Mastodon is a Fediverse platform. Lemmy is too. Anyone can host their own Mastodon and Lemmy instances.
When a user uploads an image or video on Mastodon instance 1, and a user of Mastodon instance 2 is following them, that image or video is copied over to Mastodon instance 2 - because that’s where that user resides.
This means content gets replicated and duplicated across every shared-network instance. Resulting in resilience, but also exponential, excessive storage needs.
OP is suggesting that media files should be shared across platform instances so that they don’t get duplicated many times. This would significantly reduce storage and bandwidth needs and use for the platform instances themselves, offloading and centralizing media file concerns.
When a user uploads an image or video on Mastodon instance 1, and a user of Mastodon instance 2 is following them, that image or video is copied over to Mastodon instance 2 - because that’s where that user resides.
The same thing happens with text posts, right? I don’t see an exponential expansion, just linear in the number of nodes. It sounds like the decentralized way to do things. Hmm. Anyway, thanks for the explanation. It saves some storage but doesn’t save bandwidth, it sounds like. Rather, the bandwidth requirement gets concentrated at the shared server.
Depends on the implementation. OP talked about potential of a CDN serving the shared resources. The instance servers wouldn’t have to proxy the content. Which would allow caching and not duplicating content and transmission through multiple endpoints.
Yeah, I guess it’s linear rather than exponential in growth. From an instance hoster point of view, it just never ends though, and not very predictable.
Mastodon is a Fediverse platform. Lemmy is too. Anyone can host their own Mastodon and Lemmy instances.
When a user uploads an image or video on Mastodon instance 1, and a user of Mastodon instance 2 is following them, that image or video is copied over to Mastodon instance 2 - because that’s where that user resides.
This means content gets replicated and duplicated across every shared-network instance. Resulting in resilience, but also
exponential, excessive storage needs.OP is suggesting that media files should be shared across platform instances so that they don’t get duplicated many times. This would significantly reduce storage and bandwidth needs and use for the platform instances themselves, offloading and centralizing media file concerns.
The same thing happens with text posts, right? I don’t see an exponential expansion, just linear in the number of nodes. It sounds like the decentralized way to do things. Hmm. Anyway, thanks for the explanation. It saves some storage but doesn’t save bandwidth, it sounds like. Rather, the bandwidth requirement gets concentrated at the shared server.
Depends on the implementation. OP talked about potential of a CDN serving the shared resources. The instance servers wouldn’t have to proxy the content. Which would allow caching and not duplicating content and transmission through multiple endpoints.
Yeah, I guess it’s linear rather than exponential in growth. From an instance hoster point of view, it just never ends though, and not very predictable.