Cloud Burst

Warranted or not, the great delicious.com shutdown scare of December 2010 teaches us all an important lesson about the sustainability of cloud services.

If you’re not paying for a product, you are the product.

This quote paraphrased from blue_beetle on metafilter is very apt. Companies that offer free services to their users do so in the hopes they make a return on their investment to run a service. If you’re not paying for a subscription to use a service then you’re paying with your perceived attention through advertising. Let’s ignore the fact that there are multiple ways that Yahoo could have used the delicious corpus of annotated links to increase value across their network. The fact is nothing comes for free and if the owner can’t figure out how, they are within their rights to pull the plug. We all need to be aware of this fact in the same way that we need to read the fine print on any free checking or 0% financing deal we get in the mail.
The old adage still applies. Make sure that everything you put in you can take back out. Delicious has an export feature built into their API so that if you every get twitchy, you can grab your data and take it elsewhere. I would never put my blog onto a platform, hosted or not, that wouldn’t let me pull it back out. Beware of one-way streets.

So all this moving about and looking for a new home for your stuff brings up an old debate. Should you host your own data?

In the past this was not a realistic option for most. The costs and complexity of setting up your own domain and software was far to difficult and expensive. But it’s worth revisiting. The cost of hosting your own server has come down dramatically. Maybe it’s time has come when we can push things to the edge. Stephen Hay challenges us to ask why not:

What if we flipped this all on its head? What if we hosted our own data, and provided APIs for all these webapps so that they can use our data? … So instead of having our own websites aggregate our own data from other people’s websites, we’ll let other people use the data from our own websites. Photos, meaningfully tagged, can be pulled in by Flickr via our own personal API, if you will. We provide the structured data, Flickr provides the functionality. The sharing. The social. Why not?

Imagine a world where I pay $100/year to host all my stuff (blog posts, bookmarks, status updates, photos, videos, etc) which can offset that by services paying me for access to that content. Each service can “pay” me by providing a service that plays to their strength:

• Yahoo pays me for access to my hosted photos with the collective photo tagging and geo-location tools on flickr.
• Google pays me for access to my blog by sending me traffic and offering ad rev share.
• Facebook pays me for access to a feed of my posts and likes by offering a social layer of my friends.
• Twitter pays me for access to a feed of my status updates with distribution.

For the cost of a single month’s cable TV bill I now can pick and chose which service I use and turn off and on each one at will without fear of migration. Today I “host” my money at a bank and point all my monthly bills to that bank, why can’t I do the same with my digital savings? Bundling blog software with a hosting account was a business I helped set up at Six Apart. Adding email, photos, videos, and bookmarks and putting a nice user-friendly front end onto it could be a real opportunity for an clever hosting provider.

One final thought. If each of us hosts our own data, companies would be much more likely to standardize how they integrate with our data and make it easier to mix and match datasets. It’ll be in their interest to offer the best tools for data to flow in and services to flow out. With users and data aggregating into just a handful of large players, it is not in a company’s interest to offer these tools, it’s better for them to lock users and data up to prevent loss of audience and attention used to monetize those users.

That’s a topic for the next another post (which is posted here)