search

storacha-network / project-tracking

🐾 Used as central/default repo for project management, backlog, etc.
0 stars 0 forks source link

Design and spec out generic billing #62

Open reidlw opened 3 months ago

reidlw commented 3 months ago

We need to design billing solution that is decoupled from individual operations for which we bill. There is RFC draft here https://www.notion.so/w3sat/Align-on-allocation-client-signal-to-upload-the-bytes-and-billing-with-Blob-3b6fa0373aa444ad98802e492a8f6941

reidlw commented 3 months ago

@Gozala - I'm kind of confused here. Title of this is to design and spec out generic billing. And description links to a notion doc that seems to be specific to making billing work properly going from store/add to blob/*. Then there's a statement at the top of that notion page linking to this merged PR for implementation: https://github.com/web3-storage/w3up/pull/1342#pullrequestreview-1982238270

So help me unwind... Is there still an issue needed in the backlog here? If so, could you explain a bit more what this is independent from the changes around blob protocol?

hannahhoward commented 3 months ago

@reidlw I believe the ultimate decision was to implement billing for blobs the same way we do for store/add and not add a generic billing approach. That work is now tracked here: https://github.com/w3s-project/project-tracking/issues/48

I am guessing this was to do follow on work to really figure out what generic billing, presumably something that we might want in the future related to egress billing (plus other billing -- most platforms like S3 charge per operation), would look like.

Gozala commented 3 months ago

Exactly what @hannahhoward said

reidlw commented 3 months ago

@Gozala @hannahhoward - thank y'all for the extra context. I'm struggling to know what to do here. Egress billing is super important but tracked elsewhere. Aside from that I'm not aware of any business driver for major re-working of billing, thus would probably move to icebox for now. WDYT?

hannahhoward commented 3 months ago

I am inclined to icebox this as a ticket, and see if the design of egress billing indicates the would be an ideal and fast way to implement it, at which point I'd pull it out.

That said, we are going to have to do a lot of intentional design when we introduce pay per operation via crypto, and I wonder if that is the ideal moment to really look at how billing works. (mind you that design work is happening in a month or two anyway)

hannahhoward commented 3 months ago

long and short, icebox it for now.

Gozala commented 3 months ago

I think I would make this part of the egress billing personally, perhaps because I'm more familiar how it works now and know it's not going to work out without putting proper thought into this.