CAPTIVE, COLD, HIND, HOT, WARM
WHICH KIND OF SITE IS BEST?

By Bob Johnson

Determining the best form of alternative processing for your data center, should it become incapacitated, is a difficult task and, it is never completed. The conditions that influenced yesterday’s alternative processing decision have changed! Is that change sufficient to cause the alternative processing question to be reevaluated? YES!
The first and most important aspect of making the alternative processing decision is that it must be fully documented. Ideally, in some form of a matrix that provides values to each of the factors considered. Then, as new factors are added or the value of original factors require adjustment, it is a smoother (and swifter) process to measure the impact of those changes upon the current strategies.
What ever method you currently utilize or choose to employ, you must be prepared to reevaluate. Many feel that either Hot or Captive sites (see Table 1, Definitions) are the pinnacle. This simply is not the case. While they are the most responsive, several companies have switched from one to the other. Additional organizations have downsized their needs by shifting less critical and non-critical workloads to less responsive (and less costly) alternative processing methods.

The in mode today is multiple alternative processing methods. This results in an even more complex situation which screams for effective documentation of the decision and periodic review of all components. The costs associated with the various forms of alternative processing are great; especially those that are not passive. However, the passive solution is frequently more expensive during a disaster.
All too often, the decision regarding alternative processing methods is made through the all approach. In this situation, all processing to be accomplished in backup mode will be done using a single alternative processing method. Frequently the all approach results in a cost that is excessive when considering active solutions. Not infrequently, a company will then decide to use a passive solution or hind site.
The place to start is with the applications being supported by the current system. What do they consume? What hardware/software support do they require including communications? What is their importance to the users the application supports and, are there other means of getting the job done? What financial impact will there be if the application is not supported? What embarrassment or regulatory impact is there? Will shareholders of the public be concerned?
Take the time to develop a standard questionnaire. Failure to do this results in less than adequate documentation and the final analysis will be reviewing dissimilar data. Because questionnaires are subject to interpretation, they should be completed through an interview process. If more than one individual will be conducting the interviews, all should be trained and provided the same basic information relating to rationale, reasons, background, etc. Whenever there is more than one individual that has knowledge about an application, then all should be interviewed. The composite will be far more factual.
In many cases a serious analysis of the applications will result in only a small fraction of the total requiring an immediate or near immediate alternative processing capability. Even within the financial community there are many applications or segments of applications that are not required on a daily or online basis. The emotion that must be overcome by the users of those applications can be a problem. However, that is not a reason to avoid the issue. It is simply one of the items to be dealt with.
It must be understood, that disaster recovery is addressing corporate survival. While the likelihood of any single business realizing a crippling disaster is small, it none-the-less is possible. The decisions regarding applications and what is truly critical impacts the bottom line figure which determines whether the appropriate alternative processing method is cost effective.
Those that have made the hard decision using the all approach, and have chosen one of the costlier active solutions, have much to gain by reevaluating and adopting multiple alternative processing methods.
The analytical cost associated is higher, but the savings that can be achieved will be greater than that amount if either the captive or hot site method is being used.
For many companies, decisions regarding alternative processing methods are being made with insufficient data and often is based on emotion. Considering the risks and associated costs, this does not make good business sense.
Here is an opportunity for all individuals concerned with disaster recovery for data centers. Perform a detailed analysis of the company’s requirements for processing capacity, to include end user requirements and develop an alternative processing plan that employs several methods based on the true needs. You will discover that there is substantial money to be saved.
Those just approaching this aspect of their plan or that have previously decided not to incur the expense of the all approach, can be confident that the final result will be a plan that will meet the company’s needs, is cost effective, and stand a much greater chance of obtaining the needed funds.

This article adapted from Vol. 2 No. 1, p. 22.

Disaster Recovery Worldİ 1999, and Disaster Recovery Journalİ 1999, are copyrighted by Systems Support, Inc. All rights reserved. Reproduction in whole or part is prohibited without the express written permission form Systems Support, Inc.