even Secrets to 24x7 Exchan  e

Availability

Exchange during a disaster, some data is permanently lost, the IT staff has to

intervene to bring Exchange back online, and more IT resources are needed to

recover or back up data. In contrast, disaster-tolerant solutions enable Exchange

(and other Windows applications) to operate through disasters, providing continuous

service to end users and complete data protection.

To choose the right technology, evaluate Exchange in business continuity terms of

recovery time objective (RTO) and recovery point objective (RPO). In this context, the

RTO is the amount of time your business e-mail system is down. The RPO is a

measure of how out-of-date your data is once e-mail is back up and running.

Technologies vary widely in their ability to reduce RTO and RPO. They also vary in

their cost to implement and manage.

Companies that keep backup servers on-site, pre-staged with Exchange installed a

so-called hot standby methodology can reduce RTO to hours while RPO may be

several days. Companies can also configure redundant servers into a cluster, with

automated failover from an active server to a standby server in case of a fault or

disaster. Under ideal circumstances, clusters can reduce RTO and RPO to tens of

minutes. But during those minutes, application state is lost employees are idled,

customers are frustrated, and money is lost.

Some technologies, such as traditional data backup, are focused on reducing RPO.

Although backup has a low acquisition cost, if offers no protection for the application.

RPO can be as low as 24 hours if your IT staff can access the backup media, set up a

new system, and reload the data in that time. More costly streaming data-replication

technologies can replicate to an off-site vaulting service, reducing RPO to several

hours. Synchronous data replication can reduce it further. However, these

technologies interrupt service to end users, lose transactions, and don t protect the

application. As a result, customers orders or other vital information may be lost and

revenue may be jeopardized.

 

TWO: CLUSTERING CANNOT PROVIDE DISASTER

TOLERANCE

Clustering for Exchange cannot provide disaster tolerance because it requires a

failover process. During the minutes of the failover, Exchange is not available to

users, application state is lost, and data that was in transit during the disaster may be

lost. If system administrators have not been meticulous about keeping the redundant

systems identical in all aspects, the failover will not work at all, and the RPO and RTO

could be hours to days. Furthermore, clusters require a significant investment of IT

Exchange clustering occurs at the

Windows level. Exchange s

sensitivity to failures results in a

significant number of false failovers

to the point that enterprises have

chosen to remove the clustering

after implementing it.

- Microsoft Exchange Server 2003

Gartner Report