Surge in transactions + 1 MB blocks = what?

[Cubic Earth]

It's November, 2015.  Recent media attention and a rapid rise in the price of BTC are likely causes of the daily transaction count decisively moving above 150,000.  There looks to be a good chance of exceeding 200,000 daily transactions by the end of the month, if not sooner.

There will soon be more transactions submitted to the network than can be cleared by blocks.

It's clear that a transaction fee market for will develop, for better or worse.  But what about mempool growth?  And are there other backlog issues?  The stress tests came and went, but we might be entering a sustained period where transactional demand exceeds the networks capacity.  What will this look like with nodes and network we have today?

[1713268466]

1713268466

[achow101]

It's clear that a transaction fee market for will develop, for better or worse.  But what about mempool growth?  And are there other backlog issues?  The stress tests came and went, but we might be entering a sustained period where transactional demand exceeds the networks capacity.  What will this look like with nodes and network we have today?

The Core Devs are working on a solution for the mempool issue (in fact it has already been merged to the master branch). This fix will officially be released in 0.12 sometime in early 2016 (feb I think). Basically the fix limits the mempool size and it will drop the transaction with the smallest fee in order to keep the mempool at a certain size which can be user defined.

Currently, the only fix is to change the minrelaytxfee which will prevent transactions under a certain fee and dust threshold from being accepted to the mempool and from being relayed by that node.

[nicked]

Maybe they should just raise the blocksize limit.

[Was]

We knew this day would come. Round Two will be soon.... perhaps people will be more ambitious in implementing a solution this time around.

[silverleafy]

They are having a Scaling Bitcoin conference in Hong Kong in December, but the likelihood of anything actually being done an consensus being reached is pretty low. I think part of the problems is that of the current proposals, only a few of them have actual implementations of what they are proposing.

[Cubic Earth]

It's clear that a transaction fee market for will develop, for better or worse.  But what about mempool growth?  And are there other backlog issues?  The stress tests came and went, but we might be entering a sustained period where transactional demand exceeds the networks capacity.  What will this look like with nodes and network we have today?

The Core Devs are working on a solution for the mempool issue (in fact it has already been merged to the master branch). This fix will officially be released in 0.12 sometime in early 2016 (feb I think). Basically the fix limits the mempool size and it will drop the transaction with the smallest fee in order to keep the mempool at a certain size which can be user defined.

Currently, the only fix is to change the minrelaytxfee which will prevent transactions under a certain fee and dust threshold from being accepted to the mempool and from being relayed by that node.

Thanks for the synopsis.  The minrelaytxfee seems pretty straightforward to understand with respect to its effects on network.  If a tx does not meet the threshold, the tx will not propagate.  Is the behavior of the network if nodes have very different mempools well understood?