Social Media Architectures and Their Consequences
This is an outcome of a session I ran at last week’s “Logging Off Facebook – What comes next?" unconference. We explored what technical architecture choices have which technical, or non-technical consequences for social media products.
This table was created during the session. It is not complete, and personally I disagree with a few points, but it’s still worthwhile publishing IMHO.
So here you are:
|Facebook-style ("centralized")||Mastodon-style ("federated")||IndieWeb-style ("distributed/P2P")||Blockchain-style|
|Moderation||Uniform, consistent moderation policy for all users||Locally different moderation policies, but consistent for all users on a node||Every user decides on their own||Posit - algorithmic smart contract that drives consensus|
|Censorship||easy; global||one node at a time||full censorship not viable||full censorship not viable|
|Software upgrades||Fast, uncomplicated for all users||Inconsistent across the network||Inconsistent across the network||Consistent, but large synchronization / management costs|
|Money||Centralized; most accumulated by "Facebook"||Donations (BuyMeACoffee, LiberaPay); Patronage (Patreon)||Paid to/earned by network nodes; value fluctuates due to speculation|
|Authentication||Centralized||Decentralized (e.g. Solid, OpenID, SSI)||Decentralized (e.g. wallets)|
|Advertising||Decided by "Facebook"||Not usually||Determined by user|
|Governance||Centralized, unaccountable||Several components: protocol-level, code-level and instance-level||Several components: protocol-level, code-level and instance-level|
|Search & Discovery|