Thursday, July 17, 2008

ERM-BC-COOP: Financial loss

Interesting situation before me.

I have a data base that is a repository for data from not one or two organizations but many organizations.

There is both intra-group and inter-group sharing.

The 800-pound gorilla group guesstimates that if its personnel can't access the shared data base, it will "cost" the company about 14k units of currency-per-day.


Let's "assume" that the smaller fish in the pond figure their losses to be 10k units of currency-per-day (for easy computing later).

Let's also assume that, in addition to the gorilla wading in the pond, there are 8 small frys (total 14k + (10k*8) = 94k units of currency-per-day.

The gorilla insists that the data base can be down for a maximum of two days.

Back to the basic math: 94k*2=188k units of currency.

Now we have a value for the data base.

Because (a) the data are critical and (b) we know media fails (that's why the maximum hard drive guarantee or warranty is 3 years), we have found space on a compatible machine.

If the machine on which the data base resides fails (or the facility goes away or . . . pick a risk), and since we backup-to-tape every day, and since we can retrieve and deliver yesterday's tape to the alternate site within "about" 36 hours, and since we know we can (assuming there are no Write errors) load the tape's data onto the alternate machine within 12 hours, we'll be up-and-running before the agreed-to 48 hour maximum downtime expires.

Question: How much value do I place on the data base for all days beyond Day 2?

Or, put another way, how PROBABLE is it that both the primary and alternate boxes will be lost at the same time? (They are housed at different regions.)

Realistically, the probability is minute. Small, even.

Bottom line time.

I have a box that, fully configured, is worth - say - 10k units of currency. RAID 5, all the Bells-N-Whistles, wireless mouse and keyboard; the server equivalent to a fully dressed out Harley.

I have a data base "worth" 94k units of currency-per-day to the user community that I know can be recovered - must be recovered - within 48 hours. I'll even concede that some work-aholics will access the data base from a Burger King WiFi access point on the Fourth of July.

What, given all that, is my daily financial loss for a month?

My take - and I'm willing to entertain other opinions - is

(94k * 2) + 10k + (1k * 30) = 228k units of currency.

What is the (1k * 30)? Manpower - possible, operative word is "possible," overtime for the alternate site personnel to nursemaid the temporary resident data base. Worst case expenditure.

Granted, 228k units of currency can make a major dent in some organizations' budgets, but I think it can be proportionally shared and absorbed by the 9 user groups depending on the shared data base.

One parting thought. While the techies are scrambling to get the alternate site up and running, the users can share their thoughts via email, telephone, fax, and courier so they are not necessarily out of business for the duration.

What bothers me is the once-a-day data backup.

Is it enough?

Do users keep a copy of their data on local machines? On external hard drives?

If I was writing the rules, I think I'd create a policy that states: Critical data will be stored on local media for (at least) 24 hours after being copied to the shared data base.

But I'm not "writing the rules." (Too bad; enterprise risk management very much includes development of policies and procedures.)

John Glenn, MBCI, SRP
Enterprise Risk Management/Business Continuity
Planner @

No comments: