There are eight important steps you should take to create a disaster recovery plan.

1. Secure support from senior management. A successful disaster recovery plan requires funding, coordination across departments, and the authority to carry out the plan. Depending on the size of your organization, you may also want to establish an interdepartmental planning committee.

2. Conduct a risk assessment. Prepare a business impact analysis that covers a range of potential disasters. Analyze major business units and attempt to determine the impact of different disaster scenarios. Evaluate the safety of critical documents as they are stored, in either print or electronic format. Make sure to include a worst-case scenario, such as complete destruction of your company's facility. Assign cost estimates to each disaster scenario; this will be helpful in justifying your budget.

3. Assign priorities to operations. Determine how users utilize network resources and the importance of those resources. This will allow you to make an informed decision regarding an acceptable amount of downtime. Understanding the business value of every network component will lead you toward more intelligent decisions. Begin planning by assigning priorities to network and data components.

4. Explore all options for disaster recovery strategies. Research and evaluate the most practical solutions. Budget for all equipment and services required to carry out each strategy. Potential options will be dictated by the priorities assigned in step 3. For example, you'll need a much different solution for an e-commerce site—where you can't afford to miss a single transaction—than on your print server. Typically, complexity and cost increase as the acceptable amount of downtime decreases. Don't fall into the money pit of trying to guarantee 100 percent uptime for every network resource.

5. Collect data. Conduct a detailed inventory of all network resources, from user workstations to routers. Collect as much information as is feasible. Physical details, such as make, model, serial number, and purchase date, will be important for filing insurance claims. Configuration information is important to design redundancy into your network and to prepare for any potential replacements. Make a list of information such as memory, disk, OS, and applications. It's important to keep the inventory current, so specify the documents' creation and revision dates. Pay attention to the human aspect of your business, and be sure to include contact information for all users, vendors, and service providers.

6. Create a comprehensive written plan. Construct a full outline and use it to gain initial approval. Make a road map of the final plan and divide the work among multiple departments (if you are part of a large organization). The plan must contain detailed procedures for before, during, and after a disaster. One procedure should address maintaining and updating the plan itself as systems and workflows change.

7. Develop test procedures. What good is a detailed plan if you don't know whether it will work? It is imperative to test and evaluate the plan thoroughly on a regular basis (at least annually). The more thorough the test plan, the more certain you'll be that you've addressed every contingency.