backups should be local, not to the cloud

One of the lessons of Friendfeed’s buyout by Facebook is that the cloud is not a good place for backup. In an era of the sub-$100 terabyte, the idea that the best place for our data should be anywhere other than right at home is a strange one. Cloud backup is useful as a meta-backup – for example, using Jungledisk and Amazon’s S3 service to backup your local backups – to guard against catastrophe, but should never be your primary repository.

For data like photos, this is pretty much a moot point, as everyone keeps their originals on their disk and uploads select photos to Flickr/Picasa etc (and at lower resolution than the originals). But for text, like blog posts and tweets, most people simply leave their content in the cloud – which includes leaving your wordpress database at your hosting provider rather than on your local disk. I haven’t yet found a good solution for local wordpress database backups but I have written previously about various backup strategies for twitter. Sarah Perez at RWW just did a piece on 10 ways to archive your tweets as well, but most of these are again cloud-based solutions. Marshall Kirkpatrick has a guide to using Google Reader along with Dave Winer’s new OPML tool to consolidate all your tweets and your friends’ tweets, but this isn’t a true backup solution either, as I point out in comments. The point however is as long as the data is in the cloud, it’s not really backed up – data wants to be imprisoned, not set free.