Read an Excerpt
Chapter 1: The Network Storage Landscape
Storig and sharing data over computer networks is hardly a new idea. Most readers of this book are familiar with basic client/server network computing built around the idea of a file server that provides storage services to a group of client workstations. As network computing has evolved over the years to include distributed processing technologies such as message-based middleware and clustering, the requirements for the storage devices and subsystems have also changed. This chapter analyzes some of the primary forces pushing host-oriented storage technologies in the direction of host-independent storage networking technologies.The Changing Role Of Data As A Business Asset
A great deal has changed in the computing world over the last 20 years. Inexpensive, powerful computers and networks of computers can do the work that required an expensive mainframe 20 years ago. One thing that has not changed, however, is the importance of the data that computers process and produce. If the data is lost, all the computing power at hand is virtually worthless. So, one of the challenges for the data storage industry has been to provide the type of reliability and protection that is required for 24x7 operations on networks of inexpensive systems.
This is easier said than done. Managing storage in a network environment has proven to be generally difficult. A single business location can have several different hardware/software platforms, each with their own systems and storage management utilities. This is an extremely difficult environment for system administrators who have to deal with all this diversity with little or no margin for error.
There are two approaches to managing storage in network environments. The first is to manage it through a server-provided interface and the second is to manage it through a storage-specific direct connection. The latter approach is the most common one taken by companies in the storage industry, and it leads to a very interesting conclusion: Data is an independent asset, separate from the computers that access it, and requiring a management system that is independent of host systems management.
As data is increasingly thought of as its own free-existing entity, not necessarily belonging to any particular system, it is also being viewed as a corporate asset, similar to capital or intellectual property that needs to be preserved and protected. Likewise, storage networking products and architectures, as platforms for data protection and storage management, are also being elevated as a strategic asset worthy of planning and budget meetings formerly reserved for systems and software.
Establishing the Relative Worth of Data
Storage networking products contain data that is a corporate asset. Therefore, one might want to know what that data is worth to the organization. Both qualitative and quantitative attributes should be weighed in making this determination. In general, qualitative attributes can't be measured and are rarely agreed upon but quantitative attributes can be measured in detail if one chooses to do so. We'll try to work through aspects of both to help the reader calculate their data's "net" worth.
Measuring the Qualitative Value of Data
Qualitative measurements are the most fun and the least reliable. Questions like, "So what is going to be the return on our $250,000 investment in this advertising campaign?" illustrate the nature of qualitative worth. Nobody can really say exactly what the qualitative worth is, but it's clearly a lot more than zero. At a minimum, one ought to be able to get into some inspired discussions about qualitative data.
The appeal of a company's web site is qualitative; so is the friendliness of its automated voice response system. E-mail systems have both qualitative and quantitative values. If a company loses all its e-mail, there is likely going to be a measurable quantified productivity loss, but there is also going to be a qualitative impact in the way customers, partners and employees perceive the organization.
When calculating the value of data, it's important to estimate the impact of parting ways with customers, partners, and employees who lose confidence in a company while its computers are out of commission. Frustrated customers don't necessarily notify their existing vendor to tell them they're done doing business with them and are buying instead from a competitor. The cost of retaining a customer is far less than the cost of winning them back.
Table 1-1 lists several qualitative data types along with some suggestions about how the reader might calculate their relative value. The percentages and calculations are based on personal experience with data recovery planning and on pure conjecture. The reader may find it useful as a planning guide.
Measuring and Prioritizing Data
Measuring the quantitative worth of data requires rigor and persistence. However it is done, at some point a detailed and comprehensive disaster recovery plan should be written. An important result of the disaster recovery planning process is a thorough prioritization of the recovery sequence for applications and systems. This involves meeting with business managers to determine a pecking order for computer services. (It can be somewhat entertaining and flattering for an IT worker to observe business managers who depend on the computer systems arguing over the relative importance of their systems and applications.)
From time to time different groups, including professional accounting organizations and the U.S. government, study the results of disasters and downtime on business and employment. The results are never good and are usually fairly chilling...