Topic: Bitcoin 20MB Fork - page 113. (Read 154790 times)

I dont know if someone actually made a similar poll since search engine is down so here we go..

Bitcoin fork proposal by respected Bitcoin lead dev Gavin Andresen, to increase the block size from 1MB to 20MB.

Interesting read about it: https://bitcointalksearch.org/topic/fork-off-919629 (locked )

Feel free to discuss it further.

You mis-interpreted my post. My point was with 20MB blocks, you will no longer be able to host a node on consumer Internet Plans. The important number was not the 2Mbps download, but the 4Mbps upload. BTW, to get those numbers, I simply multiplied my former node's bandwidth usage by 20.

My current ISP has reduced the speeds it is offering for consumer access. The fastest upload offered on a consumer plan is 3Mbps. Business plans have an option with 5Mbps of upload. (Experiments like 250Mbps down and 15Mbps up are being grandfathered)

Telus, the local phone company, is better with it's VDSL offerings. You can get 5 or 10Mbps of upload speed. For independent providers (using the same lines), like the one I was with, your maximum upload speed is 5Mbps. To get 10Mbps or more, you have to go to fibre, costing $$$$. VDSL and cable may technically be able to handle that, but the companies involved don't want to cannibalize their lucrative fibre offerings.


My overall point is that if we go ahead with this fork, we have to do it with the understanding that the era of running a full node on a consumer Internet connection will be over. Due to extra CPU overhead (due to transaction processing), the days of running on a VPS may be over as well.

Now, with 20MB, I expect dedicated hobbyists and small businesses will still be able to run a node, but it will be a considerable expense: hundreds if not thousands per month. Webhosting may be cheaper, but then you can't keep an eye on your box; or plug hashers directly into it.

Bitcoin can be changed in a backward-incompatible way and still remain Bitcoin. It was done by Satoshi with the version checksum change, for example.

Hopefully there won't be any huge hard fork controversy in the future. It'd be a big mess if people had to actively decide between one fork or another. If this does happen, then I will endorse the most correct version of Bitcoin, and this version is what I'll mean when I say "Bitcoin". In particular, these are some principles that any potential hard fork must not violate:
- The network must remain substantially decentralized.
- The inflation schedule must be the same or lower/slower. (Though I'm not 100% sure whether lowering inflation would be OK.)
- No one should be allowed by design to steal your money.
- As much as reasonably possible, no one should be able to prevent you from spending your money.
- Anonymity should be at least possible.

I will oppose any unsafe hard fork, even if it's proposed by Gavin. I and the sites I have some hand in are independent of the dev group, the Bitcoin Foundation, and other companies/organizations. I don't know whether Gavin's current proposal is safe, so the only thing I'm doing now is recommending caution.


I wasn't a fan of the whole idea of giving miners any special say on the issue. (Though it wasn't actually much of a vote, since miners could only confirm/reject P2SH.) Miners are basically employees of the network, and it should be the actions of users and businesses that influence what miners do, not the other way around. It would have been possible and better for users and businesses to (at a reasonable pace) force miners to accept the P2SH change.

code needs to be in place so I can run a full node and mine (as solo or pool operator) without the entire chain on my machine.

Why are you being so shy and throwing out insinuations? Just be honest and direct. Do you suspect I'm a paid shill from the Philippines ?

What?!?! This is the dumbest thing ever so

1: It's going to cause bloat to allow blocks larger in size?
2: It's going to cause centralization how? That the blocks are ABLE to get to 20mb if there are enough people out there using Bitcoin
3: You rather not fork and join a minority that would cause a huge hit to Bitcoin in the hopes of what a dual chain that would kill the ease?

Seriously it's stupid.

Don't fork fine, your just hurting the network as a whole if it goes through and you dig your head in the sand.

Well, it's either this, or 'intellectually challenged'.

Why are you being so shy and throwing out insinuations? Just be honest and direct. Do you suspect I'm a paid shill from the Philippines?

It seems someone is indeed astroturfing. Not as a point of Bitcoin promotion, but promoting a certain kind of future for Bitcoin. It is very likely the grammatically correct, technically word salad posters are hired.

I don't really understand your (nor Mircea's) point. Are you trying to say that some overt Bitcoin supporters are plants from PR companies doing positive astroturfing/Bitcoin brand amplification? Not just the standard negative trolling of style similar to the Something Awful?

This is a very important point.

Eh, all in Costa Rica and seemingly none of them have support for English-speaking customers. This isn't a representative sample of the worldwide Internet. And by a margin so wide that it makes your statement laughable.

From my experience trying to support the use of Bittorrent (both commercially and individually) and various RPC stacks (commercial support for home workers) I can say that the "complaints" don't equal the real ISP problems. Only somewhere between 1% and 5% are the real provider-caused blocking/interference/mishandling. The remaining 95% to 99% is on the customer-caused misconfiguration or hardware/software faults.

Many ISPs, used advanced filtering products to protect their network and clients like http://www.cisco.com/c/en/us/products/security/security-monitoring-analysis-response-system/index.html , so even if they aren't overtly blocking TCP on port 8333 but their firewalls and filters are because it is detecting a suspicious transfer.

You can test for 8333 being opened here - http://www.yougetsignal.com/tools/open-ports/ and if you don't have more than 8 connection in bitcoin QT you are not accepting any incoming connections on port 8333.

Sometimes its a local firewall blocking you or you need to either allow UPnP or port forward to get QT functioning as a full node but in many other cases its the ISP themselves.

Regardless of the exact specifics with each case, most users aren't acting as a full node and the numbers back my claim. The average user isn't going to call their ISP , or create a forwarding rule in their router and make sure their firewall is configured properly.

I can only talk about what I have personally tested:
Cabletica, ICE, Kolbi, Claro, Moviestar all need to be called and asked to open this port.
I have to call to open many ports blocked by default for BTC or other uses like CCTV.

I also see many other posts around the world making the same complaints.

I've noticed certain repeating characteristic in the writing of many members of this forum: they construct grammatically correct sentences but absolutely disregard the underlying semantics: incoming vs. outgoing, local vs. remote, source vs. destination, etc. Here in regards to TCP/IP ports, but I observed that in regards to pretty much any technical issue.

It reminds me of the http://en.wikipedia.org/wiki/Word_salad , but doesn't go as far in the unintelligibility. It is more akin to somebody just memorizing sentences and phrases without any sort of comprehension, something that actors have to do well.

What would be the real underlying psychological mechanism at work here? Conformism? Or maybe there is a physiological explanation, like some sort of milder http://en.wikipedia.org/wiki/Korsakoff%27s_syndrome ?

I'm really puzzled, because I've noticed this also in some very visible and high-level people, eg. core developers talking about the hardware design instead of the software design.

mircea_popescu: it was highlighted by intel about two weeks ago, and we're generally tracing it to philippines "top quality" content farms working for a number of (mostly ct and wash based) pr firms.
mircea_popescu: wasn't goping to say anything, but since it's public now..

Anyone can name an ISP that blocks outgoing TCP/IP connections to port 8333? And where such a block isn't a voluntary one (some sort "Internet safety" option) and the unblocking isn't a self-service "turn it off" checkbox?

How about incoming TCP/IP connections to port 8333?  And where such a block can be lifted and isn't a technical restriction stemming from the use of CG-NAT or DS-Lite?

My understanding was that the 1MB limit was only ever a temporary measure to limit spam when there wasn't much real transaction volume.  Pretty sad if this ends up holding Bitcoin back.  Guess altcoins will be the future then.


Well you'd think people could just learn to use Electrum or similar, rather than downloading the whole blockchain.

Why do we average less than 7k nodes worldwide at any given time? https://getaddr.bitnodes.io/

If and ISP blocked a particular port, one would get zero connections on it.

It would not surprise me at all to have a max cons set by and ISP.  That would explain the why one has difficulty.  Each connection uses network resources in a router (yours and your ISP's respectively.)

This is what drives me nuts about ignorant people making simplistic calculations of capability based on the download speeds (which are artifacts of a marketing department usually anyway.)  Ya, you might be able to get some decent percentage of a pipe in use but only by using multiple streams, and that is particularly the case with TCP on a wan.  Anyone who has tried to SCP a large file can tell you this.  Nobody is going to be thrilled about Joe Sixpack using a thousand sockets, and crappy consumer grade routers will have difficulty.  Probably max-cons is set by the ISP to throttle those trying to do torrents.  Same trick here.  This certainly illustrates the point I'm making about being at the mercy of network infrastructure providers.


IIRC it's even less than that by now.  Anyway, I don't have to guess about why this is.  It's obviously because for some bizarre reason Satoshi didn't choose to reward transfer nodes in his implementation.  Why I have no idea.  It's one of the strongest observations supporting the hypothesis that Bitcoin was designed to fail (or at least become centralized and under control.)  Not that I believe this to be true, but it is one of the many hypothesis that I continue to hold open.  Probably he just ran out of time and patience.  If he was ignornant enough to not forsee the difficulties with the block size hard-fork then he may also have imagined that transfer node rewards were something that could be tacked on later.

Save the end of the world prophesies for the speculation subforum please.

We could try a proof of troll system, keep randomly saying the same thing over and over and every time a new variation is found its a new block

Noticed the same with the ISP's, Ireland's really lax on internet rights and there's lots of weird stuff going on, outages followed by slowdowns on certain sites, everything loading fine except images taking ages, that kind of thing. Obvious what's going on, lots of sites seem to be going offline lately, nothing big but a lot crypto related.

Have to get back onto my ISP again now, hadn't realised they'd blocked 8333 again. Second time they've done that, will have to tunnel out to a VPS. That's what the dumbasses don't get, that which doesn't kill us makes us stronger.

We could try a proof of troll system, keep randomly saying the same thing over and over and every time a new variation is found its a new block. Save the end of the world prophesies for the speculation subforum please.

Noticed the same with the ISP's, Ireland's really lax on internet rights and there's lots of weird stuff going on, outages followed by slowdowns on certain sites, everything loading fine except images taking ages, that kind of thing. Obvious what's going on, lots of sites seem to be going offline lately, nothing big but a lot crypto related.

Have to get back onto my ISP again now, hadn't realised they'd blocked 8333 again. Second time they've done that, will have to route out to a VPS. That's what the dumbasses don't get, that which doesn't kill us makes us stronger.