QUOTE(Tarball @ Feb 15 2025, 00:53)

Is there a rough guideline for an optimal "storage to data transfer limit" ratio for a H@H node?
Like between a node with large disk space (1-2 TB) but lower data transfer limit (<5 TB), and lower storage (100-200 GB) but lots of data transfer (10-20 TB), would one be more beneficial (or generate more Hath)? Obviously unlimited data transfer is ideal but not always possible.
Hard to give a concise answer, since it varies a lot by region and how old the client is. Traffic per range varies by a factor of 100 between the lowest-traffic region and the highest-traffic one. For the low-traffic regions like Oceania and South America, you won't get much traffic unless you have a lot of ranges, which means you need a lot of disk space.
These numbers are a bit higher than the real ones, but I can give some rough numbers for the main regions based on the bandwidth limiter estimates. For Asia, clients average around 8.5 KB/s per assigned P1 range, each of which requires 500 MB of assigned space. Followed by North America with around 2.4 KB/s per P1 range, and Europe with around 1.4 KB/s per P1 range. But there won't generally be enough assignable P1 ranges to make full use of a large cache, so a fast connection with a smaller SSD will probably perform better than a slow connection with a larger HDD.