Quote from: Holliday on September 16, 2015, 12:27:34 AM

Quote from: adamstgBit on September 15, 2015, 08:56:51 PM

should bitcoins traffic limits be based on what 10% of typical home connection can handle?

maybe 20% ? maybe 50%? surly anything demanding 20-10% isn't going to impact number of full nodes.

in anycase i feel that there can be much improvement as to how data is shared across the network to dramatically decrease bandwidth use for full nodes, which should allow for proportionally bigger blocks.

I don't have the answers to your questions, but I will ask you a few instead.

Should I be able to run a fully functional node while using the same internet connection for other daily household uses?

I pay near $100 per month for the highest tier internet speeds my ISP will offer.

I've built a dedicated machine to run Core (modern quad-core CPU, 16GB RAM, modern SSD).

To give you an example of my resources:

Quote from: adamstgBit on September 15, 2015, 09:26:28 PM

i'm currently syncing a full node and it feels like my computer is about to explode! cpu is running hot! hot! Hot!

You are still catching up from a partially synced state after at least 24 hours? I can sync the full chain from scratch in about 14 hours.

Yet, I can't run a full node without "putting the brakes on".

I'll ask a second question. Will a larger block size limit (which we agree will lead to larger blocks) make it easier or more difficult for me to run a full node?

I already have to gimp my node in order to have a functional connection to the internet. I already have dedicated hardware in order to keep my daily use computer "up to speed". What future steps will I (someone with fast computer / fast internet) need to take in order to keep running a full node?

I guess I could cross my fingers and hope my ISP (which has zero competition in my area) to just pump up my bandwidth for no additional cost.

I could continue to gimp my node when I know the network would happily accept more data from me.

Or, should I just give up on running a full node?

Quote from: adamstgBit on September 15, 2015, 08:56:51 PM

I know someone that turns on his full node like once a week, just to download the lastest blocks and then turns it off again, so that when he makes a TX it doesn't take him long to sync up first. is this node useful? should ALL users be asked to run a full node?

I haven't done a Bitcoin transaction in over 2 years. Yet, I have multiple full nodes with damn near 100% up time. Maybe I'm the sucker for continuing to bother. If it ends up taking even more resources, I may simply have to stop.

I'm currently catching up from a about 1 year of down time. anyway:

"I already have to gimp my node"
dose gimping your node reduce its usefulness??
that's the thing! there is a real demand for some kind of simi-fullnode.
which is why i believe there should be different levels of nodes

SPV, simi-full node, full node, supernode, miner node.

currently the network is configured in a completely random order, with everyone relaying TX's to everyone else, increasing bandwidth requirements needlessly. and on top of that the way new blocks are propagated also increases bandwidth needlessly.

If improvements were made such that your full node requires X4 less bandwidth would you not agree that a 4X block increase would be OK?

block propagation can be improved up to 250X faster ( there is a running network which currently utilizes this improvement, sadly your node does not currently benefit from this ) there are probably all kinds of other improvements to be made.

you could say you are already running a sort of simi-fullnode because you limit its bandwidth usage. it would be much better if a real simi full node was developed and optimized to sever a specific function on the network.

bottom line, we need improvements, But still the answer is No you shouldn't aim for running full nodes while running a full household worth of devices. I believe we should keep the limit such that a typical home connection can handle it with about 80% of the its bandwidth being utilized. this will be the upper limit which should never be crossed. and at the same time offer simi-full nodes that do something useful with only 10-20% of a typical home connection.

some improved incentives will also help, what's the point of running a "super node" that relay GB's of data per day? maybe miners would be willing to pay a small fee to connect to this node so that it gets blocks faster? or something!