I killed the server. People are still struggling with basic transacting, so I didn’t like the idea of adding another obstacle to getting transactions confirmed.
When it was operational, it was opt-in and over tor. Usage wasn’t enough to make it worthwhile yet. There is less than 1 transaction per minute in Grin, so even if every transaction used GrinJoin, the server would have to wait 15-30 minutes to have any kind of meaningful aggregation.
@oryhp I have not yet, I will soon.
@david I think the only viable way to do aggregation with current usage is to make aggregation windows of 12 hours, even up to a week. My preference is 2-day periods. It could:
It would mostly be Intended for users to self spend, not for real economic transaction. If someone can point me to how it can be done with an easy UX (and not too complicated to build), I’d love to make it happen and maintain it.
The downside of fixed-size aggregation is that it could be effectively (although at some cost) undone by broadcasting lots of dummy transactions around the same time.
A fixed time window doesn’t suffer that downside. Perhaps it would make sense to have separate hourly, daily, and weekly aggregation servers, to accomodate different tradeoffs between speed and traceability.
Yea, only time-based aggregation windows should be used. Fixed-size aggregation requires much higher fees to prevent dummies.
At this point though, running a coinjoin server is little more than an academic exercise. We know coinjoins are possible (in fact, trivial in mimblewimble). And nobody using Grin at these usage levels should expect any level of privacy. The best way to improve privacy right now is through increasing adoption or someone coming up with a clever crypto solution.
Starting up my GrinJoin server just creates more work for me with no practical gain. When we see more adoption, I’ll fire it up again.