And yet, this was already established.drew.phillips wrote:- Peering with other news providers. You must establish relationships with other platforms to send/receive messages and for what hierarchies you exchange with each other. For example one peer may agree to send you everything from a particular sub, like sci.* and not for others. It's a complex set of relationships that dictates what messages you will and won't receive from other peers. It's a complex, distributed, server-to-server type of network.
And yet, this was not an issue since the vast majority of the retention was stored remotely.drew.phillips wrote:- Retention. A few TB of storage could be a huge understatement. Depending on your retention period and subs, this can be in the petabytes. Even consider text only, with a hypothetical 30 MB of news per day to be retained for a period of 1 year (low) would require over 10 TB of storage (just for the content itself - not to mention filesystem overhead etc).
Low quality of service is better than no service. Again, this was not an issue.drew.phillips wrote:- Network bandwidth. You have to consider (even with a small number of users) the fact that a handful of people may pop on at the same time every couple of weeks and download a huge amount of content. If you serve media content in alt.bin imagine a few users on high speed connections trying to download a movie at the same time. They can either get their content quickly (requires a large amount of bandwidth) or it can take forever, resulting in a low quality of service.
All things considered, what Sonic had worked quite well and didn't require building an entire "modern Usenet platform" nor require pulling "substantial staff resources" to maintain. What's the real reason for killing Usenet at Sonic?