Bitcoin has two major kinds of graph obfuscation tecnhiques: coinjoin and coinswap. The former aggregates regular transactions, while the latter swaps around ownership of coins.
The proposed aggregation services for Grin similarly fall into two camps.
The original proposal was for a coinjoin service, while this thread proposes a coinswap service.
Each has their pros and cons. Coinjoins require no extra bandwith/fees but require patience and a backup publishing mechanism if the service fails.
Coinswaps can be run at leisure right after or inbetween regular transactions but require extra bandwidth and fees. While a weekly coinjoin service makes little sense, this would make a reasonable option for a coinswap service.
Having minimal impact on the number of kernels is very nice. A final advantage is that mathematically modeling a coinswap service and its requirements is much easier, which may boost its research and development.
On the whole then, I tend to prefer the idea of a coinswap service, and to focus our attention on those first. Once coinswap services are succesfully deployed, we can come back to revisit additional coinjoin services.
6 Likes