Thoughts on day 1 of the Milan Scaling Bitcoin conference."I just got back from day one of the conference. Everybody else went to a bar to drink and eat, but since I'm still jet lagged I came back to my hotel to get some rest.
I was also at the "Satoshi's Vision" conference about 2 weeks ago, and the two conferences are very different in many ways.
First off, the vibe here seems much more social. It seems that most people that are here are well connected socially. There are hardly any "loners" here. Contrast this to the Satoshi's Vision conference where most people there had never met anyone else there before.
During the presentations, a lot of people seem to be not paying very much attention. A lot of people are on their computers and not really listening to the speakers. I also hear a lot of chatting amongst the attendees. At the Satoshi's Vision conference, I got the feeling most people were there to listen to the speakers, and social fraternizing was not on the top of anyone's reasons for being there.
During the breaks, everyone would go outside to the snack tables and chat with each other. There is a lot of chatter. Everyone is talking with everyone else, and there is a high level of noise. When I say "noise" I don't mean people are being loud, the noise comes from so many people talking. After the break is over, people file back into the conference room, and the speaker has to quiet down the crowd before the speakers can begin. "Ladies and Gentlemen, can I have your attention.... We would like to get started....Can I have your attention..... Please everyone get seated.... Please take your seats we are about to get started". I don't mean to be demeaning, but it kind of reminds me of school children coming inside after recess. It takes a while for people to calm down and get ready to listen, everyone's conversations continue until the MC can get everyone to shut up. Maybe this is a function of the Scaling Bitcoin convention being much larger (almost 10x as many people are here as there were at Satoshi's Vision).
The content of the presentations seem to be loosely scaling related. Maybe Day 2 will have more scaling content, but today hardly any scaling content. For some reason, there was a lot of talks on fungibility and there was even a presentation on Timestamping. What does timestamping have to do with scaling? It seems like these Scaling Bitcoin" conferences are turning into general bitcoin presentations, not strictly scaling related. This is a stark contrast to the first Scaling Bitcoin conference which was pretty much completely blocksize limit proposals (BIP100, BIP101, BIP102, etc).
Also, it seems that there is very little new information coming from the presentations. The LN talks sounds like every previous talk on the LN. The Sidechain presentation sounded like it could have been the exact same presentation given 3 years ago. Very little progress seems to have been made in these areas. At least at Hong Kong segwit was announced, which added an air of excitement. This time around there isn;t hardly any excitement for any breakthrough.
Overall, the day was pretty boring. I'm not particularly interested in making friends here, as I'm not interested in BTC for social reasons. I'm only here to feel the vibe and get a more close up look at everything plays out." - freework
Layer two is not scaling Bitcoin."It's scaling layer two." - DIGITAL-not-Virtual
BitFury slide : " What the real topology of LN would be ???"There is still no sensible topology" - realistbtc
The only topology for which the routing problem is sovable is a single hub (or a small closely cooperating set of hubs, that work as one), and every LN user has a payment channel to/from that hub. Then the hub knows at all times the clearance of every channel (funding minus net payments made). The routing then is trivial, all paths have two hops, fees are minimized, and payment success is almost certain because the hub is always up. But that of course takes away all the advantages of bitcoin."Any amount of decentralization seems to make the problem much harder. A topology with a few loosely cooperating hubs is already hard enough, because each hub A needs to know the current clearance of every channel between every hub B and every client C. That does not seem possible, because only B and C have that information, and they may take too long to send it to A. Then paths found by the routing algorithm are not guaranteeed to have the necessary clearance, and the payment may have to be retried with a different path. Also the paths may have two or more hubs, which will all want to be paid a fee. The more distributed the topology, the worse these problems become. Also, if a client has to open more than one channel, it has to split his bitcoins between them beforehand; and that will very likely limit his payment capability. That is, if he has 3 channels with 1 BTC each, he cannot make a 2 BTC payment." - jstolfi
hardly matters,"it will never work" - paulh691
regarding the current situation:"Blockstream wants to pocket all those fees them self, instead of having them paid to the miners. So they "invented" all kinds of obscure technologies (like side chains or LN) to reroute those fees to them self. The unimaginable greed of Greg/Adam/Blockstream is capable to destroy Bitcoin." - coin-master
It is a major threat"indeed" - Ant-n
Right now, bitcoin uses less than 1% of my CPU time and bandwidth allowance, and less than $0.10/month of hard drive space. That's on a 2009 vintage desktop. We can easily scale by 100x and still run on typical home computers."100x more transactions means something like 100x more users. With that many users, they can share the cost of a heavy-duty server. For security of the network, you want there to be many independent nodes. With that many users, someone will set up a pre-packaged server, with the blockchain already loaded, that any data center can host. If 20 or 50 people access "their" server through a user account, to receive and push their own transactions, I don't see where the problem is. Also, with that many users, many companies, credit unions, etc. will host a node for their customers, so the network will stay distributed." - danielravennest
The point for me is to use bitcoin for financial transactions."Running a node myself isn't a burden, so I do it to have a high level of security. When the cost of running a node myself gets to be higher than the cost of a checking account ($12/mo typically), bitcoin loses one of its supposed advantages - that of low fees. I would then look for less expensive solutions. Sharing the cost of a node among a group of local people would do it. Given the assumption of 100 times more transactions and therefore lots more users, I could find local people to share the cost." - danielravennest
THe point is that as a whole we can create a type of P2P money that is censroship resistant (no one person CEO or gov can take ,freeze or seize your funds)."I think this was outlines by Satoshi, no?" - zcc0nonA
If scaling on chain is not necessary because 2nd layer are used, there is no point in limiting the first layer."If your second layer are as good as you say they are, this won't happen. Additionally, this is Satoshi's vision and this is what people bought into. If you want something different, that's fine, but do it in an altcoin." - deadalnix
Currently, a novice or basic VPS would be able to process 8MB blocks with ease. Couldn't help but noticed that there are about 4000 nodes and a single BTC being worth... The only likely issue could be long term storage and bandwidth. With bandwidth being addressed( 0.13) only hard-drive space remains and with my local South-African Hosting providers being able to supply dedicated hardware with 4-core 2T HDD's and ample Bandwidth for less than 90$ we should seriously consider Hetzner/link or start considering nodes in hosting environments and wallets linking on to those nodes. This should, increase block transfer times globally and allow individuals access to their BTC without waiting for their node to start and sync-up. Ease of access, faster nodes more TX per second == more use cases == more value of the BTC network."And above all, less agitated users!" - Poseidonn77
Fifteen million purchases a day amounts to 4.5 GB of blocks a day. This would require an average blocksize of 62 megabytes. My five year old desktop computer (which cost $600 new) processes about 4 GB of blocks per hour, so it would still be loafing along."The data rate required to process these blocks is 400 kbps. My pathetic rural DSL has 8 MB/s bandwidth down and 1 MB/s bandwidth up, so it would not limit my node either. The rate for Visa comes from consumer demand, not network capacity. They could increase their capacity the same way that Bitcoin's capacity can be increased: faster communications links and faster computers. Ultimately, the demand comes from economic transactions that people wish to make." - tl121
2000 tx/sec = 600MB blocks
1Gbps internet = $70-$300 per month
2.5TB storage = ~$100 per month
140W of electricity for CPU validation = $10 per month + ~$600 for CPU
One time cost for 128GB RAM for utxos = $650
Total annual cost of fully validating node at 600MB blocks ~$4500
Times 100,000 nodes = $450 million per year
Visa annual expenses ="$4.8 billion per year" - peoplma
In my opinion Satoshi meant it this way. The processing of millions of transactions is trivial and easy on today's hardware."We can very very likely scale up faster than demand for the simple reason that we are at maybe 0.5% of the max of simple hardware today. And software and hardware will very likely grown in the time that it takes for the demand to fill this up." - technicaltony
