The case for network storage appliances

Potential benefits include centralized management, data protection, and support for heterogeneous platforms.

By Daniel P. Petrozzo

Click here to enlarge image

In the traditional design model for most corporate computing environments, storage is considered a peripheral element to the infrastructure and is typically implemented as a dedicated resource of a host computer. Until recently, adding storage meant adding new storage devices, which was acceptable because changes in capacity requirements tended to be predictable and gradual.

In the traditional model, technological barriers made sharing of resources impractical because islands of data were distributed around an enterprise. The technological barriers, combined with the security and administration preferences of IT organizations, reinforced the model of host-based storage.

However, with the advent of e-business and other storage-intensive applications, IT strategists generally agree that traditional storage approaches no longer are adequate for a number of reasons:

  • An exploding demand for information to be made available across the enterprise, regardless of where it may physically reside.
  • Competitive demands on enterprises to leverage information to improve productivity, customer service, and decision making.
  • Senior-level executive pressure to make company information work by managing it more efficiently and effectively.
  • Global business requirements for 7x24x365 network availability and business continuity.
  • The growing shortage of skilled IT profes sionals to handle the increasing complexities of heterogeneous computing environments.

To proactively address these challenges, enterprises are now taking a broader, more strategic view of storage and are building infrastructures to support global networked business.

Beyond the traditional: NAS and SAN

The elevation of storage from a peripheral component to a primary focus requires a re-examination of the traditional design and implementation of storage services. The intensity of focus on storage grows as IT spending shifts from servers to storage. Based on IDC and Dataquest estimates, sales of network storage products are expected to increase from $4.4 billion in 1999 to $19.8 billion in 2004-a 34% compound annual growth rate.

In response to the challenges of globally networked business, the storage industry has developed two alternative approaches that separate storage from servers: network-attached storage (NAS) and storage area network (SANs).

NAS attaches directly to the LAN using traditional protocols and serves files to any host or client also attached to the LAN. Because of its simplicity, it is a logical approach for storing files that must be shared by standard office productivity applications and executed on desktops.

A SAN is a centrally managed, high-speed storage network that provides a link between users and a single pool of storage, enabling faster and more productive access to data throughout an enterprise. A SAN consists of a separate network of storage devices linked by switches, usually using the high-speed Fibre Channel protocol.

Both NAS and SAN offer benefits. But in practice, today's implementations fail to address many of the most important network storage issues. While users like the simplicity of NAS, they're concerned about its scalability, availability, and functionality. And most current SAN solutions are actually single storage-array frames that restrict IT organizations to proprietary products, preventing administrators from choosing best-of-breed products from multiple vendors.

What's needed is a network storage solution that simplifies the complexities of storage re-source management.

Network storage appliances

Increasingly, storage integrators and users are turning to SAN or network storage appliances. These devices give users a single platform from which to manage, protect, and share critical business information from anywhere on the network, regardless of the host, storage, or network infrastructure. Users can simply plug a storage appliance into the existing infrastructure to manage rapidly growing volumes of data.

A network storage appliance depends on three inter-related components: hardware, software, and services. Or, put another way: appliances, applications, and people. By introducing this level of intelligence, a storage appliance promotes a number of functional advantages not currently possible with SAN or NAS implementations.

Universal access and sharing

One of the most valuable capabilities of an intelligent storage infrastructure, enabled by a storage appliance, is virtualization. Through software, a storage appliance can overcome the physical constraints normally encountered in host-based storage by allowing any or all storage devices to appear as a single pool or by dividing the "virtual" pool into smaller units to be assigned to specific users or groups. This functionality is independent of the host and storage systems.

The simplified, centralized storage management provided by virtualization forms the basis for easier storage management, which relieves many of the challenging and costly demands of storage administration.

Improved security

With the benefits of networked storage comes the challenge of maintaining security and proper data-access controls. Uni-versal access is beneficial, but providing access in a controlled way to only those authorized systems is a requirement. A network storage appliance can play a pivotal role in providing the right levels of security via software.

Support for growth

Essential to supporting business continuity is the ability to create exact replicas of stored data in real time among storage systems-commonly called "data mirroring." This function provides quick recovery

and ready access to business-critical data. Data mirroring also enables data migration, which is the movement of data from one device to another without disrupting applications.

Smarter backup

A point-in-time image, also referred to as a snapshot of data, can be made instantly to enable off-loading of functions and traffic from hosts. The point-in-time image can be used for applications such as server-less backup, data mining, and systems testing-with minimal impact to host systems.

Increasingly, enterprises are looking for better ways to manage the complexities of huge volumes of business-critical data, making that data more readily accessible for operational, strategic, competitive, and profitable uses. Net-work storage appliances accomplish this goal by wrapping the essential functions of storage management into a single package-one that accommodates NAS and SAN architectures.

By integrating the hardware and software required to intelligently move and control information across the network, storage appliances can help reduce deployment time and costs, enabling organizations to turn the explosion of data into a competitive edge.

Daniel P. Petrozzo is president and chief operating officer at StorageApps Inc. (www.storageapps.com) in Bridgewater, NJ.

This article was originally published on November 01, 2000