Well, well, well, now we know what Jihan Wu’s been up to.

[Killerpotleaf]

Personally, my biggest concern about large block sizes is, to put simply: Datacenters instead of home users running full nodes.

Thats already a reality. Vast majority of nodes are currently run in datacenters.

The 300GB/mo bandwidth it takes to run a node is a lot for any home users connection. Even unmetered internet has a "fair usage policy", which is an undisclosed bandwidth limit that the ISP can change at any time. If your running a node at home and have multiple people streaming HD videos, you're going to get a call from your ISP about your "excessive bandwidth usage" regardless of what your limit is.

hi
welcome to my datacenter.

i think what you mean to say is if a user wishes to have a node with 99.9% uptime, which is constantly trying to connect to as many other nodes as possible, serving them the blockchain without any speed limit on individual connections, you'll hit 1T-300GB / month depending on the internet speed.

pretty sure if you we're nutty enough to connect to all ~6000 node and serve highspeed blockchain syncs at will, with a datacneter like speed connection, you'll probably manage to use 4TB of bandwidth / month

i forget what was your point?

[1715457118]

1715457118

[1715457118]

1715457118

[Killerpotleaf]

now lets take a look at what a maxed out bittorrent node can do.


such centralization!

[Quickseller]

they would not be pushing for the max block size to be lowered

luke JR wants 0.5mb blocks.. just to correct you a little

Luke JR has wanted smaller blocks before his involvement in blockstream. Also, the HF that Luke JR proposed is not a real proposal as all sides will be against it.

First item is to understand why Blockstream refused to accept 2MB blocks along with SegWit.

It is not that they feel that 1MB is the optimal number (if the max block size was currently 2MB, they would not be pushing for the max block size to be lowered to 1MB), and it is not that they are against 2MB blocks in itself, it is that they are against any max block size increase ever.

From the core roadmap https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2015-December/011865.html

Finally--at some point the capacity increases from the above may not
be enough.  Delivery on relay improvements, segwit fraud proofs, dynamic
block size controls, and other advances in technology will reduce the risk
and therefore controversy around moderate block size increase proposals
(such as 2/4/8 rescaled to respect segwit's increase). Bitcoin will
be able to move forward with these increases when improvements and
understanding render their risks widely acceptable relative to the
risks of not deploying them. In Bitcoin Core we should keep patches
ready to implement them as the need and the will arises, to keep the
basic software engineering from being the limiting factor.

They want improvement and security BEFORE raising the blocksize, nowhere does it say 1mb forever.

It is not unusual to see fairly long periods of nearly 1MB blocks in 2015 and technology has improved in the last two years. The max block size has been in place for ~7 years, and technology has drastically improved since then.

They can say that they are in favor of a block size increase in the future all they want, however their actions speak much louder than their words, and they clearly will never accept any kind of max block size increase. There is also no technical reason why we could not HF to 2MB blocks today and get a SegWit SF in 2-3 years...

[anonymoustroll420]

It is not unusual to see fairly long periods of nearly 1MB blocks in 2015 and technology has improved in the last two years. The max block size has been in place for ~7 years, and technology has drastically improved since then.

Internet connection speeds have not improved. In the US, the average internet connection speed has been dropping recently due to increased loads due to HD video streaming etc and not enough investment in new infrastructure to counter this.

A big cause of this was many years ago ISP's set up coaxial networks instead of fibre to try and save money, and it's just not economical to rip out the coaxial and replace it with fibre now. Another cause is that in many areas people only have the choice of one ISP. That ISP has a monopoly and has no incentive to invest in a better network as there is no competition. Government regulations prevent new ISP's from setting up in the area, due to the ISP's lobbying to get this regulation past.

[Quickseller]

The 300GB/mo bandwidth it takes to run a node is a lot for any home users connection.

I am not quite sure where you are getting this number.

The current max block size is 1 MB, and the difficulty adjusts so that on average there will be 144 blocks found every 24 hours. This is 144 MB per day if you are using the 'blocks only' feature.

144 MB per day times 30 days is 4.32 GB per month if all you download is blocks and do not accept incoming connections.

If you start to accept incoming connections and have 8 peers at any one time, then your bandwidth grows to 34.56 GB per month.

If you want to receive 0/unconfirmed transactions as they appear on the network, then you would roughly double required bandwidth to ~75 GB per month for each of upload and download -- you would want to obviously disable RBF, and would want to set your minrelayfee (or whatever it is called) high enough so that you do not receive transactions that are unlikely to confirm.

You would need some unknown additional amount to account for serving new nodes old blocks.

Even unmetered internet has a "fair usage policy", which is an undisclosed bandwidth limit that the ISP can change at any time. If your running a node at home and have multiple people streaming HD videos, you're going to get a call from your ISP about your "excessive bandwidth usage" regardless of what your limit is.

Looking at cox communications, it looks like that their 'fair use' threshold is at 1TB per month, and effectively only applies in a few cities. This is 3x above your claimed 300 GB per month requirement to run a full node, and ~13x of my above calculations.

I would also point out that it is unlikely that 3MB (or 13 MB) blocks will be full over the long run currently, so even if the max block size were increased to these amounts, you would likely not be in danger of hitting the 'fair use' limits.

It is not unusual to see fairly long periods of nearly 1MB blocks in 2015 and technology has improved in the last two years. The max block size has been in place for ~7 years, and technology has drastically improved since then.

Internet connection speeds have not improved. In the US, the average internet connection speed has been dropping recently due to increased loads due to HD video streaming etc and not enough investment in new infrastructure to counter this.

I am not sure about this. Over the past ~4-5 years, my internet speeds have been growing at a much higher rate than my internet bill. (I don't have specific numbers).

Longer term trends show that residential internet speeds are growing faster than what moores law would anticipate. Again, looking at cox pricing, it currently costs ~$100/month for 300 mbs of internet download speed, and 30 mbs upload speed.

If you counted the price of both the additional required phone line and the ISP service, it would cost ~$50 for a 28.8k connection about 20 years ago. Using round numbers, the unit cost of internet speeds (download) have decreased by a factor of ~5000x over the past 20 years.

[iamTom123]

Fucking over everyone using bitcoin is great... Potential user sees a group that's able to game the system with an advantage, and want's to run the other way. Was it illegal, who knows. Amoral, unethical, fucked up central banker shit absolutely.

 
Well I would agree that given the context of Bitcoin that everyone should be able to compete fairly, a patent should be unenforceable, but I wouldn't blame Jihan for trying to mine.  Everyone using ASICs is trying to gain an advantage.

Ridiculous imo to try to change the PoW... its the last thing Core should be doing.

Having an exploit and using it is one thing. Especially when that thing is a sizeable advantage. But an entirely different thing to manipulate the environment, in order to maintain that advantage. You should evolve, not.stagnate everyone else.  I hate this about DC politics, hate to see it in BTC as well

I also hate politics because in many cases politicking is just promoting one's self-vested interest and advantage and many times to the disadvantage of the other many stakeholders. This situation here is quite an insult to the ideas of Bitcoin...we are avoiding centralization and regulation yet we are facing people whose greed is showing for all the world to see. I hope these people will realize that if in the process they stagnate Bitcoin in the end we might not have a Bitcoin to speak up. Hope these guys can level up and not let their profit-oriented minds do the talking all the time. The whole Bitcoin community should condemn these people and have them sent to the digital prison called ConBits.

[Killerpotleaf]

this is highly speculative but seemingly flawless logical reasoning
I think you're attributing deliberate intent based on some grand master plan, which only exists in your head.  
I think the reality is more transparent and random coincidence then you make it out to be.
I also think that giving everyone asicboost for free means that no one will have the advantage, and so now one will stand to lose anything by going with bigger blocks.

[Quickseller]

First item is to understand why Blockstream refused to accept 2MB blocks along with SegWit.

It is not that they feel that 1MB is the optimal number (if the max block size was currently 2MB, they would not be pushing for the max block size to be lowered to 1MB), and it is not that they are against 2MB blocks in itself, it is that they are against any max block size increase ever.

From the core roadmap https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2015-December/011865.html

Finally--at some point the capacity increases from the above may not
be enough.  Delivery on relay improvements, segwit fraud proofs, dynamic
block size controls, and other advances in technology will reduce the risk
and therefore controversy around moderate block size increase proposals
(such as 2/4/8 rescaled to respect segwit's increase). Bitcoin will
be able to move forward with these increases when improvements and
understanding render their risks widely acceptable relative to the
risks of not deploying them. In Bitcoin Core we should keep patches
ready to implement them as the need and the will arises, to keep the
basic software engineering from being the limiting factor.

They want improvement and security BEFORE raising the blocksize, nowhere does it say 1mb forever.

But Blockstream's stated logic above makes no sense. The real reason Blockstream doesn't want 2MB blocks now can only be because of what I explained earlier (because there is no other plausible reason that makes any sense).

It makes no sense from both a technology standpoint (i.e. afaik SegWit adds no significant reliability and security that would alleviate issues that a 2MB block could cause, instead the improvements of SegWit are bugs fixes for things that have nothing to do with block size)  and also because it is quite clear that they know Bitmain can covertly block all block size increases, so they know that the 2MB offer is actually a clever ploy of Bitmain wherein Bitmain's own customers will end up blocking the block size increase (so Bitmain can deny that they did it) and also because Blockstream knows they become irrelevant once the market has SegWit and block size increases. The market won't be in any rush to accept further protocol changes from Blockstream. So Blockstream is fighting for their control against Bitmain. So then if you understand all this I have written, then you understand that Bitmain is really against block size increases, but they have such a clever strategy that they can make it appear that they are for larger block sizes. The covert aspect of asicboost is a very key component to the masterful chess that Bitmain is playing. Bitmain's goal is to keep the protocol immutable and to destroy Blockstream. Very simple. But they have a clever strategy will obscures their actual intentions. Blockstream's goal is to be in control of the changes to Bitcoin's protocol.

Read again Bitmain's statement with this new analysis in mind.

I enjoy figuring out these mental chess challenges. That is why I do it. Not because I am for or against anything. I am trying to help readers understand the facts.

First item is to understand why Blockstream refused to accept 2MB blocks along with SegWit.

I haven't actually looked to see their official stated reason, but I can guess that it must be because they know they each time they want to "upgrade" the protocol, they need have some carrot and stick that forces the market to want their upgrade. So keeping blocksize very constrained is necessary so that Blockstream remains in control of the protocol.

In other words, Blockstream can't accept 2MB blocks, because they know their business model depends on them being in control of the protocol of Bitcoin. Also factor in that maybe Bitmain has told them behind the curtain that Bitmain will block all their future protocol changes after the 2MB+SegWit. Blockstream probably couldn't repeat that in public, because Bitmain would deny it and accuse Blockstream of fabricating lies.

Am I correct so far?

First item is to understand why Blockstream refused to accept 2MB blocks along with SegWit.

I haven't actually looked to see their official stated reason, but I can guess that it must be because they know they each time they want to "upgrade" the protocol, they need have some carrot and stick that forces the market to want their upgrade. So keeping blocksize very constrained is necessary so that Blockstream remains in control of the protocol.

In other words, Blockstream can't accept 2MB blocks, because they know their business model depends on them being in control of the protocol of Bitcoin. Also factor in that maybe Bitmain has told them behind the curtain that Bitmain will block all their future protocol changes after the 2MB+SegWit.

Am I correct so far?

It could be a matter of opinion. Or maybe you have perceived one actors tactical methods used in pursuit of their strategic goals.

If you had the chance to get your most significant work approved (SegWit) and only had to accept a 2MB block, why wouldn't you do it?

It makes absolutely no sense. Can you think of any other rational explanation than the one I offered?

Even if Blockstream had an opinion that they don't like 2MB blocks, they wouldn't block their own work over such a small quibble. There has to be a more significant reason they refuse 2MB blocks with SegWit. I offered my reason. Is there any possible other reason that makes any sense?

(note I am so very sleepy, so we may have to continue this after I sleep)

Presuming I am correct that Bitmain knows that Blockstream will never accept a 2MB+SegWit proposal, then you can then re-analyze their statement knowing full well that he knows there will never be big blocks.

Furthermore Bitmain is challenging the community to make AsicBoost patent licensing accessible to everyone. And to enable AsicBoost in the most optimum way. But Bitmain already stated that larger blocks make AsicBoost less efficient.

And remember some "hacker" (which is actually Bitmain but nobody can prove it) can release the covert upgrade so all Bitmain's [existing customers who own] miners on the planet could increase their efficiency and side-step the patent (and also they would be against large blocks).

So Bitmain is total BS about wanting larger blocks. Bitmain is trying to destroy Blockstream by never allowing them to get any protocol changes at all. That is why they proposed the 2MB+SegWit at HongKong. And that is why @Gmaxwell is wanting to block the covert AsicBoost.

But Bitmain has checkmated Blockstream. Think it out. I am too sleepy to explain all the detailed reasoning.

So I hope you understand that I don't think Bitmain nor its customers are currently employing asicboost in mining. Rather Bitmain has an installed base of customers worldwide (where the patent is a liability for the customers) whose hardware could suddenly be made more efficient if Bitmain were to release the software/firmware upgrade to enable asicboost.

Thus Bitmain has a poison pill on any block size increases which won't pin the blame on them for the failure of any soft (or hard) fork which attempts a block size increase. Because a "hacker" (who is actually Bitmain) could release the covert asicboost upgrade over the Internet, and then all those Bitmain customers would be economically incentivized to employ it and to block both block size increases and @gmaxwell's overt BIP (which is supposed to block the covert form of asic boost but not the overt form). Or even some clever people might reverse engineer and design the upgrade from scratch and release it. The point is the customers of Bitmain wouldn't want to run the overt asic boost because they wouldn't want patent liability. And those customers would also not want block size increases because asic boost looses efficiency with larger block sizes.

So Bitmain can block any block size increases without it even being provable that they did obstruct it. It can be made to appear that Bitmain was the good guy and was supporting 2MB blocks, but in fact behind the scenes Bitmain has already told Blockstream the truth and Blockstream is realizing they are fucked and trying to find a way to fight back, but it is too late, they are already checkmated by Bitmain's very clever strategy.

Bitmain urges the community to make the patent licensing available to all (not their patent but the patents held by others on asicboost) so it appears Bitmain is the good guys. Yet in reality, Bitmain knows that the patent licensing won't be forthcoming because you can't organize such a thing given there are multiple patent holders in different jurisdictions. Thus Bitmain knows the covert upgrade is the poison pill that blocks any thing Blockstream would attempt to do. Blockstream is now frantic and desperate.

Given the Chinese now know they have us by the balls, they are always manipulating the LTC markets as well. Using all this to pump up and crash the LTC price at will and taking more BTC from all of us.

I am hating PoW even more. I want to make those ASICs irrelevant. I want to eliminate PoW. I am so tired of all this fighting and manipulation. Before I supported Bitcoin because I thought it was a stable and politics-free reserve currency of the altcoins which I could hold without losing my time on this shit. Now these fuckers have turned the Bitcoin is a fucking noise that makes it just not worth it.

I will kill this shit. I hope you guys understand what is really going on.

Why would Bitmain not want a larger max block size? Larger blocks will ultimately lead to additional total tx fees (and in turn higher overall block rewards), both of which will lead to greater profitability of those who run their equipment, which would theoretically lead to higher market prices for their equipment, and most importantly, more profits for them.

[anonymoustroll420]

I am not quite sure where you are getting this number.

https://bitcoin.org/en/full-node#minimum-requirements

Plus my own experience with running a node.

Yes, you can limit this a lot by setting everything to minimum, but then you're not really running a full node. And I believe it will still be above your estimates.

144 MB per day times 30 days is 4.32 GB per month if all you download is blocks and do not accept incoming connections.  

Bandwidth from newly announced blocks is not the only consumption of bandwidth resources. You need to account for:

Transaction relay
New nodes syncing old blocks
Upload and download bandwidth is not uniform.
The impact of SPV wallets
Frequency and magnitude of worst-case-scenarios and attacks
The impact of orphan blocks and reorgs.
The impact of stale blocks.
Blocks may arrive before the last one is received or processed.
Bitcoin does not use bandwidth in a uniform manner, tends to be in spikes.
Bandwidth availability can vary depending on general ISP congestion.
Transaction backlog and mempool requirements.
The processing load of transactions is not identical for every transaction.
The impact of block propagation techniques such as invertible bloom lookup tables.
The impact of other bitcoin features, such as replace-by-fee.
The number of peers connected to the node.
The acceptable latency of blocks when 'catching up' in a series of good luck from miners.
The effect of pruning nodes vs full nodes.
Latency of multiple global hops may become nontrivial.
Other consumers of bandwidth

If you counted the price of both the additional required phone line and the ISP service, it would cost ~$50 for a 28.8k connection about 20 years ago. Using round numbers, the unit cost of internet speeds (download) have decreased by a factor of ~5000x over the past 20 years.

Try the last couple of years.
https://www.theverge.com/2012/5/1/2990469/average-global-internet-speed-drop-us

[iamnotback]

I also think that giving every asicboost for free means that no one will have the advantage, and so now one will stand to lose anything by going with bigger blocks.

Not everybody has Bitmain's hardware. Not everybody is willing to risk the liability of running a (potentially) patented thing covertly. It is just simple economics. Miners are forced to support what gives them an economic advantage.

[franky1]

It is not unusual to see fairly long periods of nearly 1MB blocks in 2015 and technology has improved in the last two years. The max block size has been in place for ~7 years, and technology has drastically improved since then.

Internet connection speeds have not improved. In the US, the average internet connection speed has been dropping recently due to increased loads due to HD video streaming etc and not enough investment in new infrastructure to counter this.

A big cause of this was many years ago ISP's set up coaxial networks instead of fibre to try and save money, and it's just not economical to rip out the coaxial and replace it with fibre now. Another cause is that in many areas people only have the choice of one ISP. That ISP has a monopoly and has no incentive to invest in a better network as there is no competition. Government regulations prevent new ISP's from setting up in the area.

lol

internet speeds are dropping due HD video streaming?
lol
HD video streaming was not really a thing in 2009 but now its every day used.
meaning internet streaming is now more then possible due to improvements.
EG 56k-> ISDN-> ASDL-> Fibre..
even mobile/cellular has moved on. now we have 5G being in the telecommunications 5 year plan.

im laughing.
whats next tell twitch/youtube and google hangouts to cut their losses because the internet is dead and they wont get more than 7000 users able to cope with livestreaming.

i know people in china, russia, africa, australia urkraine, canada india and thailand that all are NOT crying about the internet.

oh wait let me guess the usual reddit script rebuttle "gigabytes by midnight"
response
dont be a fool, dynamic implementations take a natural growth scaling based on what nodes can cope with.
even blockstream have accepted 8mb is safe but went for 4mb to be ultra safe. (even if that limit wont be utilised due to flaws)

p.s

a 500kbit/s = 37megabyte/10 minutes.
this allow 4mb to be sent 8 times in 10 minutes
a 1mbit/s = 74megabyte/10 minutes.
this allow 4mb to be sent 8 times in 5 minutes
a 2.5mbit/s = 187megabyte/10 minutes.
this allow 4mb to be sent 8 times in 2 minutes

average upload speeds are about 3mbit/s on both landline and 4g cellular.
and thats before even talking about fibre/5g

so get it out of your head that 2mb is bad. thats a very old reddit script


P.S

Try the last couple of years.
https://www.theverge.com/2012/5/1/2990469/average-global-internet-speed-drop-us

lol stats of 2011....... dang you really are grabbing the old old OLD reddit scripts

[anonymoustroll420]

i know people in china, russia, africa, australia urkraine, canada india and thailand that all are NOT crying about the internet.

https://www.theverge.com/2012/5/1/2990469/average-global-internet-speed-drop-us

Whats actually funny is that the internet speeds in developing countries are actually increasing, because they are just now investing in their internet infrastructure, while 1st world countries are stuck with old outdated networks.

it currently costs ~$100/month for 300 mbs of internet download speed, and 30 mbs upload speed.

Just a note, thats not MB, megabytes, thats Mb, megabits. A common misconception people have. There are 8 bits in a byte. So its 8x less in MB, in reality you can expect it to be around 10x less, as thats the 'maximum' which the ISP will never go over.

[Killerpotleaf]

I also think that giving every asicboost for free means that no one will have the advantage, and so now one will stand to lose anything by going with bigger blocks.

Not everybody has Bitmain's hardware. Not everybody is willing to risk the liability of running a (potentially) patented thing covertly. It is just simple economics. Miners are forced to support what gives them an economic advantage.

but Bitmain said he would want to work with other patent holders to null and void the patents and make it available to everyone.

his reasoning was, its better for bitcoin's security, to have an extra 20% hashpower over all, then drop hashing power by disabling his (potential) advantage.

you might  need to do a little correction to your conspiracy theory.

remember the episode of star trek voyager where 7 of 9 assimilate to much data and comes up with crazy conspiracy theories

Star Trek: Voyager
The Voyager Conspiracy
Season 6 | Episode 9

you should go watch this now, and then GET SOME SLEEP!   i should do that to...

[iamnotback]

Why would Bitmain not want a larger max block size? Larger blocks will ultimately lead to additional total tx fees (and in turn higher overall block rewards), both of which will lead to greater profitability of those who run their equipment, which would theoretically lead to higher market prices for their equipment, and most importantly, more profits for them.

1. Because killing Blockstream and removing their ability to destroy Bitmain (and Bitcoin) by controlling the protocol is more important.

2. Because SegWit can be added on Litecoin where they can still earn all those fees, and Blockstream will be impotent.

It all about removing the coup d'etat of Blockstream.

In that way I like Bitmain, but as I said the Chinese are fucking with us and making it a fucking mess. I really blame PoW, because PoW makes all of this shit possible.

Can there be a better way? Maybe not. Maybe yes. I have ideas I need to go work on.

So are there are any objections to my analysis of this fucking mess?

[anonymoustroll420]

so get it out of your head that 2mb is bad. thats a very old reddit script

I don't think 2MB is bad. 2MB is safe enough I reckon. I think that Bitcoin 'unlimited' is bad, or actually limited to 72MB because thats the maximum the code currently supports.

From what I've read, the situation with internet in the US is getting better as a whole except for some areas, however there are mnay countries where the situation is getting worse

https://www.rte.ie/news/2016/0706/800618-internet-speed-ireland/

My main point is, internet speeds do not magically get better with better technology. End users do not always get access to that better technology.

[iamnotback]

I also think that giving every asicboost for free means that no one will have the advantage, and so now one will stand to lose anything by going with bigger blocks.

Not everybody has Bitmain's hardware. Not everybody is willing to risk the liability of running a (potentially) patented thing covertly. It is just simple economics. Miners are forced to support what gives them an economic advantage.

but Bitmain said he would want to work with other patent holders to null and void the patents and make it available to everyone.

his reasoning was, its better for bitcoin's security, to have an extra 20% hashpower over all, then drop hashing power by disabling his (potential) advantage.

I already refuted that:

Bitmain urges the community to make the patent licensing available to all (not their patent but the patents held by others on asicboost) so it appears Bitmain is the good guys. Yet in reality, Bitmain knows that the patent licensing won't be forthcoming because you can't organize such a thing given there are multiple patent holders in different jurisdictions. Thus Bitmain knows the covert upgrade is the poison pill that blocks any thing Blockstream would attempt to do. Blockstream is now frantic and desperate.

Can anyone find a flaw in my analysis?

[franky1]

i know people in china, russia, africa, australia urkraine, canada india and thailand that all are NOT crying about the internet.

https://www.theverge.com/2012/5/1/2990469/average-global-internet-speed-drop-us

Whats actually funny is that the internet speeds in developing countries are actually increasing, because they are just now investing in their internet infrastructure, while 1st world countries are stuck with old outdated networks.

it currently costs ~$100/month for 300 mbs of internet download speed, and 30 mbs upload speed.

Just a note, thats not MB, megabytes, thats Mb, megabits. A common misconception people have. There are 8 bits in a byte. So its 8x less in MB, in reality you can expect it to be around 10x less, as thats the 'maximum' which the ISP will never go over.

30mbit/s=2.2gBYTE/10minutes
more then you need, what ar you complaining about

[franky1]

so get it out of your head that 2mb is bad. thats a very old reddit script

I don't think 2MB is bad. 2MB is safe enough I reckon. I think that Bitcoin 'unlimited' is bad, or actually limited to 72MB because thats the maximum the code currently supports.

you dont even know how dynamics works.
hint
its wont be 72mb over night.
seriously, stop reading reddit, its filling you with FUD and out dated data..
yep still laughing at you pulling out the 2011 stats as your rebuttle

[anonymoustroll420]

you dont even know how dynamics works.
hint
its wont be 72mb over night.
seriously, stop reading reddit, its filling you with FUD and out dated data..
yep still laughing at you pulling out the 2011 stats as your rebuttle

Miners essentially decide the limit. No I don't think it will go to 72mb overnight, however, the block size needs to be kept in check, otherwise the network will become centralized. The best way to do that is any blocksize increase is scheduled and agreed upon by economic majority.

Even though the stat is from 2011, it proves that internet speeds don't always magically go up. They can drop 14% in the course of 3 months, as shown.

My internet connection has gone to complete shit over the past 5 years. From 250mb to 50mb, when it works. I do not have the choice of a different ISP.

[Quickseller]

I am not quite sure where you are getting this number.

https://bitcoin.org/en/full-node#minimum-requirements

Plus my own experience with running a node.

Yes, you can limit this a lot by setting everything to minimum, but then you're not really running a full node.

144 MB per day times 30 days is 4.32 GB per month if all you download is blocks and do not accept incoming connections. 

If you start to accept incoming connections and have 8 peers at any one time, then your bandwidth grows to 34.56 GB per month.

Bandwidth from newly announced blocks is not the only consumption of bandwidth resources. You need to account for:

Transaction relay Accounted for in my 8 peer, receive unconfirmed transactions calculation
New nodes syncing old blocks True, I am not sure how to best measure this
Upload and download bandwidth is not uniform. True, however my 8 peer, receive unconfirmed transactions calculation would be upload, which would be the greater consumer between upload/download
The impact of SPV wallets Each SPV wallet should count towards my 8 connections
Frequency and magnitude of worst-case-scenarios and attacks True, however I don't think these happen frequently, especially not every month
The impact of orphan blocks and reorgs. This should be minimal, maybe 1% increase at best, especially considering that SPV mining reduces orphans/reorgs
The impact of stale blocks. see my response to orphans/reorgs
Blocks may arrive before the last one is received or processed. see my response to orphans/reorgs
Bitcoin does not use bandwidth in a uniform manner, tends to be in spikes. The spikes are due to downloading the entire block once one is found/relayed. This is accounted for in my calculations
Bandwidth availability can vary depending on general ISP congestion. I am not sure how this affects total download/upload amounts when running a node
Transaction backlog and mempool requirements. the minrelayfee should be set to an amount so that you will not relay transactions that are unlikely to get confirmed semi-quickly(maybe 72 blocks) -- I don't see the benefit to relaying these types of transactions for most home users
The processing load of transactions is not identical for every transaction. each transaction that is confirmed still counts against the 1MB per 10 minutes limit. Also this is more a CPU issue, not a bandwidth issue
The impact of block propagation techniques such as invertible bloom lookup tables.
The impact of other bitcoin features, such as replace-by-fee. This is not a feature I support, and as such any node I run will not relay RBF transactions -- I acknowledge others may disagree
The number of peers connected to the node. I understand that it is fairly common for home users to limit their connections to 8
The acceptable latency of blocks when 'catching up' in a series of good luck from miners. This is more of an internet speed issue, not a total bandwidth per month issue as the long term average block time should be about 10 minutes
The effect of pruning nodes vs full nodes. Yes, running a full node might result in more new nodes asking your node for old blocks if there are a lot of pruning nodes
Latency of multiple global hops may become nontrivial. Latency is likely to improve over time, not deteriorate, with the exception of the possibility of a fairly major war breaking out
Other consumers of bandwidth

See my responses in red above.

If you counted the price of both the additional required phone line and the ISP service, it would cost ~$50 for a 28.8k connection about 20 years ago. Using round numbers, the unit cost of internet speeds (download) have decreased by a factor of ~5000x over the past 20 years.

Try the last 1-2 years.

It doesn't look like there is very much data available for this measurement. It looks like unit costs of internet declined by ~1/3 in 2015 and 40% in 2014 though. I would be surprised to see those levels of price declines to entirely reverse in the span of two years.