Tumblr’s most recent outage lasted for over a day, and while that may not seem to be very long, a day on the internet is an eternity of bad PR. It highlights the main problem with a _”cloud”_ system— it’s *quite vulnerable* and can screw everyone over if not managed properly.
For those unfamiliar with the term _”cloud,”_ it refers to having all the data stored on a remote, third-party server, rather than locally or on a private server. Tumblr is an example of this, the whole app along with its databases is stored on Tumblr’s own servers. This simplifies everything for less tech-savvy users, but it brings a risk.
If, for example, I screw up a file or database on my own installation of WordPress, I’m the only one who gets fucked over. Compare this to a system like Tumblr, where one mistake fucks over *everybody* using it. Not very reliable at all, especially if you host your portfolio/business on Tumblr, like Cameron Moll. Imagine the reaction a client would get if they saw a “We’ll be back soon” on your portfolio, rather than the content they’re interested in. They’ll just move right on to somebody else with a functioning site. Sure, that’s a bit exaggerated, but it’s an entirely realistic scenario.
I’m not saying don’t use Tumblr (unless your blog consists of memes or has “fuckyeah” in the title), just be careful. If you host important information or well-written, carefully thought-out blog posts on it, they could all be lost in the blink of an eye. Worse still is that Tumblr offers no easy way to export entries, and Google Cache can only go so far. And these are only mere blogs, imagine what would happen if entire OS’s stopped working because of a database error on a central server.