Michael McNamara (pictured) writes: The growth of business data continues to explode along with the need to store it. Workers generate more and more e-mail messages and file attachments, users demand instant access to data like never before, IT managers install more storage-hungry applications, and aging paper-based data continues to be converted into digital form. Information growth is so intense, in fact, that spending on data storage is expected to outstrip server spending.

However, with IT managers facing flat or shrinking budgets, the pressing challenge for them is to do more with less - to squeeze the most data storage out of every IT dollar. To achieve this objective, they must start by assessing all data storage costs - those tied to initial equipment acquisition, as well as those for resource management, capacity use, and most importantly, system downtime.

Three options exist today for managing data: Direct Attach Storage (DAS), Network Attach Storage (NAS), and Storage Area Networks (SAN).

Direct Attached Storage (DAS) represents the status quo in many organizations that aren't aware of the hidden costs or technology limitations related to this form of implementation:

• Difficult to Manage - Data is dispersed over many servers, which increases the personnel cost for supporting the organization with online configuration management and backup/restore capabilities.
• Limited Asset Utilization - Since each server owns the storage connected to it, DAS makes it almost impossible to share storage assets across multiple servers.
• Low Scalability - Server scalability is limited by the number of I/O buses supported and the SCSI bus maximum of 15 devices.
• Limited Distance - SCSI implementations typically have a 12-meter limit, which doesn't provide flexibility or let storage assets be located in secure locations in a facility or on a campus.

Network Attached Storage (NAS) is an attractive alternative to general-purpose computers, but has limitations that constrain customer configurations:

• Performance Constraints - Based on workload, the NAS box's performance can be constrained by CPU power, network throughput, and storage I/O bottlenecks.
• Bandwidth Requirements - Network bandwidth for the NAS server can compete with the other computer resources on the network.
• OTLP/Database Bottlenecks - NAS excels at file-based access but can be bottlenecked on OLTP applications and database block-level driven applications.

Storage Area Networks (SANs) represent a topology for connecting storage assets directly to the network and establishing a peer-to-peer server/storage implementation. SANs have historically been based on Fibre Channel and can now also incorporate iSCSI as a method of server/storage communication. SANs solve multiple issues for large enterprises with data centers and remote facilities and meet the IT requirements of SMB environments.

For years adding storage meant buying additional servers, tape libraries, and disk enclosures to attach to the server - a costly and inefficient tactic that left large amounts of storage capacity and computing power unused. Today, SANs - high-speed networks that connect multiple storage devices so they can be accessed on all servers in a local area network (LAN) or wide area network (WAN) - have proven to reduce management costs as a percentage of overall storage costs. Other benefits include:

• Increased disk utilization
• Reduced data center/rack floor space
• Improved data availability
• Improved LAN/WAN performance
• Reduced storage maintenance costs
• Improved protection of critical data
• Reduced CPU loads on servers freeing up computing power

See Figure 3

One way to combat the increasing cost and complexity of storage is to consolidate it in a single pool with fewer storage devices shared among multiple servers. By consolidating storage in a SAN you can:

• Reduce the number of physical devices to manage
• Reduce complexity
• Centralize storage management tasks
• Simplify growth and expansion
• Maximize storage utilization and return on investment