic host configuration protocol (DHCP),

bootstrap protocol (BOOTP) or reverse address resolution protocol (RARP) server

to receive an automatic Internet protocol (IP) assignment. If you opt for manual

IP address assignment, look for an easy-to-use installation wizard that will

help you step through the task quickly. The better a NAS server supports your

existing workgroup environment, the less time and effort it will take you to

configure and integrate the new NAS server at installation.

Intuitive Web administration: Since a NAS server is

specialized for a single function, it requires limited maintenance. Any routine

administration or customization of the NAS server is easily performed via a

simple Web browser from any client on the network or from anywhere on the

Internet. While evaluating a NAS server, check for an intuitive, easy-to-use

graphical user interface to the Web browser.

Cross platform file sharing: Cross platform file

sharing can be a headache for the IT manager in workgroup environments where

there is a mix of Windows, Apple Macintosh and Unix or Linux workstations

requiring support for Microsoft, Novell, Apple and Unix networks. To address

this problem, some NAS file servers have been pre-configured to support more

than one network protocol for cross-platform file sharing. A NAS server should

appear on the network like a native file server to its clients with files saved

and retrieved in their native file formats. To protect your investment in

existing hardware and software, choose a NAS server that is pre-configured to

provide concurrent support for those network protocols and workstation types

already in use in your organization. Conformance to industry standards and

compatibility with existing equipment will ensure an easy installation and the

long-term use of the NAS server.

Network security: If security is a concern on your

network, choose a NAS server that integrates into your existing network security

by supporting NT Domain or NetWare Bindery for passthrough security. Integration

with NT domain security for passthrough user authentication provides a superior

solution over password security for network shares.

Compatibility with backup policy: After storage and

file serving, backup is the next most common use for a NAS server. There are two

types of backup usage that should be considered during your evaluation: using

the NAS server as a backup device on the network for workstations files; and

using the NAS server as part of an archival solution for enterprise backup.

Check to see if the NAS vendor offers a solution to back up workstation files

onto the file server. In addition, the NAS server should be compatible with

leading network backup application software used, or anticipated, in your

organization.

Improved performance under load: A NAS file server,

when all resources are optimized for the single task of file hosting, can offer

superior performance over a general-purpose server. Because a file server is

primarily I/O bound, criteria traditionally used in measuring PC server

performance may not always apply. For example, processor speed and memory do not

create as great an impact on a file server as they do on an application server.

When a file server is under load, data throughput is the most important

measurement. Investigate how many concurrent users the NAS server has been

routinely tested to support. A file server should be able to comfortably support

thirty or more concurrent users in a workgroup of average size without

experiencing a severe degradation of response or drop in network connections.

When performance is an important consideration, choose a product that supports

advanced disk configuration such as disk striping (RAID 0), where two or more

disks are combined into one large volume and the data striped across the disks.

High reliability: Single-purpose NAS servers are more

reliable than general-purpose servers that are designed to handle diverse tasks

are. With NAS servers, all unnecessary components have been removed, resulting

in a streamlined, highly integrated product. A higher degree or integration

enhances the NAS server’s ability, and fewer components reduce the risk of

failure. The embedded operating system has been optimized for file system I/O

and networking.

When reliability is important, look for NAS file servers that

support data redundancy features like disk mirroring (RAID 1) where the content

of one disk is duplicated onto a second disk. Data striping with parity (RAID 5)

is an important feature to look for in NAS servers configured with three or more

disks.

With NAS servers, users have access to their data even if the

general-purpose network server is unavailable. This in turn makes the network

more reliable, reduces the load on the general-purpose server, and minimizes the

exposure of a server outage.

Portable storage: Using the NAS server as a portable

or temporary storage device enables a number of new applications. For example,

the NAS server can be installed temporarily on the network as a scratch server

while the general-purpose server is taken off line for maintenance. Or the NAS

server can be used to store large graphical presentations and easily transported

and installed at trade shows, customer sites or remote offices. If the NAS

server is being considered for applications like these, look for a lightweight

device with a compact footprint.

Courtesy: Quantum Corp