I believe that a centralized model would be ideal for recovery of a high impact organizational "disaster". Reasons for choosing the centralized model are:

1. Allows for streamlined command and control of personnel coordination. Technical resources are always limited and must be utilized optimally to expedite the recovery process.

2. Faster recovery of core systems. Resources can be concentrated on recovery of key systems which must be brought online within the designated RTO.

3. Reduced possibility of redundant effort or conflicting recovery measures. These types of incidents are common due to poor communication and may lead to wasted resources, system misconfiguration & extended recovery time.

4. The effort required to manage one large recovery effort is less than what is required to manage several smaller ones.

Warrick makes some valid points. If you use a decentralized model it will be less cost effective for your company. It will take more time to try and figure out whatbsystems and aplications need to be brought back online after a disaster. Even if you have documentation of all the upgrades, patches, applications, etc.; it is more difficult to take care of a decentralized system. Due to the fact that many companies are allowing employees to work from home thru the use of a VPN or some type of Internet connection they may want a decentralized model. But when you do an assessment and look at the cost to benefit ratio I think the centralized model will be the model of choice, for now.

Disaster recovery can be performed as a result of so many unforeseen circumstances, some of which can range from just data corruption, lost of critical servers or even an entire building or business complex. When implementing a Business Continuity Plan, you have to plan for the worst and unexpected events. Finding a recovery plan for some of the disasters that may occur can sometime be unthinkable. When cost is factored, it is even worse to think of how a business will recover from such a disaster.

Today, businesses rely more on data generated and stored using varieties of technology infrastructures than years back. Considering how critical such data have become and the threats and disasters that loom around the technology industry, it cannot be overstated that data recovery is a key feature of technology deployment.

With the topic under discussion, I would prefer to use a decentralized model as my recovery method. Yes, critics might argue that it will cost more, but the mere fact is no amount of dollar can pay for data. A decentralized model would be one of those methods that will have your organization recover and back in business within the shortest possible time frame. When you get to think about the worst disasters, a decentralized model would be one of those things that will save an organization time and money when recovery is needed. Hot, warm and cold sites have all become common recovery methods used today and these are all decentralized models of recovery. When you can’t find a solution to recovering from the worst unforeseeable disaster, a decentralized recovery model is the way to go.

For the business survivability, it is best to use decentralized approach. Like they say,"don't put all your eggs in one basket." Let's take the situation of 9/11's tragedy where some firms went bankrupt. They were not able to recover after the 9/11's disaster. Some of the firms had their backup in different floors of the tower or in different buildings within a block or two away that were destroyed. This is where the decentralized model comes in play. Dispersal of your assets or backup data is essenteal for organization survivability. If your secondary site is destroyed how are you to recover? Decentralized is expensive compare to centralized, but this is a matter of business survival. There are also other benefits of decentralized model such as different IT departments. IT department can update patches and fixes much faster because they don't have to go through organization's bureau.

Nate,

I agree with you in certain situations.  But if you have multiple facilities not all the systems can work independently of a server hosting it.  For example, if the database server goes down, having the other facilities on fat client may not do any good because 80% of your business may happen on that database server.  Of course it is all about the specific scenarios that you encounter.  So what you are saying is completely valid :-).

-Dave

I agree Tim, However, in choosing a centralized or decentralized system , we need to take into account the type and size of the business in order to determine the system that makes the most sense. Data recovery may be faster and more cost effective on a centralized system but what if your backup is corrupted?