The problem with cloud computing is two fold.
One, the assumption based on the marketing spiel and hype that surrounds it that nothing will ever go wrong with data stored on and accessed via the Internet and web. Two, when something awry does occur, the almost complete inability for cloud services and products to continue functioning with even reduced functionality. Google Gears and Adobe’s AIR may offer some relief, but the WordPress outage for over 2 hours this morning (SL time) was a sobering reminder than reliance on hosted / cloud solution comes at a price – when things go wrong, one is at the complete mercy of the service providers.
CMEV’s blog went down with the WordPress outage. We’ve been getting a fair number of page views on it after yesterday’s interesting elections in the Eastern Province. It was annoying to see that WordPress was down this morning and even more so when the time given for downtime turned out to be a gross misrepresentation of the time it actually took to get WordPress.com up and running again (messages on my account kept saying 21 minutes more, went down to 1 minute and then kept going up to 18 or 19 minutes for around 2 hours!).
With everyone talking about cloud computing usurping our desktop centric storage / access / dissemination paradigm I really wonder if we’ll hit a stage in which downtime is guaranteed to not occur, or whether it is possible with science to give such a guarantee?
While we wait, some short common sense strategies for redundancy and the preservation of sanity during downtime esp. for ICT4Peace applications:
Diversify media. CMEV also has a Twitter channel that I was able to post updates on the outages to. CMEV also hosted it’s maps with election violence updates on Google Maps. Combined with Twitter, I could have run the updates even without WordPress. I also used Twitterfeed to update the Twitter channel with content from the CMEV blog automatically. I also hosted the podcasts outside WordPress on Internet Archive, giving access to them even when the site was down.
Diversify access. To the extent possible, make sure critical applications aren’t reliant on a single ISP or connection. I used SLT ADSL and SLT 3G HSPA interchangeably over election day to update the CMEV sites with content.
Backup and mirror. All CMEV sitreps and incident updates were mirrored on the old CPA site.
Email. Plain old email updates work just fine when you want to alert a few key people on new content, who then spread the word in a viral fashion.
Expect the block and outage. Plan at the outset for the worst case scenario.
Use RSS. It helps gets the message out. Don’t expect people to come to your website to get the news.
Don’t be old school and stingy. Full RSS feeds may reduce the number of those who visit your site, but they get the content out in a manner that’s accessible even if the site goes down for a bit. Combined with a service like RSSFwd, can be a powerful way to disseminate content that makes it virtually impossible for Governments to clamp down on (unless they go the route of Myanmar, but that’s it own defense).
Go mobile. Make your site accessible to mobiles. It pays off. There are a couple of plugins for WordPress that do this if you host the site on your own. Mofuse is a great (free) service is you use a hosted WordPress (or any other blogging) service / site.
Need to write a book about these strategies in more detail sometime…