[Lightning-dev] Research on proactive fee free channel rebalancing in the friend of a friend network / and roadmap for a protocol extension

[ZmnSCPxj]

Good morning again Rene,

An observation I would like to make is that, as I understand it, the model inherently assumes that all nodes are equally likely to be payer and equally likely to be payee.

While in a complete economy this is to be expected (the "customer" of a "merchant" will be an "employee" of some other "merchant") in an equilibrium state, it seems to me that transient states might exist where some nodes will demand more incoming funds.
For example, consider a manufacturer that has just released a new model of their product; generally the expectation is that customer interest in the new model will be high when just recently released, thus transiently such manufacturers will need more incoming liquidity than in equilibrium.

Thus I think JIT Routing is still more useful in practice than a proactive rebalancing policy, as it is more resilient to such transient shocks to the system.
Every rebalance is effectively a speculation that capacity in one (set of) channel(s) is less useful than capacity in another (set of) channel(s).
Targeting some kind of "perfect balance" assumes that the equilibrium state will be that payments will generally be equal in all directions (an equilibrium state that I expect to be true, actually, once entire economies switch to the Lightning Network), but I think we should provision our nodes to also be resilient against transient shocks.

Though I could also be overestimating the effects of such transient shocks, in which case equilibrium probably implies that targeting a perfect balance proactively would be better.

Regards,
ZmnSCPxj