iSCSI: When It's the Right Fit for a SAN

Over the last decade organizations large and small have seen an explosion in the volume of business-critical data that is being captured, processed, stored, and manipulated. Efficient data management is becoming one of the most important aspects of business-critical decision making. Ever-increasing reliance on applications, ranging from business intelligence and decision support to electronic messaging, and increasing use of graphics (as opposed to text) contribute to this trend.

In response to these demands for efficient data storage and management, shared storage solutions are being deployed at all levels to take advantage of the inherent benefits of a networked approach. These benefits include increased availability, scalability, and potentially performance. By bringing networking technology to data storage, a shared storage solution can also provide increased flexibility and centralized management with a lower total cost of ownership.

Today, with the mainstreaming of the storage protocol known as Internet SCSI (iSCSI), there is now a viable alternative for small and midsized organizations that allows them to reap the proven benefits of storage area networks (SANs).

In addition, large organizations may choose to use iSCSI instead of, or in addition to, Fibre Channel in workgroup and departmental environments. The relative simplicity of iSCSI uniquely positions it to meet the needs of those direct-attached storage (DAS) environments that, through business expansion, now require shared storage features at minimal cost.

But, beyond just storage, where does iSCSI fit in the wide-ranging space of business applications? What is the best match, architecturally speaking, for iSCSI in the typical mix of applications found in small and midsized organizations? Finally, does an organization have to decide between Fibre Channel and iSCSI, in a "Hobson's Choice" either-or sense, or can they use both, simultaneously, within a single infrastructure and a single storage array?

Where iSCSI Makes Sense

Specific characteristics that indicate a good fit for iSCSI are as follows. The more indications that are present, the better the fit:

• Microsoft® Windows®, Linux®, and/or VMware® operating systems, running on x86-based or x64-based servers
• All DAS for business-critical data and applications such as database, email, customer resource management (CRM), and manufacturing, as well as user-level file and print services
• Gigabit Ethernet for server network infrastructure
• Full-time general purpose IT staff (albeit small, typically shared with all other workgroups)
• Preference for a comprehensive solution (as opposed to construction by parts)
• IT budget (annual spend) between $140,000 and $500,000
• Purchases made primarily through a value-added reseller (VAR) or other indirect channel
• Interest in moving from DAS to higher functionality and scalability of SAN to enable business expansion or FTE efficiency
• Tendency to postpone purchases until a business-critical need occurs

Business Requirements

• Reduced capital investment (key differentiator) on servers and server-related hardware (e.g., I/O adapters) relative to Fibre Channel and reduced operating expenses relative to DAS
• Lower transaction response times and greater I/O per second (IOPS) performance relative to DAS
• Centralized management of data storage instead of the decentralized, server-by-server management that characterizes DAS
• Scalability over time without having to couple the purchase of servers to the purchase of storage capacity or performance

Summary

Read The Full Article