Skype, Webservices and System Redundancy

So, like the rest of the planet we have had a Skype-Outage for a lot of the last few days....its sort of back up now, with a few people appearing and disappearing here and there. However, I read of companies that have stopped using landlines and are now in dire straits. What were they thinking?

There is an engineering (as opposed to management) view of redundancy, i.e. that extra options are good. Especially for networks, where the last thing you want is to lose service if a network drops out.

So, for comms that means having at least one other option - and better a non Internet one in case you lose all IP. And better also a non wire backup in case the phone line goes down. For full resilience we use VoIP, IM, POTS and Mobile communications. So should you if you depend on comms in any way.

The same rule also applies with Webservices, which is why we don't recommend any service that does not allow you to do stuff offline. So, yes all these webwidgety things are dreadfully fashionable right now, but let Skype Out be a warning.

Think about what you are using and what happens if:

(i) You are offline, and / or
(ii) The Webservice loses your data

In addition, think about what happens if the webservice exposes (or abuses) your data. It can, and does, happen, and will happen increasingly as more people put more valuable stuff online, making it worthwhile to hack it. For this reason, keeping your stuff in more than one place is good.

Here endeth the lesson...the weekend beckons

Postscript - read quite a few comments on various blogs moaning about Skype.....our reply would be that if you put all your eggs in a free consumer service...well, you gets what you pays for.....

Post-Postscript...Om Malik asks where were eBay during all this. Certainly not blogging about it.....maybe they only had Skype phones as part of the drive to get synergies on their $2.6 bn acquisition

Post Post Postscript - it would appear it was mass reboot of PC's following a Windows patch wot did it. This clearly has interesting implications for what people who have infrastructure layer assets can do to application layer plays.....yet another reason for going for redundancy.