Storage virtualization and the full impact of storage disruptions: relief and ROI.

On the surface, most IT professionals would agree that the most worrisome causes of downtime are natural disasters and unexpected failures. But get a little more specific and mention "storage-related disruptions," and they'll agree that most of their downtime headaches come from the myriad of planned outages needed to manage storage resources.

Every change made in the storage environment has a direct impact on applications that need to be up and running constantly. Whether it's several minutes for rebooting, a few hours for backups, or several days for reconfiguration, these interruptions have a quantifiable cost to the enterprise in terms of lost opportunities to conduct business. The business cost of planned downtime needs to be regarded with the same seriousness that unplanned outages currently garner.

When IT administrators groan about the pains of storage management, it isn't just about complex cabling, tedious physical reconfiguring of hardware, long backups, and learning several different management interfaces. All these problems are cumbersome to cope with, but in the past, most administrators have managed to avoid unduly impacting business operations by sacrificing nights and weekends. Global 24x7 operations and rapid data growth have changed all that, and there is no longer any "safe" time to handle routine maintenance tasks.

In a study of e-commerce Web sites, Boston-based Aberdeen Group estimated that about 30 percent of all user-perceived Web server outages are caused by planned storage reconfiguration. The research also shows that frequently unavailable Web sites can expect to lose 7 percent of their customer base every year. Clearly, a case can be made that eliminating known storage-related disruptions has a real monetary value and can contribute significantly to a corporation's return on investment (ROI).

Storage virtualization may provide many companies relief from the headaches--and costs--of planned downtime. As a flexible, highly automated solution for addressing many of the most fundamental storage management activities, virtualization can significantly reduce the need for shutting servers or the network down.

The Sources of Storage Related Downtime

There are a number of common storage management activities that cause critical business applications to either be taken offline or run at less than full potential:

* Adding raw disk capacity

* Allocating storage to servers

* Removing storage from servers

* Backing up data

* Replicating volumes

* Migrating data

* Adding/maintaining servers

* Testing disaster recovery readiness

Some of these items are often overlooked when estimating anticipated server downtime because not all of these tasks completely disable servers. For example, replicating a volume might consume approximately 35 percent of a server's processing power--a seemingly small price to pay. But even partial system performance degradation jeopardizes the ability to operate the business at full potential. 35 percent less available power can translate into 30 percent fewer transactions that the server can handle--and 30 percent fewer customers.

Direct-Attached Storage Pays the Biggest Price

Naturally, enterprises that attach storage devices directly to servers are the most exposed to storage-related downtime and feel the most urgency to change. To add, move, and modify storage in a direct-attached environment means rebooting the server at best, and taking it out of commission for several hours or days at worst.

Strategic Research Corporation, based in Santa Barbara, Calif. estimates that the banking industry places a value of about $2.6 million on every hour of data center processing. Even spread across 100 servers, that's a hefty price to pay for disabling a server to add more capacity.

Stamford, Conn.-based Gartner Inc. studies reveal other industry averages for the value placed on an hour of processing (See Table).

The numbers in the table emphasize the intimate relationship between technology and business, and the importance of minimizing interruptions--regardless of whether they are planned or unplanned. Avoiding these opportunity costs can quickly provide more than adequate return to justify a storage virtualization investment.

A Common Approach: Bring in the Big Iron

Seeing the frightening potential for lost business, some companies with deep pockets have elected to alleviate the impact of disruptions by moving to large, consolidated shared disk arrays. These shared arrays clear the jungle of direct-attached devices, put all the storage resources in one big box and provide an advanced storage subsystem with features for on-the-fly volume allocation and point-in-time copies of production data. In this way, shared arrays offer effective relief from several types of disruptions. Companies that have made these investments find that these capabilities go a long way towards ensuring that business operations run smoothly and continuously.

However, the high-end features of consolidated arrays also come with high-end price tags, proprietary technology, and vendor dependence, which have their own ramifications on the bottom line. At one time, these premiums were tolerated--the economy was more upbeat, corporate growth required storage expansion and IT budgets were more accommodating.

Current economic conditions put pressure on IT departments to trim costs, making expensive consolidated arrays difficult to justify. In addition, the features provided by these storage subsystems are limited to individual hardware frames and each can only connect to a relatively small number of application servers. When storage demands exceed the capacity or connectivity of a single unit and more frames are required, some of the management difficulties of direct-attached storage re-emerge.

Virtualiation Can Help

Storage virtualization provides advanced capabilities to shield applications from the many changes that take place in the underlying storage environment. A virtual pool of storage behaves similarly to a consolidated array, and in many ways complements it. However, instead of confining these functions to a single array, the advanced storage management services span multiple arrays from potentially different suppliers.

Depending on the features and configuration, it's possible to either entirely eliminate disruptions or significantly reduce their impact on business. Solutions that include features like a robust high-availability architecture to protect against a single point of failure and disaster recovery options, like mirroring, maximize the storage investment.

The rollout of a virtualization strategy can all but eliminate storage-related interruptions, keeping applications constantly running at full potential to deliver a competitive edge.

Figure 1

Industry         Application            Avg. Cost/Hour of Downtime

Transportation   Airline reservations   $89,500
Retail           Catalog sales          $90,000
Retail           Home TV shopping       $113,000
Media            Pay-per-view           $1,150,000
Financial        Credit card sales      $2,600,000
Financial        Brokerage operations   $6,500,000

www.datacore.com

Calvin Hsu is a product marketing manager at DataCore Software Corp. (Fort Lauderdale, FL).