9 replies
by shanghai sam
I have to agree with Gartner! Whenever working with external vendors (and even with internal departments) it is critical that you look out for your own interests. If disaster recovery is important to you, then always specifically spell it out in the Service Level Agreement (SLA). The SLA is the legal yardstick by which your vender can be measured in court. If you don't have the details spelled out, you may find yourself up the proverbial creek without a paddle! Good disaster recovery options usually come with a noticable price tag.
As for personal experience, I did have a severe server crash! The redundant fail-over server kicked in beautifully using the replicated database, and was relatively seamless to the end-user community. Disaster planning is essential! When you have carefully planned contingencies, you can truly avoid a real disaster.
by itdebate
by the chad
Having planned and executed business continuity plans (BCP) over the past 3 years, I can say that unless those plans are actually tested on a regular basis, they are next-to-worthless.
The two largest points of failure I've seen have been when the plans fail to take a critical process into account (either because the plans are out of date, or the plan author did not interview the business process owners properly), or recovery personnel fail to follow the recovery procedures.
Only by performing mock tests can one shake out the "bugs" -- waiting until a tornado takes out your datacenter is too late.
I would insist that the outsource company plan at least one major BCP test per year, with at least someone from my company present to participate (perhaps testing connectivity at our end). If the company was unwilling to do that, then I would not do business with them.
See, it doesn't matter what they put in their contract vis-a-vis uptime and guaranteed recovery. If they fail then YOU are stuck with the consequences of having no data center and scrambling to find another vendor. Maybe you'll get your data recovered in time, but what if there is a regional failure? How many other companies will be looking for a new data center at that time?
Some BCP shops have more clients than room. For that reason, we contracted with a multi-state vendor and reserved two major locations in different areas. Should one go down due to earthquake, it is unlikely that the other will also be down.
Would the data center outsourcer be able to do this? Would they be willing?
by itdebate
by plantogo2000
We do have a formal recovery plan for outsourced vendor services for our small operation. There has not been an actual disaster experience save for one virus attack that did not result in disaster. We do know exactly what would happen in a disasterattack and the reasons for such awareness are noted below. My specific role is to coordinate the contact network (names and phone numbers at home and at work including backup contacts) of those persons that would be involved in the disaster recoveryeffort.
Agreement with Gartner?s position is not absolute. Having a contract that explicitly identifies all of the requirements for recovery is 80% of the solution and the remaining 20% of the solution is going to take 80% of the actual physical effort required. By order of magnitude, if the disaster recovery plan took ten hours to write the test will take 40 hours. The effort required is a twofer. That is, the resulting action is effective in either a physical disaster or a virus attack. N.B. Those companies that developed effective Y2K plans should dust them off and get ready to do the next step if they have not already.
The next step is to conduct an actual recovery exercise. A real life simulation of the events and procedures thatwould be required for every type of conceivable (within risk limits) recovery effort is absolutely required. A practice run, more than once depending on the results of the debriefing effort, is mandatory. This training exercise will put you in position to evaluate not only your effectiveness to respond to a disaster but evaluate the ability of the outsource service to meet their contract disaster recovery requirements.
by itdebate
by plantogo2000
We do have a formal recovery plan for outsourced vendor services for our small operation. There has not been an actual disaster experience save for one virus attack that did not result in disaster. We do know exactly what would happen in a disasterattack and the reasons for such awareness are noted below. My specific role is to coordinate the contact network (names and phone numbers at home and at work including backup contacts) of those persons that would be involved in the disaster recoveryeffort.
Agreement with Gartner?s position is not absolute. Having a contract that explicitly identifies all of the requirements for recovery is 80% of the solution and the remaining 20% of the solution is going to take 80% of the actual physical effort required. By order of magnitude, if the disaster recovery plan took ten hours to write the test will take 40 hours. The effort required is a twofer. That is, the resulting action is effective in either a physical disaster or a virus attack. N.B. Those companies that developed effective Y2K plans should dust them off and get ready to do the next step if they have not already.
The next step is to conduct an actual recovery exercise. A real life simulation of the events and procedures thatwould be required for every type of conceivable (within risk limits) recovery effort is absolutely required. A practice run, more than once depending on the results of the debriefing effort, is mandatory. This training exercise will put you in position to evaluate not only your effectiveness to respond to a disaster but evaluate the ability of the outsource service to meet their contract disaster recovery requirements.
by itdebate
by itdebate
This discussion is closed to new replies.
Trending IT Employment posts
Sort by Trending