Metrics - Avg Transaction Amount, Open Blocks Rate

A brief search didn't turn up past threads, but has there been discussion of nodes tracking average tx amounts and/or the rate of new account creation/open blocks?

Seems this is doable and would provide some benchmarks for PoW difficulty adjustments, acct/pending deletion/archiving, and if ever needed min. acct. balances.

Large transactions would need to be capped to prevent bad actors from skewing the avg TX amount upward. Zero provides the downward bound.

Acct creation could be based on Moore's law or some other reasonable estimate of future hardware capacity.

1 Like

Can I judge the silence as a "no" or as a lack of interest in this idea?

Maybe some examples will help.

Accounts with frontier blocks "X" months old or older and a balance <1% of current average transaction amount can be dropped from the ledger.

"Recent" account creation has exceeded 10x's the desired/expected account growth rate. Increase account PoW required to "X" times current network PoW.

Network is saturated and block transaction amount is <0.1% of recent avg transaction amount. Increase PoW required to "X" times current PoW. "X"^2 if <0.01%.

Values deserve further investigation, but the idea is to have some additional metrics to determine when there are a significant number of transactions happening that are well outside the normal network function (i.e. likely spam). I believe these could be framed in a way such that honest users aren't penalized under normal conditions for opening an account of marking a small transaction. All using the existing spam mitigation tool of DPoW.

This is an interesting idea, in-line with PoW multipliers for different types of transactions, but based on network activity instead. I'm imagining something like, "for every 10 open_block_cps, increase PoW by 100%". Or just do it for all transactions (100% PoW for every 10 CPS), since you can send to unopened accounts, and it could act as a pseudo base PoW balancer for Moore's law + time. It wouldn't really be "base PoW" though, since it could only be used for prioritization/DPoW (due to different nodes seeing different CPS)

The average transaction amount metric would be difficult to use, since it can be skewed both ways (too high and too low), and there could be legitimate reasons for both

Experimental pruning already takes into account time, so usable metrics would just allow for more complex rules/conditions for pruning. Not sure that's necessary or worth the effort though

That being said, I'm not sure this idea targets the real issue. Even with the recent 70+ CPS spam, the network as a whole was not out of capacity. Should we be actively trying to increase the costs/"fees" if the network is not really saturated? Pre-emptively increasing costs because low end nodes get out of sync seems like somewhat of a backwards solution to me. Maybe this idea addresses the wrong problem?

You mean intentionally skewed by a bad actor? I do believe a "cap" on how large TXs would be needed for this reason.

Are there legitimate reasons why the network avg tx amount (say based over 2 weeks) suddenly drops 90% or rises 1000%?

Thinking this has utility in that is provides a loose metric for acct values in non crypto currency.

I envisioned these items only kicking in if the network was saturated and/or new acct growth is clearly abnormal. Regarding the last spam attack, isn't it hard to rationalize millions of new accounts being created per day? The above would mainly be focused on ledger size control.

I still believe DPoW works fine for run of the mill "loop" spam attacks.

Thanks for the thoughts, look forward to more.