Extending storage over WANs

Disaster recovery and business continuity often demand long-distance storage transfers, but users have to weigh the cost-performance tradeoffs between the various methods of data protection.

By Gary Johnson

With the heightened attention on business continuity and disaster recovery, many organizations are looking to extend their storage over long distances using a wide area network (WAN). However, extending storage over a distance of 100 miles or more brings a unique set of considerations, including latency, data integrity, and bandwidth utilization.

This series of articles will describe how these issues affect the performance of a storage network, review the technologies available to address these issues, and provide guidelines on how to choose the best technology for your specific requirements.

This first article in the series provides an overview as well as a discussion of disk versus tape considerations when dealing with distances of 100 miles or more. Subsequent articles will examine the issues of latency, data integrity, and bandwidth utilization, and how to deal with these issues as you consider extending your organization's storage network over long distances.

The business continuity mandate

Companies today increasingly rely on computer operations and information access as the lifeblood of their organizations. Disruptions to these operations can have an immediate negative effect on the business. It is therefore extremely important for many organizations to design redundant systems for non-stop operation of critical functions and to have a reliable business continuity strategy.

Much has been written about the need for disaster recovery and business continuity plans that are justified by the potential cost of downtime. Such plans are now under revision by many organizations to improve recovery period objectives from, say, 72 hours to less than one hour, and recovery point objectives from 24-hour loss of data to almost no loss of data. In fact, "continuous-availability" designs are becoming mandatory for the core operations of enterprises in many industries.

These business continuity plans cannot be accomplished without the ability to effectively extend storage over long distances.

Another factor in the need to extend storage is that, as organizations expand, their various IT sites may be hundreds or even thousands of miles apart. Consequently, remote data replication solutions must be able to span long distances; integrate multiple network systems, protocols, and interfaces; and provide robust network management capabilities.

The combination of storage and networking is now an integral part of making information available across the enterprise. By replicating data across LAN, MAN, and WAN networks, companies can better protect, copy, migrate, distribute, and manage their information assets, wherever they're located.

Click here to enlarge image


The Gartner Group predicts that the growth rate in data replication investments will go from 25% this year to 75% in 2005. And some storage vendors are already reporting that up to 30% of the storage they sell is being used for remote purposes.

At the same time, the exponential growth of data and the need for non-stop operations have driven the rapid adoption of Fibre Channel storage area network (SAN) "islands" and, more recently, isolated iSCSI departmental-level SANs based on IP networks.

While the challenge for IT managers used to be managing corporate information locally within the data center, now it is how to manage information between data centers, which might be hundreds or thousands of miles apart.

A key question is which data replication to use. The distributed nature of IT that has resulted from business growth, regulatory compliance, and events such as September 11, means that connecting storage over distance is now a common requirement.

When deciding which technologies to use for replicating data, it is helpful to think in terms of a "continuum" of business continuity solutions (see figure).

Moving data around today's extended enterprise requires a sound storage networking strategy and a safe, cost-effective network infrastructure. Any bottleneck or disruption in the flow of corporate assets can have serious implications, which may ultimately affect the bottom line.

In this pyramid, different types of data require different levels of protection. For example, the most important mission-critical data—is at the top. This data must be protected at the highest possible level, and some companies use clustered, remote hot-site facilities for this purpose. Other options, depending upon the criticality and volume of the data, include remote or local disk mirroring, electronic tape vaulting, and off-site tape backup storage. Note that cost increases significantly as you go up the pyramid.

Many vendors offer data replication solutions to match the different priorities for business continuity. In a subsequent article we will highlight enabling technologies that address the issues associated with deploying storage over WANs—a key requirement for any business continuity plan.

Virtually all large and mid-sized enterprises recognize the need to have at least one copy of their data in a location outside of their main data center to protect against an event that could put that data center out of business.

Many firms have implemented MANs, using point-to-point fiber optics as the data transport medium. While this method of data transport works well, in practical terms it is limited to distances of less than 100km, or 62 miles. Many other companies are reviewing contingencies that include moving data to storage facilities that are beyond the reach of MAN technologies. These contingencies are intended to protect against natural disasters, such as earthquakes, floods, and hurricanes, which have the potential of affecting an entire metro area—or beyond.

Other companies want to leverage their multiple data centers, which can be hundreds of miles apart. In the case of backup or archive tapes, these companies could use planes, trains, or trucks for transport, but this manual method has a negative impact on recovery time objective (RTO), data availability, and security.

However, when IT managers try to discuss extending storage devices (tape and disk) over a WAN with contingency planners and storage specialists, many of them remain skeptical about the reliability of electronically moving business-critical information over long distances.

These concerns can be allayed by technologies that provide safe and reliable data transport over a variety of WAN signaling and protocol interfaces.

The next article in this series will discuss the key issues surrounding storage over WANs: latency, data integrity, and bandwidth utilization.

Gary Johnson is vice president of solutions at CNT (www.cnt.com) in Minneapolis, MN.

This article was originally published on July 01, 2003