IT Employment

General discussion


Disaster Recovery Standby Server classifications

By cmapp ·
Putting together Disaster Recovery Policy and would like to include definitions for the DR Standby Servers classifications as depending on the business application criticality, we will be directing the level of DR capability required.

Here is what I have put together but would like feedback on whether they are appropriate.

Cold-Standby Server ? A server that is pre-assigned for provision of DR capability for the associated application, although it is utilized for other purposes during normal operations (e.g. development/testing environment for associated environment). This server will need to be re-configured for use if required for disaster recovery purposes.

Warm-Standby Server ? A server that is dedicated to providing DR capability for the associated application. This server may need the latest version of the application software loaded and the application data restored before it can be used for disaster recovery purposes.

Hot-Standby Server ? A server that is dedicated to providing DR capability for the associated application. This server is fully configured; the application software and data is kept in synchronized with production and server is able to take over production processing load with minimal manual intervention in the event of a disaster.

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Comments

Collapse -

Technical vs Business

by CfK In reply to Disaster Recovery Standby ...

Your classification of servers is accurate and concise - from a purely technical standpoint. I would not fault this approach in describing them since it is farly standard stuff.

Are these offerings to the business to get their buy in for DR? If they are then you will not attract much positive feedback.

The approach to DR needs to be driven out of the business needs - going to each business unit, understanding the applications they use and the criticality of those applications to that business unit. Use a ranking system for each application to indicate the criticality of the application to the business need.
Create from these interviews a composite map showing how many Business unit use an application and the criticality it has to them. Rank the applications based on the results of this analysis so you can see clearly which are absolutely required and therefore rate putting on the Hot-Standby servers, which fall below this level but are still quite critical, and which are obviously not required on a day-to-day basis.

Using this approach you can then present to the business a (graphical) presentation of their requirements withing the organisation as a whole, and the requirements for the servers can be demonstrated out of this. Developing the costs for the servers and DR solutions as a whole becomes a fairly easy exercise following on from this.

Collapse -

Technical vs Business

by cmapp In reply to Technical vs Business

Thanks for your response. The policy I am building is purely a technical one that will support our business continuity management policy (which covers off the process for classification of applications from a criticality perspective). What I am trying to do is state that if your application has a certain classification, then you must have a corresponding level of DR capability, which is where the server definitions come in.

Collapse -

Technical Classification otions

by CfK In reply to Technical vs Business

There are a couple of extensions you may want to make to the list you have.
On-site vs off-site. Include this as a sub-classification of each category of server to make it clear where the server is located, and define why.

A Cold/Cold Standby Server. This is a fully provisioned server that has the application loaded, but will require restore of a backup copy of the data itself. This server would not be turned on at all except for a disaster. This can happen in an application where it is not possible to have two servers running it (some older databases and applications are machine-specific with hard-coded addresses etc) but some sort of alternative is required.

Cold/Warm is your Cold Standby option of using a 'warm' machine that is allocated to a sacrificial function. You need to be careful that what one department regards as sacrificial may be essential to another, so this is not a favoured option.

Warm/Cold - similar to Cold/Cold except that it is up and running and only needs data restored to it. For practical purposes the issue is more about time required to restore the data and be up and running than availability. This is generally a slower restore option (ie tape).

Warm/Warm - Where the backup option is from disk instead of tape, hence reduced time taken, but there is still a significant lag.

One measure to include is the estimate on how long it will take to get the application up and going on that particular scenario, and without going into the applications themselves that may be hard to do. However you can make the distinction between the Warm and Cold scenarios to getting the machine available for DR duty.

Collapse -

Your right

by dafe2 In reply to Disaster Recovery Standby ...

As you say (& another post here),just attach the business drivers toeach class.

Of course, you already know that every business unit and worgroup application is absolutely critical right?:-)

Related Discussions

Related Forums