Achieving Scale in the Decentralized Web

[Paul Frazee]

One of the on-going questions of decentralized tech is, how will it scale? Which is another way of saying, is it viable?

The Web has always been about scale. If we’re not able to produce Web-scale applications, then it’s time to pack up and go home.

This post covers how P2P Web applications can match the scope of the .coms, while still maintaining the valuable attributes of decentralization. It does this with a hybrid P2P/Service model.

http://pfrazee.github.io/blog/achieving-scale

Many people here probably know my previous project, before Beaker, was called Patchwork. It was the hypothetical "Fritter" discussed in the post, and it had the problems of Fritter. We found various solutions, like automatically downloading the messages of Friends of Friends, but it worked too inconsistently for my tastes. (I've since fallen out of sync with the project, but they're extremely smart hackers, so I suspect they've improved on performance since my time there.)

With Dat and Beaker, a lot of my past concerns are solved. (I briefly cover the differences here.) Combining pure P2P with crawler services gives a solid path to scale, and makes me confident we can tackle the centralized web.

-prf