Disaster Recovery RTO

[David Teh]

Hi folks,

RTO (Recovery Time Objective):

From the time a disaster is declared to the time the systems are restored and handed over to End Users to verify and perform lost and orphan data entry.

The normal practice seems to be to establish an RTO for each individual system. This can range from 2 hours (for those on hot standby DR) to 48 hours (for those using tape recovery).

But when disaster strikes the entire data center, and operations fail over to the alternate site, is the individual systems’ RTO still valid?

I am thinking of the situation where the source system’s RTO is much shorter than the receiving systems. Data synching needs to take place, and that can’t start until receiving apps systems have been recovered. Should the users of the source system still expect the system to be returned to them by RTO? Or is that RTO valid only when there is failure of that one single system alone?

TIA.

Topic

[Robert Gordon]

CNN 09/05/07 Nuclear weapons flown over US.   Think again about getting a license key and not having phones, better yet the server’s nic card gets fried and the license key you got ahead of time no longer works.  Hope we don’t have to go to the extent of having a lead shield bunker for computing with fall out.

[David Teh]

Hi Robert,

Well….anything can happen. 911 certainly woke *some* (not all) up…the following is a ‘fun’ (if not sobering read):

http://www.operationaldynamics.com/reference/papers/Sept11/

Robert Gordon wrote:

CNN 09/05/07 Nuclear weapons flown over US.   Think again about getting a license key and not having phones, better yet the server’s nic card gets fried and the license key you got ahead of time no longer works.  Hope we don’t have to go to the extent of having a lead shield bunker for computing with fall out.

[Author]

Replies