Disaster Recovery Experts Speak Out (continued) 7

<b>THE DISCUSSION (continued)</b><P><b>Hot Site Criteria</b>

InformationWeek Staff, Contributor

December 1, 2001

6 Min Read

THE DISCUSSION (continued)

Hot Site Criteria
thread by Bob Gorski (08-Nov-01 4:28 PM GMT)

Our organization has had numerous discussions lately regarding systems/applications that should be considered for a "Hot Site." Everyone seems to have a different set of criteria and consensus does not appear to be imminent. I would like to hear from the experts as well as anyone else who may have an opinion on what determines if a system/application requires a disaster hot site or gets put on the either the rebuild ASAP or the whenever we get to it list.

Re: Hot Site Criteria
by Smise (08-Nov-01 8:12 PM GMT)

Our organization is in the process of interviewing all of our departmental managers to understand completely what functions they perform, their importance level to the organization. What other internal groups they work with as well. We are then going to number them 1 to 4 as a Vital down to support-non vital role. This will assist us in implementing who gets the "hot" site and who gets left behind so to speak. Working in Manhattan we had the impetus placed on us to come up with this rather quickly. We are all waiting for the other shoe to drop. One thing to consider as well if your organization is building these sites in a relatively close geographical vicinity is whether they are on the same utility grid as your other sites. If your normal site is down will your hot site be as well. This applies to data, power, transportation lines as well. Something we had not really thought of until we realized that one of our "hot" sites in Greenwich, CT data lines were routed back down through downtown NYC right under 7WTC. We would have never known until it was too late. Now we are getting utility maps for all of our locations.

Re: Hot Site Criteria
by Martin Garvey (09-Nov-01 0:55 AM GMT)

I talk with every sort of customer with all sorts of apps. Brokerages, hospitals, 911 centers, utilities, and some others don't have much choice. Most companies should protect at all cost any customer facing apps. Tick off a customer and they might never come back. Next I'd tie each app to revenue and how much per hour is lost if the app is down. I'd base my threshold on that. Some apps get live replication, others can come back in an hour or two. Finally, some apps are in the archived stage, and hours to days is tolerated, as long as the information isn't lost.

Re: Hot Site Criteria
by Martin Garvey (09-Nov-01 0:58 AM GMT)

Thanks for the feedback. It's excellent advice. One of the great things about the IT community is the open exchange of information. You provide a shining example.

Re: Hot Site Criteria
by Bob Gorski (09-Nov-01 7:24 PM GMT)

Thanks for the response. Two questions: 1. Are the interviews being conducted using a standard set of questions? If so could you share them with me? 2. What will be the factors that determine if a function gets a (1) Vital, (2), (3) or (4) Non Vital ranking.

Re: Hot Site Criteria
By Ray Dmochowski (12-Nov-01 9:02 PM GMT)

I think most "experts" would agree that that the systems/applications that support the cash-flow and revenue streams are probably most critical to business continuity.


Re: Hot Site Criteria
by Bob Williamson (09-Nov-01 2:10 PM GMT)

Congratulations on taking a well thought out approach to building your DR plan. Too many companies jump into implementation without first doing the Business Impact Analysis that is needed to put the correct plan in place.

A question - have you taken the step to document all of your job functions and processes as you are doing the interviews? Something that we don't like to think about but should is loss of key people in a disaster. If the job functions and processes by which the jibs are done are documented, it can help to mitigate some of the confusion that occurs with loss of key personnel.

Real time recovery
thread by Mark (08-Nov-01 7:21 PM GMT)

I would like to know if products like EMC SRDF and Veritas Volume Replicator are comparable and if either could be used for real time recovery?

Are there any pitfalls with either? Do both work with Informix databases?

Re: Real time recovery
by Martin Garvey (09-Nov-01 0:48 AM GMT)

I'll answer based on what I know, but I'll double check with both vendors. I believe they can both be used in mirroring or real time replication.

A pitfall could be high cost. Any real time mirroring, though, will be expensive. You have to measure the cost of the product and architecture vs. the cost of business downtime. For many companies, the product cost is well worth it. Since Informix has been a stellar RDBMS for a long time, and he products have been available for a long time, I thing they both support Informix. But I'll check on the real time Q and the Informix Q with both vendors. Shoot me an e-mail when you have a minute: [email protected]

Re: Real time recovery
by Bob Williamson (09-Nov-01 1:55 PM GMT)

Hey Mark,

Yes, these products, along with others, can be used to ensure that you have a copy of your data a remote site to support real-time recovery.

These two products operate in very different ways to accomplish the same tasks. The EMC product runs only on EMC's Symmetrix storage line (thus it's name, Symmetrix Remote Data Facility) and uses code in the storage unit's firmware to handle the data replication. The Veritas product is purely software-based so it is storage independent. I am certain that the Veritas product does work with Informix, and I don't think that the SRDF product cares at all what type of data it is replicating, so it should work as well.

We have similar products at my company, SteelEye. Take a look at Extended Mirroring for the Windows environment and LifeKeeper Data Replication on Linux. You can see them at www.steeleye.com

Data mirroring is a critical component to any remote-site real-time recovery plan.

Re: Real time recovery
by Rick Weaver (09-Nov-01 4:06 PM GMT)

Not only does SRDF not care about the type of data you mirror, but it can also support any platform you attach it too. Theoretically you could attach Distributed Systems applications via SCSI adaptors and Mainframe applications via ESCON adaptors to the same Symmetrix box, and mirror the whole lot at once. Personally I've never seen this done, but the EMC hardware supports it.


<< Previous Page | Next Page >>

Read more about:

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like

More Insights