essay, Uncategorized

A critical look at decentralization

Hey federated social web people! I read a great paper yesterday: A Critical Look at Decentralized Personal Data Architectures. If you’re working in this space, you should check it out. It’s a quick read, just five pages.

It comes from some pretty high powered authors, including Helen Nissenbaum and Dan Boneh. They step back and look objectively at accepted wisdom in our community, such as “people should own their data” and “decentralized and open always win out,” and attempt to reconcile with the fact that centralized social networks have exploded, but long-standing decentralized alternatives like OStatus have not.

Everyone wears rose colored glasses occasionally, including our own community, but we shouldn’t overindulge them. The network effect, ie “all my friends are on Facebook,” isn’t the only adoption hurdle. There are real technical, economic, and product drawbacks to decentralized social networks, and they matter to users.

I’m fond of the Indie Web movement, which encourages us to focus on the positives and evangelize having your own web site. However, we should be honest about the negatives. That’s the first step toward allaying FUD, talking openly about the tradeoffs, and maybe even solving some of the real problems.

If you’ve already read the paper, any thoughts?


11 thoughts on “A critical look at decentralization

  1. awesome, thanks! sad timing w/the app engine outage, but oh well. i was glad to hear evan talking about that summit, looks like a high powered group. i’ll have to get a brain dump from you sometime. via Google+

  2. Looks like a good one.

    I think one if the issues is YET.

    Every existing proprietary system in the past has fallen away to the Internet.

    Compuserv, AOL, etc.

    One of the confusing issues here is cloud computing.

    It isn’t just the decentralized aspect of data but there is the fact that Facebook is free, scales well, and does a lot of it in the cloud.

    The solution to this might be for hosting providers to boot up EC2 nodes for their users directly.

    The problem THEN is how to pay for them…

  3. i wish i agreed! sadly though, i’m not sure history is on your side.

    yes, early functionality like email and information browsing transitioned from proprietary to open, but that trend stopped around the time mainstream consumers and businesses adopted the internet. instant messaging, payments, and shared document editing are just a few examples that have been around for a while. sure, we have jabber and bitcoin and wave, etc., but they’re all still bit players. AIM, paypal, and google docs (or whatever) still largely own those markets.

    you’re right to think about facebook’s other advantages, many of which are arguably due to centralization, and alternatives like VM-per-user. the paper discusses all of those ideas and many others.

    let me know what you think of it!

Leave a Reply

Your email address will not be published. Required fields are marked *