Storing and managing data is an important business need for an enterprise. P. P Subramanian shares his thoughts on the various storage alternatives enterprises have

According to the National Storage Industry Consortium, the amount of data stored on the world’s network is now 1,000 Petabytes

The ultimate goal for IT organizations is to view and manage all its storage and share data across all types of platforms. But to access the same data on the same disk is an elusive goal

Storing and managing data in an enterprise is not just a technology need but an important business need as well. It's imperative to keep abreast of storage mechanisms and techniques to deal with the problem of managing storage. There is a deluge of technologies to choose from. And a whole bunch of new technical jargon that goes along with it. This jargon doesn't make it easier to understand new technologies.

Out of the many storage technologies that have emerged, let me throw light on the confusion created by SAN (Storage Area Network) and NAS (Network Attached Storage).

Storage innovations
SAN is not a new technology. It is one of the two storage innovations that have emerged in response to the ever-increasing demand for data storage (the other being NAS). SAN and the less costly NAS capitalize on low storage costs and fast network speeds. They rely on high-speed networks to build fast and highly available systems to store and retrieve large amounts of data.

The large demand for storage arises from the increased use of Internet and Intranets, applications like data warehousing and video, and the potential of e-commerce. According to the San Diego-based National Storage Industry Consortium, the amount of data stored on the world's network is now 1,000 Petabytes (or 1 million Terabytes).

Just like variety of mobile technology options create confusion among mobile phone users regarding which technology is the right one for them, people readily interchange SAN and NAS, creating confusion and slowing adoption. The confusion partly arises because the two technologies share many similar features.

Both are used for scaling storage capacity and performance, and can be expanded step-by-step as the need for greater storage arises. Both topologies enable companies to purchase storage separately from servers, which means organizations can tailor their purchases to their needs. Both are capable of making data available to a variety of users across multiple OSs. Both are highly available. And a thing to note is that the availability of stored data isn't necessarily dependent on the availability of an organization's multi-purpose server. Both can reduce long-term operating costs by offering centralized storage management. Organizations don't need separate storage for data generated by Unix systems, NT systems, Novell servers, and other proprietary networks.

The differences
But there are significant technical and cost differences. The key element of a NAS system is a dedicated 'thin' server that manages the storage and flow of data across the network between clients and an array of disks. NAS relies on standard LAN/WAN connections. And since it supports multiple communications protocols, it can store data from a variety of heterogeneous servers.

A SAN on the other hand, is a dedicated, high-bandwidth, fiber-channel network that connects storage disks to one or more servers and incorporates SCSI, fiber-channel or ESCON (Enterprise Systems CONnection A mainframe technology) links. SANs are typically more complex and difficult to operate in a heterogeneous environment, but they are highly reliable and offer greater speed and better integrity of data.

The two technologies are targeted at different applications and environments. NAS is emerging as a popular choice for organizations of all sizes that want rapid access to files across heterogeneous servers. SANs are more widely deployed in highly specialized applications like multimedia and video, where the volume of data storage is very large. It can also be used in business-critical environments, where redundant and highly available storage is crucial. SANs are faster than NAS systems.

NAS systems are generally less costly than SANs. NAS systems are used where budgets call for systems in the sub-$10,000 range. SAN installations, including disks and fiber-channel networks, start at $25,000.

Both technologies are expected to gain a wide following. Market research firm IDC estimates that SANs, including mainframe ESCON and Digital VAX VMS-type SAN systems (Compaq proprietary), will generate $11.4 billion in revenue in 2002, up from about $3 billion in 1998. Dataquest estimates that the market for NAS appliances, as the thin servers are called, will reach $10.5 billion in 2002, up from $900 million in 1998.

Some businesses view NAS as a stepping-stone to SANs, which are more difficult to operate in a heterogeneous environment. Yet while SANs are faster than NAS systems, NAS is not exactly slow.

Pre-assessment for better understanding
Ultimately, the goal for IT organizations is to view and manage all its storage and share data across all types of platforms. But enabling multiple users to access the same data on the same disk is an elusive goal. This is because NAS uses file-system protocols like NFS (Network File System) for Unix and CIFS (Common Internet File System) for NT to access and share data across systems. In a SAN, the process of sharing data across multiple OSs is a more complex proposition. Whichever technology is used, a careful assessment of the pros and cons of both is required.

SANs provide high-speed access via fiber-channel links to large stores of data like data used by digital video and audio applications. They are designed for applications where data integrity, availability, and reliability are a must. Mirrored sites, for backup and recovery, can be located up to 10 Kms apart. SANs use reliable SCSI-type connections, not network protocols such as NFS, FTP, HTTP and CIFS.

The high-speed features come at a steep price. SANs cost more, with installations starting at $25,000 and rising steeply, depending on the size and complexity of the solution required. SANs are also usually limited to LANs due to the cost of running fiber-channel links over longer distances.

Data sharing in a SAN is a difficult proposition which requires that each OS has knowledge of the others' data and metadata formats. SAN installation is typically more complex than NAS installation and may require custom configuration.

On the other hand, NAS is a simple and elegant solution for fast, wide-ranging access to data storage. It requires few or no changes to many organizations' existing networks. NAS enables access to data from multiple servers with different OSs. It's relatively inexpensive you can get disks and a thin server for $10,000. The thin-server technology can be distributed across a large network, but centrally managed. NAS systems can work with lower-speed connections.

The disadvantages of NAS include a lack of built-in security. Organizations must use firewalls to tackle this. NAS isn't well tuned for storing enormous amounts of data because of the load that it puts on the network. Disaster recovery or tolerance requires a customized solution.

It's a good idea to ask your peers to help you decide which products will best suit your needs. Chances are that someone will have tried and tested it before you. Remember the old maxim: don't re-invent the wheel…improve it!

P. P Subramanian is Country Manager (India), Hitachi Data Systems and can be reached at subbu.subramanian@hds.com