As government agencies increasingly rely on information systems, they generate vast quantities of data that must be reliably stored and easily accessible by numerous users.

Experts have predicted data volumes will grow by at least 50 percent a year thanks to electronic commerce, data warehouses and information-intensive applications focused on such things as enterprise resource management, multimedia and e-mail messages.

Network-centric data storage has doubled annually for the last several years, and will continue this type of growth for several more. By the end of 1999, demand for network data storage will approach 60,000 terabytes, a quantum leap from the 5,000 terabytes needed last year.

Major vendors, such as Compaq and Dell, recently announced enterprise storage-system strategies, signaling their entry into the network-storage arena.

As volume increases, keeping information both accessible and safe is difficult. Every organization needs a comprehensive data-storage strategy that meets their daily activity and backup requirements.

Most government agencies look for products that support a wide range of applications with varying performance and capacity requirements; have high reliability, enable installation of additional storage without further straining general-purpose servers or requiring network downtime, provide faster recovery and high-performance backup solutions; and deliver cost-effective fault-tolerance with RAID storage subsystems.

RAID Your Data

A Redundant Array of Independent Disks (RAID) system is the most common way to protect network content. Investment in a RAID system typically starts at more than $10,000 but pays dividends by keeping applications and data available to network users. While there are six formal RAID levels, 0 through 5, many vendors offer proprietary RAID configurations with levels 7 through 10, or higher.

Today's RAID storage systems often sit outside the server, and can connect to more than one server to keep network storage available. However, adding a RAID system can double the cost of a server, making agencies confront cost and benefit decisions.

The backup and restore process comes down to two elements: the media and software. Despite advances in CD technologies, nothing has replaced tape for economical storage. However, even tape media decisions involve balancing cost, capacity and speed.

A tape backup device is still the simplest defense against data loss. Costs, capacities and performance vary. Picking the right tape backup device is as important as choosing a hard drive and a RAID system for your server.

Finding the Right Architecture

Network Attached Storage (NAS) servers are increasingly popular options with IT managers struggling to handle massive data growth. They relocate data-storage tasks from traditional servers to physically separate boxes, and can be plugged into a network where needed. They are highly scalable and can save data to clients of varying operating systems and platforms. In contrast, hard drives have limited scalability and operate only in a homogenous environment. NAS devices range in size from 4 to 6 terabytes.

One major shortcoming of NAS servers is that the network on which they run is also used for data access by clients to retrieve data from the file server or to communicate with an application server. The data movement between the disk and tape servers also goes over the same LAN. An increase in users could create a network bottleneck.

To overcome some of the shortcoming of NAS, another core technology architecture, Storage Area Network (SAN), is finding its way into enterprise data centers. Being a separate dedicated network, it avoids any traffic between clients and servers. A Fibre Channel-based SAN combines the high performance of an I/O channel and the connectivity (distance) of a network.

Adapting SAN technology through use of Fibre Channel and hubs and switches allows high-speed connectivity using a separate network infrastructure. SAN enables concurrent access of disk or tape arrays by two or more servers at high speeds across Fibre Channel, providing souped-up system performance.