Profiling your Disaster Recovery Plan

WHITE PAPER
Pro? ling your
Disaster
Recovery
Plan
Table of Contents:
AbstractIntroductionSeven key metricsUsing the InMage Pro? ler tool1 - Measuring total storage capacity requirements2 - Determining bandwidth requirements for near-zero RPOHeadquarters: 3 - Evaluating Whether current bandwidth is suf? cient for a InMage continuous DR implementation 3255-1 Scott Blvd, #102 4 - Estimating the storage required for a desired Retention WindowSanta Clara , CA 95054 5 - Understanding how data changes are distributed over time Phone: 1.800.646.3617 6 - Calculations bandwidth requirements for a given RPO Local Phone: 408.200.3840 7 - Evaluating the provisioning effects of protected server selectionFax: 408.588.1590 ConclusionEmail: info@inmage.net The InMage differenceWeb: www.inmage.net
© 2007, InMage Systems, Inc. All rights reserved.Abstract
It's a fact: nearly 48 percent of ? rst-time Disaster Recovery (DR) plan implementations fail due to inaccurate estimates p2 of necessary resources. Due in large part to faulty methods of estimating resource requirements, such as analyzing logs generated by tape backup software, organizations come to conclusions that are drastically wrong. Since a mistake in this area can cost months of time and countless dollars to recover from, it is worthwhile doing accurate resource estimates early in the process.
InMage now offers a ? eld-tested method for determining the real answers to the top seven metrics that will ensure a successful implementation. This white paper covers each of the seven key metrics. It shows how the InMage DRScout Pro? ler can be used to quickly and accurately determine real-time resource requirements, and give an in-depth under-standing of the true Recovery Point Objectives (RPOs) that can be achieved at different points in the day.
Introduction
One of the challenges encountered in implementing a successful continuous Disaster Recovery (DR) plan is to ac-curately provision WAN bandwidth to meet the desired Recovery Point Objectives (RPO). IT organizations tasked with DR implementations lack comprehensive and accurate help in making these decisions. This is re? ected in the fact that nearly 48 percent of ? rst-time DR plan implementations fail due to inaccurate estimates of necessary resources. Recov-ering from these mistakes is a costly and time-consuming process that sometimes lasts several months.
A commonly used method to estimate the WAN bandwidth requirements is to analyze logs generated by tape backup software, where logs record the changes between differential or incremental jobs that usually run on a daily basis. There are three major problems with this approach:. These logs do not capture the data changes that happen throughout the day on a real-time basis. They do not record changes that occur within ? les at the byte level. . They give no understanding of the true RPO possible at different points in the day given a certain amount of WAN bandwidth.
Bandwidth estimates based on this approach can be off by 600 to 800 percent. Because WAN bandwidth is the single largest recurring cost in a DR plan, such a large margin of error could ground the entire implementation.
This white paper presents a method, and an easily-available tool, that will let you avoid these pitfalls and quickly con-struct a viable, true disaster recovery provisioning plan. You will be able to gather solid data on each of the eight key metrics, and build a plan that enables you to meet your Recovery Time and Recovery Point Objectives.
Seven Key Metrics
A good DR provisioning plan results from solid data that can be used to construct a usable plan: one that results from an ongoing process that can predict and preemptively provision additional requirements due to broader application cov-erage and data growth. In order to construct such a plan, the following key metrics must be calculated:
1. What is the total storage capacity required for recovery of select applications?
2. What is the bandwidth required for a near zero RPO? What is the bandwidth reduction due to compression?
3. Does the currently provisioned bandwidth suf? ce for a continuous DR implementation?
4. What is the storage required for desired retention window?
WHITE© 2007, InMage Systems, Inc. All rights reserved.PAPER5. How are the data changes... [download for more]