Pages:
Author

Topic: What are the downsides to 8MB blocks? - page 4. (Read 5376 times)

sr. member
Activity: 434
Merit: 250
August 29, 2015, 10:29:13 AM
#67
Better call Netflix to stop their 4K releases, otherwise nobody will be able to watch Netflix anymore if everything is 4K.

Let's hope Windows 11 can fit back on a floppy, otherwise we are doomed.
newbie
Activity: 42
Merit: 0
August 29, 2015, 08:23:20 AM
#66
Why would you need 2 TB SSD to store 512 MB of data? (we are talking about running a node here, right?) Even if you wanted to store the full blockchain 2 TB HDDs are cheap and would be enough for long time to the future.

How do you maintain all blockchain when block size limit will be 8 MB? Blockchain is almost 50 GB already. It'll be 63 GB before new year. Imagine we see 8 MB blocks every 10 minutes. Even 2 TB would be obsolete in 5 years.
If you don't want any problems you don't store your blockchain in HDD. You need a good SSD. Running Bitcoin Core is very expensive even right now.

I'm actually using SSD myself too, so I don't know if there would be problems using HDD. What would they be?

8MB block limit doesn't mean that all blocks will be full. But let's say they were. 5 years is still a very long time and by then larger drives would be cheaper.

Also why should everybody store the full blockchain? Blockchain size doesn't really matter (aside from sync time) if most nodes use a pruned chain that only contains the latest history.
Nas
full member
Activity: 167
Merit: 100
August 29, 2015, 07:54:53 AM
#65
Why would you need 2 TB SSD to store 512 MB of data? (we are talking about running a node here, right?) Even if you wanted to store the full blockchain 2 TB HDDs are cheap and would be enough for long time to the future.

How do you maintain all blockchain when block size limit will be 8 MB? Blockchain is almost 50 GB already. It'll be 63 GB before new year. Imagine we see 8 MB blocks every 10 minutes. Even 2 TB would be obsolete in 5 years.
If you don't want any problems you don't store your blockchain in HDD. You need a good SSD. Running Bitcoin Core is very expensive even right now.
newbie
Activity: 42
Merit: 0
August 29, 2015, 07:44:49 AM
#64
Some idiots are still running nodes on Raspberry Pis or Beagle Bones.
Exactly what is so idiotic about supporting the Bitcoin network by running dedicated devices as nodes (like Raspberry Pis) ?

Yeah, why don't you buy 2 TB SSD to everybody in Bitcoin community Mr. Donator Legendary rich fella? Almost everybody in this community wants to run Bitcoin Core as node.
Or you want to limit node usage to some secret elite society? More centralization is definitely a good idea, huh?

Why would you need 2 TB SSD to store 512 MB of data? (we are talking about running a node here, right?) Even if you wanted to store the full blockchain 2 TB HDDs are cheap and would be enough for long time to the future.
sr. member
Activity: 473
Merit: 250
August 29, 2015, 07:42:39 AM
#63
A good argument against the bitcoiners that demand 1MB blocksize because "Miners have to be rewarded" is the electricity that goes away with that. I recently read an article about a study saying that currently 1! bitcoin transaction costs so much electricity like 1.75 normal US households use in a day.

We can't let all these miners survive artificially by raising the fees. The amount of miners need to be trimmed down and the bitcoin security be more power efficient. That was always happening and it should and can't be held up artificially. The bitcoin network would be still very very secure with less miners.

Unfortunately miners are the mighty ones in the game. We only can hope that the sane one are the biggest ones. And that they see the need of bitcoin adoption for getting future profits.
newbie
Activity: 42
Merit: 0
August 29, 2015, 07:38:22 AM
#62

i have had trouble trying to get a node up and running as it is.

What kind of problems?

I'd guess bandwidth saturation while syncing? You only need to sync once and after the blockchain is downloaded the bandwidth and processor usage is minimal. If you're willing to run a node (and thus keep it up for months or years) few days to sync the blockchain isn't an issue. (IMO)
Nas
full member
Activity: 167
Merit: 100
August 29, 2015, 07:37:20 AM
#61
Some idiots are still running nodes on Raspberry Pis or Beagle Bones.
Exactly what is so idiotic about supporting the Bitcoin network by running dedicated devices as nodes (like Raspberry Pis) ?

Yeah, why don't you buy 2 TB SSD to everybody in Bitcoin community Mr. Donator Legendary rich fella? Almost everybody in this community wants to run Bitcoin Core as node.
Or you want to limit node usage to some secret elite society? More centralization is definitely a good idea, huh?
sr. member
Activity: 473
Merit: 250
August 29, 2015, 07:35:03 AM
#60
There are some problems with bigger blocks, as far as I can understand bigger blocks means:

1. You need to have more HDD place to store bigger blockchain.
2. Transactions fees will be actually higher, RIP bitcoin microtransactions.
3. Higher blocks also requires higher internet bandwidth.

Point 1 is not really the case. I mean the 8MB won't be filled magically suddenly. Where should these transactions come from. And when you would stay at 1MB only because of that then bitcoin would be unuseable since ALWAYS transactions would not be included. What kind of transaction system would it be when you can't trust that it transfers?

Point 2 is wrong. The fee will grow without a bigger blocksize. With 1MB blocksize for example everyone would pay higher fees to get their transaction included before others.

Point 3 is the same like point 1.
newbie
Activity: 42
Merit: 0
August 29, 2015, 07:29:44 AM
#59
Some idiots are still running nodes on Raspberry Pis or Beagle Bones.
Exactly what is so idiotic about supporting the Bitcoin network by running dedicated devices as nodes (like Raspberry Pis) ?


they will soon run out of disk space and I don't think raspeberries are supposed to handle so much disk accesses, so it might reduce the life of the device

Just use some external drive and prune the blockchain to size of your liking. I haven't had any problems running node on raspberry pi and won't have until bandwidth of my ca. $25 connection is not enough (would need over 100X increase of blocksize from today). Currently upload bandwitdh used is under 5kB/s for most of time (of which some might be other traffic as I also run other services on it).
newbie
Activity: 42
Merit: 0
August 29, 2015, 07:18:21 AM
#58
There are some problems with bigger blocks, as far as I can understand bigger blocks means:

1. You need to have more HDD place to store bigger blockchain.
Partly true. If you're running a node you can prune the blockchain. I run with a node with 10GB blockchain currently, but even pruning it to 512MB is possible.

Quote
2. Transactions fees will be actually higher, RIP bitcoin microtransactions.
False. Transaction fees will be lower, because there's more space in the block, thus less fee competition.

Quote
3. Higher blocks also requires higher internet bandwidth.
True. Now running a node requires some 5kB/s, so running a node won't be a problem for a long time. Propagation times for miners might be, though.
sr. member
Activity: 473
Merit: 250
August 29, 2015, 06:57:45 AM
#57
It will solve the spam attacks problem.
How is that? What prevents anyone from sending 8 times (or 100 times) as much spam as what we've seen recently?

What we need is a slight change in the default fee setting for Bitcoin nodes, that still allow very cheap payments (also micropayments) but helps against mass spam / dust transactions. Something similar to what Litecoin does, by requiring an additional small fee for each small output in a tx (and perhaps for each small input as well, while we're at it).

Will you do it? You would be poor pretty fast. Even the small KB that were needed to fell the 1MB were already expensive. Imagining that someone would spend the money to fill 8 MB blocks is hard.

It simply would not make any sense in any way to spam that way.
sr. member
Activity: 473
Merit: 250
August 29, 2015, 06:55:50 AM
#56
The biggest downside is basically the fact that the blocksize will be bigger, therefore limiting the amount of people that can access to it, or they could still access to it but it will be slower to download it.

Not really. Where should all these additional transactions come from suddenly to fill the 8MB? The blocks weren't full up to the spam attacks so why should suddenly there be so much more transactions?

If you think about spammers then no, that would simply be way too expensive.
sr. member
Activity: 473
Merit: 250
August 29, 2015, 06:53:46 AM
#55
8MB per 10 minutes is about 13.6 KB/s. What was the problem again?

Not that I think that simply increasing the block size limit to 8MB is a good solution (I'd rather see something more dynamic) but if 13.6 KB/s is actually a problem, maybe you should fix your internet, rather than debate over Bitcoin.

Miners cant wait 10 minutes to have their newly mined block trasfered to the next node.

What do you mean with that? Miners can propagate really fast. An there is a project, i'm not sure anymore what it's name was, that was created to propagate blocks faster. And they get >50% in 2 seconds or so when using that network. My numbers might be wrong but it was so fast that creating 1 transaction blocks made no real sense.
sr. member
Activity: 434
Merit: 250
August 15, 2015, 07:46:13 PM
#54
Some idiots are still running nodes on Raspberry Pis or Beagle Bones.
Exactly what is so idiotic about supporting the Bitcoin network by running dedicated devices as nodes (like Raspberry Pis) ?


they will soon run out of disk space and I don't think raspeberries are supposed to handle so much disk accesses, so it might reduce the life of the device

A 1TB drive attached to a Raspberry Pie doesn't run out of diskspace sooner then if it were attached to a Core i7.
I'm also not sure why it would die from "disk accesses".
legendary
Activity: 1372
Merit: 1252
August 15, 2015, 10:34:01 AM
#53
Sooner or later only a few people will be able to run nodes. And even if running nodes was easy, most people would be too lazy and resort to SPV or web based wallets. So the moral of the story is: No matter what they do, the % of people running nodes is always going to be a minority of enthusiasts.
legendary
Activity: 1120
Merit: 1000
August 15, 2015, 09:22:11 AM
#52
Some idiots are still running nodes on Raspberry Pis or Beagle Bones.
Exactly what is so idiotic about supporting the Bitcoin network by running dedicated devices as nodes (like Raspberry Pis) ?


they will soon run out of disk space and I don't think raspeberries are supposed to handle so much disk accesses, so it might reduce the life of the device
legendary
Activity: 1176
Merit: 1011
August 15, 2015, 09:15:58 AM
#51
Some idiots are still running nodes on Raspberry Pis or Beagle Bones.
Exactly what is so idiotic about supporting the Bitcoin network by running dedicated devices as nodes (like Raspberry Pis) ?
member
Activity: 70
Merit: 10
August 15, 2015, 08:23:55 AM
#50
This will help to solve the spam attack, and send more transactions to the network
sr. member
Activity: 434
Merit: 250
August 15, 2015, 06:45:48 AM
#49
What's the point of a "speedy 1MB blocks flashing by over 10gbps uplinks" bitcoin network, if nobody can use it.
legendary
Activity: 1148
Merit: 1014
In Satoshi I Trust
August 15, 2015, 03:56:59 AM
#48
no bandwidth problem. no storage problem.

i would like to see mass adoption. i would like to follow satoshis view of big blocks. i would like to have a permission less and cheap blockchain (cheap transactions) for everybody in the world.


look at Nielsen's Law:

http://www.nngroup.com/articles/law-of-bandwidth/

we dont talk about 100 MB blocks tomorrow - we talk about 6, 8 or 16 MB blocks in a few years (and it takes time to even fill them).
hundreds of millions of people can support such nodes today. no problem.

Nielsen's (Folk) Law is bounded by Gibson's Razor:

"The future is already here - it's just not evenly distributed."

See my lengthy, well-researched response to you here for Important Details.   Smiley

i read it. but even when you look Rustys estimation, bigger blocks are no problem:

Extracting the figures gives:

#Average download speed in November 2008 was 3.6Mbit
#Average download speed in November 2014 was 22.8Mbit

#Average upload speed in November 2008 to April 2009 was 0.43Mbit/s
#Average upload speed in November 2014 was 2.9Mbit



http://rusty.ozlabs.org/?p=551


@Soros Shorts

that could happen. but please note that blocks will probably not be full immediately.
Pages:
Jump to: