• 0 Posts
  • 4 Comments
Joined 1 year ago
cake
Cake day: June 16th, 2023

help-circle

  • A problem I haven’t seen anyone discuss: What of server costs?

    When even just 1% of the still growing 30 MILLION Threads users (300k) are interacting with an average Mastodon, Lemmy, etc. instance every day, just how much data do you think that will generate? As Threads scales and users are posting content that users on smaller instances try to interact with, the hosts of these smaller instances bear the brunt of the costs.

    Threads need only exist, and as everything scales upward and more people join the Fediverse, their sheer mass will wipe out all the smaller players just by virtue of the smaller servers being unable to cover growing server costs. 300k users creating 1kb of content each, per day, comes out to 292 megabytes of data. (But that’s not realistic. The OP contains 5,171 characters, or roughly 5kb of data.) This does not account for images, or videos, which also cost money to store. If 1% of those 300k users (1% of the 1%, 3,000 people out of 30 million) are posting images, if we assume the maximum file size Mastodon can store (8mb per image) and arbitrarily set the file size at 1mb to try to be conservative, we’re still adding an additional 3,000 megabytes of data per day in addition to the original 292, or 3.21 gigabytes of data. We’re not even yet accounting for the additional data to store the database references for all of this either, keep in mind.

    Those numbers are small. They don’t include videos, and they vastly underestimate the amount of users interacting with any of our smaller instances. Every time a reply containing an image or video is posted to Threads, if smaller instances want to keep a copy for their own users to reply to or interact with, they have to store that data.

    Server owners will be buried under the server costs – costs which Meta can easily subsidize with Instagram and Facebook revenue, not unlike Walmart intentionally under-pricing everything in a new branch in a small town right up until every local store ceases to exist, at which point they jack up the prices and put another new store somewhere else.


  • You’re looking at this from the wrong point of view. The fediverse is not just lemmy: Threads, Tumblr, even BlueSky (albeit with their own protocol, but anyone could just modify their fediverse enabled app to convert their data to be applicable to BlueSky’s protocol) are quickly setting the stage for a new norm. The more websites integrate the fediverse into their stack, the more data outside the immediate sphere of influence of these major corporations can be harvested. To what ends they’ll use it, I don’t know – but I don’t trust them with it.