Scalability Design Explorations around changing Payment Flows about zips HOT 1 OPEN

As a starting point for exploring changes from the current Zcash architecture: what if senders somehow deliver most transaction data directly to recipients, with only minimal commitment and nullifier state stored / validated globally?

This impacts trade-offs roughly like this:

• It loses some important data availability guarantees: if the recipients lose their data, they can't recover it (unless the senders resend it?)
• It gains scalability by minimizing the amount of global data. For example, users now could send very large "private memos" since those no longer are globally stored.
• It loses offline reception.
• It degrades privacy if it's possible to observe the transmission from sender to recipient.
• It may introduce consistency failures: what if the txn is committed, but the recipient never receives it or forgets it?

Maybe we can address some of the shortcomings in other ways? Perhaps a mixnet can protect network privacy and a "private mailbox" can provide data availability and allow offline reception.

from zips.