Make no mistake, data storage and data recovery are now top of the agenda for any technology-savvy company. From heavyweight financial institutions to small and medium-sized enterprises, the emphasis has shifted decisively towards networking solutions that guarantee integrity, accessibility and recoverability of enterprise data - so much so that storage now represents the largest entry on most corporates' IT budgets.
The proof is in the numbers. According to US market-research firm IDC, the worldwide storage market will be worth $46.4 bn (€43 bn) in 2003. With such hefty sums at stake, it's increasingly apparent that data storage is no longer a closed industry, but one that appeals more and more to telecoms carriers and telecoms equipment vendors as a potential source of revenue.
For the customer, of course, the priority is to ensure that any storage network maximizes the value of an organization's most compelling asset: its data. Those given the task of achieving this, the corporate IT managers, are generally expected to guarantee five-nines reliability, bringing the cumulative network downtime to just five minutes per year. But how can a company ensure that it chooses the most appropriate data-storage solution for its needs? And if it adopts one of the emerging storage technologies, how can it guarantee that the technology is here to stay?
Here and now
Fibre Channel is generally considered the incumbent transport protocol for today's storage networks, and is currently the only viable choice for storing and networking data at the enterprise level. The technology offers a reliable platform for transmitting data over distances of up to 30 km at speeds of up to 2 Gbit/s using a long-wavelength source and singlemode fibre (although it can also run over copper plant across distances of up to 30 m).
Developed specifically to handle data storage, Fibre Channel arose from a combination of other standards - Fibre Distributed Data Interface, Small Computer System Interface (SCSI), High Performance Parallel Interface and Intelligent Parallel Interface. The protocol incorporates the best aspects of point-to-point links and aggregate networks to connect devices via an any-to-any architecture.
The emergence of Fibre Channel as a credible technology in the last four years has transformed the way people regard storage and the value of information, to the extent where the IT world is shifting emphasis towards recentralization and an information-centric computing model. Traditional server-attached set-ups (often referred to as direct attached storage) are now being replaced by, or supplemented with, storage networking architectures such as storage-area networks (SANs) and network-attached storage (NAS).
A Fibre Channel-based SAN is a high-speed subnetwork of shared storage devices, attached to a local- or wide-area network (LAN/WAN). In such a system, the stored data do not reside directly on a single server, as with traditional storage methods, but are dealt with by an entirely separate server. This means that LAN server power can be better used for business applications, releasing network capacity to the end-user.
The partner to SAN, though in the past often wrongly thought of as a competing architecture, is NAS. This complementary configuration offers different strengths and provides access to different types of data. In general, SANs are seen as the tool of choice for high-volume block data transfers, whereas NAS is specifically designed to provide data access at the file level.
In practice, this means that organizations can benefit from using SANs to support mission-critical applications, storage consolidation, back-up sites and high-availability computing, while NAS permits file sharing between multiple users anywhere within the organization. In addition, SANs can be geographically dispersed and linked over Transmission Control Protocol/Internet Protocol (TCP/IP) networks.