The cloud is falling... erhh... wait...

[Jan Kyster]

But I knew it!

A disaster just waiting to happen.
The idea of having everything outsourced and having access to everything from everywhere on the globe sounds intriguing, but can we really trust the web - stability wise, I mean?

What if... here's some recent news (yesterday):

http://news.yahoo.com/s/atlantic/20110421/tc_atlantic/cloudconundrumwhatamazonsservercrashrevealsaboutdo ingbusinesscloud36892
http://www.bnet.com/blog/technology-business/amazon-outage-shows-how-web-20-becomes-web-00/10157
http://newenterprise.allthingsd.com/20110421/amazons-cloud-crashed-overnight-and-brought-several-other-companies-down-too/

:shifty:

[Torplexed]

I imagine those companies that still possess their own servers are breathing a big sigh of relief right now. It's 24 hours of failure and counting at this point.

[Pisces]

Rule one in IT: Always have a secondary system as backup if your life depends on it! So if every one is consolidating it's services on the servers of Amazon then you're asking for trouble.

[Fish In The Water]

To quote a wise old sage: "The more complicated you make the plumbing, the easier it is to plug the drain." :D

Or something to that effect...

[TarJak]

Hosting is hosting regardless of who does it. Outages will occur regardless of who does the hosting. The real problem is how to recover quickly when a problem occurs, and that depends on what the problem is.

The length of the outage is always the issue. Having redundancy is OK as long as it takes you less time to bring it up than it does to recover the primary.

Everyone thinks that disaster recovery is a simple, flick switch affair, which, when it comes to DNS and TCP/IP addressing changes as part of DR, is an absolute crock. It takes time for these changes to be recognised around the world so most times getting your primary working again is almost always quicker than moving to a failover.