The NAS solution: managing the explosive growth of data volume on business networks - Tape/Disk/Optical Storage

Computer Technology Review, June, 2003 by Wayne Arvidson, Chris Romoser

Managing network complexity is a perennial IT challenge. Like weeds in the garden, network problems can creep in whenever a business has to add hardware, tinker with the network architecture, or make ad hoc changes to accommodate new software.

These days, the growth in network traffic within most companies, and across the Internet, means that network administrators have to deal with new storage requirements. They need manageable, cost effective capacity. They want minimal disruption to users. And they want storage solutions that integrate cleanly into existing networks. Simple solutions that also simplify their network management tasks are worth their weight in gold, especially for smaller shops with limited resources and technical support.

The desire for low-cost, highly integrated, easily managed network storage is the reason that network attached storage (NAS) servers are the choice of more and more small and medium-sized companies. Last year, unit sales of midrange NAS servers in the $5,000 to $25,000 range grew by 58.4 percent; analysts project 25.6 percent compound annual growth in the category through 2007 (Gartner Dataquest, "Network-Attached Storage Market Share 2002," March 2003), driven by consolidation of traditional file servers, ease of scaling, and cost effectiveness.

In terms of simplicity and manageability, NAS servers are the dream solution for small and medium-sized businesses. NAS servers cost less, are easier to install, and simplify overall network management. Yet today, NAS servers are installed in only 12 percent of sites that have less than 100 employees and 15 percent of sites with 100 to 499 employees (Harte-Hankes study of SMB market potential, April 2003). What are the rest of them doing? And why do they need to consider NAS?

Network Capacity Requirements are Growing

While they may not all agree on the solution, IT people will certainly agree on the scope of the challenge. Corporate e-mail, which is now a critical communication channel in most businesses, has proliferated with rich content that includes images, presentations, spreadsheets, PDF files and more. Office application file sizes have mushroomed, embedded with high-resolution graphics as well as audio or video content. Internet access has driven explosive growth in the size of corporate websites, which now include customer support databases, file downloads, marketing materials, and online stores (so-called "e-tailing"). Multimedia content editing has entered the mainstream.

Storage Capacity: One Big Problem, Many Possible Solutions

Business systems are increasingly designed around the network accessibility of data, stored on the network and accessed as needed via browser-based interfaces. Workgroup activities have evolved to exploit the pervasiveness and accessibility of the network, providing print services, file sharing, WAN and Internet access, and an increasing range of corporate services ranging from document control to accounting to ISO 9000 training. In short, a lot of data is surging over the network. It all has to be stored somewhere.

Running short of disk space is like running out of oxygen; network applications need room to breathe. However, some of the more common approaches SMBs use to add storage--such as adding a new hard disk to a general-purpose server, or adding a new general-purpose server for a particular task-- can be disruptive and time-consuming. It's not just that it takes hours to configure a new server or install and configure a new disk. And it's not just that users are aggravated when their server goes offline while a new disk is installed. It's that the old system was working before, and now--with some careful manual configuration--it's been taken apart and fixed. Odds are that the fix will work, but the new hardware may also make the network more complicated to administer than before.

The Hard Disk

Adding a hard disk to a general-purpose server solves the immediate storage crunch, but doesn't address all the issues of data growth. The hardware cost of adding the hard disk may be lower, but the added network load on the server will reduce its performance in other roles. Backup and recovery routines will also be less manageable as disks are scattered among multiple servers.

There are other drawbacks. Hard disk space may go wasted on some servers and be in short supply on others, stranding unused capacity on the underutilized server and forcing the administrator to do some drive remapping. Even the disk upgrade that goes well is an added headache, both for the network administrator who installs and configures it and for the business that has to schedule the hours of server downtime. The disk upgrade, which was originally intended to add lowcost capacity, may inadvertently cause an increase in management cost and complexity.

The General-Purpose Server

Adding a general-purpose server can help lighten the load of existing servers, but it also entails extra cost, complexity, and effort. The operating system and disk subsystems will not be optimized for file serving. Many of the server's hardware and software features will go unused. Its OS and applications will require software licensing fees based on the number of users. And after a time-consuming installation and configuration, costing the administrator a day or more, the general-purpose server will also add to the organization's ongoing support costs and the complexity of the network.