« Risk assessment and the Nu M8 Child Tracker | Main | Don't lay all the blame for insecure systems on the developers »

In the cloud or up the spout

KingS There is an inevitable moment in married life where the wife will ask of her husband the question "what were the best moments of your life so far?" The expected and acceptable responses are well documented. They include "the day we met", "the birth of our first child", "our wedding day" and other such momentous events. It's all rubbish of course. While I might stick to the script for the sake of marital bliss, I'll actually be thinking "the day I learnt to play along on the guitar to Back in Black", "driving at 130mph on the autobahn in my old Lotus", "the first time I saw Norwich City play and win.." Of course, to voice such thoughts out loud would probably see the withdrawal of any associated marital favours for quite some time. However, since my wife has a highly developed mind-reading skill, I'm out of luck anyway. Oh well...

The ability to keep ones mouth shut at the right moment is a skill learnt only through experience. To know when to open it again and say something that's actually important is a quality to be admired. When you're responsible for security, saying the right thing can be the difference between being respected as a pragmatic and business orientated leader, or being avoided and ring-fenced.

So, to the point: Salesforce.com had an outage last week. It affected almost their entire customer base. It lasted for around half an hour. Almost immediately the ravens were squawking overhead dropping remarks such as "Anyone who trusts their critical business data to the cloud deserves to go bankrupt the next time something happens to any one of the multiple points of failure that are inherent in the system" (see here). 

According to the information at trust.salesforce.com the cause was: a core network device failed due to memory allocation errors. The failure caused it to stop passing data but did not properly trigger a graceful fail over to the redundant system as the memory allocation errors were present on the failover system as well. Clearly a lesson learnt.

So, is this a security issue, or should we accept the fact that occassionally data services, like electricity, will go off-line? It was a power-cut. If the service is so critical then you'll have a fail-over system in the same way as a hospital will have its own generator. If not, then stop whining, light a candle, and wait for the lights to come back on.

This outage does not, as claimed on the headline here "expose cloud's dark lining" because the risk of the occassional service cut are well known, documented and factored into the SLAs. If you can't function as a business when the lights go out and the data stops flowing then for goodness sake don't put all your eggs in one basket and have a back-up plan.


Bookmark and Share


TrackBack

TrackBack URL for this entry:
http://www.computerweekly.com/cgi-bin/mt/mt-tb.cgi/43661

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

About

This page contains a single entry from the blog posted on January 12, 2009 9:45 AM.

The previous post in this blog was Risk assessment and the Nu M8 Child Tracker.

The next post in this blog is Don't lay all the blame for insecure systems on the developers.

Many more can be found on the main index page or by looking through the archives.