What do people think the possible implications of this could be? This is also just as the devs have moved to MIT isn’t it?..
Peter Todd has poo pooed this. See the link to an interview with him in my tweet:
I think it’s a fundamental flaw in block-chain technology. Hosting the block-chain on SAFE doesn’t really solve the core problem, then Bitcoin will rely on two decentralization protocols and will be vulnerable to weaknesses of both. It would mean admitting that block-chain technology can’t scale on it’s own.
It’s not blockchain tech in particular. It’s the proof-of-work that seems to be the real bottleneck.
I don’t see how the type of proof is relevant here, it’s about the cost of running a full node (storing the full block-chain history, validating all transactions, communicating these large blocks) if the blocks become larger and larger. Only the big guys will be able to run full nodes, causing more and more centralization.
I don’t see the size as an issue as eventually storing the chain on the SAFE network becomes a possibility.
The bottleneck is the number of transactions per second (currently only 7? - unbelievable) Going to the larger block size is supposed to increase this a fair amount (which is why it is proposed), but seriously, it’s not going to make that much difference in the long run I think. Proof of Stake (NXT) and proof of resource (farming - MAIDSAFE) brings a lot more nodes online and increases the ability to process transactions.
I don’t think the Bitcoin devs and community will see that as a real solution, as the decentralization (and thus security) of Bitcoin would be dependent on the reliability and security of SAFE.
There is a lot of controversy around that proposal, because a larger block size makes maintaining a full node more costly, like I described in my previous post.
This isn’t about the type of proof, it’s about the consensus model. To make the block-chain work it is required that all miners (well, the full nodes at least) store and verify all transactions. So having more nodes online doesn’t increase the ability to process transactions, because they all have to verify and store the exact same transactions.
SAFE’s consensus model doesn’t use a block-chain, in SAFE only a few chained consensus groups of 32 nodes are involved in a SafeCoin transaction, and there’s no giant public ledger in SAFE. That means that my and your vault will usually handle different transactions. The load of handling transactions is distributed over the nodes of the network, which is not the case with a block-chain.
The type of proof in a block-chain technology is only there to decide who gets to make the next block while stopping Sybil and double-spending attacks. It doesn’t affect the processing and maintenance costs of the block-chain data itself. Proof of stake only makes the process of getting a chance to mine a block cheaper. The blocks themselves don’t get smaller because of it.
An increased blockchain size is not sustainable… 20MB blocks * (24 hours in a day * 60 minutes in an hour / 10 minutes) = 2.8 GB a day.
There’s no such thing as a perfect solution and they’ve been struggling to solve this for years now, and so far this seems the only option that actually might achieve the aims. Nothing else is even on the table (despite Gavin’s block size proposal).
The real alternative would be a SAFE concensus model for a ledger based currency alongside Safecoin, or as an option for Safecoin transactions. SAFE already supports permanently immutable data, so once Safecoin is running, it should be trivial to have the option of recording transactions in a public or private ledger.
Now that would be a solution, but the bitcoin devs are out in the cold then, so it won’t be a bitcoin driven solution, it will be an alternative that is adopted because it provides the functionality of bitcoin without the drawbacks.
With bitcoins proof of work system only a small handful of elite miners get to verify transactions - so it is about this and not about the size of the blockchain (currently). If I could still mine bitcoin I would, large hard-drives are not that expensive, but ASIC’s are.
With NXT even a phone can verify transactions. So there are many more people doing so.
The benefit of accepting a currency with a non-optional public ledger is that as an organization you instantly have proof of transparency and can thus be considered as more accountable. Don’t get me wrong, I love the fact that SafeCoin transactions can be private, and I think that we should and will have an (optional) public ledger for SafeCoin transactions somehow like you suggest. But maybe we can also think of some system that can serve a proof of transparency?
Something like a special type of SAFE account, of which perhaps the core SAFE protocol automatically publicly records all incoming and outgoing SafeCoin transactions. I bet @Warren would love that.
The type of proof affects centralization, increased block size also affects centralization. That doesn’t mean that the matter of block size is about the type of proof. Yes, ASIC mining in proof of work is a problem, but it’s not the problem this thread is about. It’s about the block size (see the title), and the type of proof in a block-chain doesn’t affect the block size.
I read this article yesterday. The proposal is to deal with the low transaction rate of bitcoin. The type of proof affects the transaction rate. So it is fully relevant.
Then you know something that I don’t, as far as I know the type of proof has nothing to do with transaction rate?
No, after looking, I think you are right, I’m confusing proof of stake with transparent forging (a NXT feature to increase the transaction rate) and once I confused that I then made the assumption about bitcoin proof of work. I’m not 100% sure I’m wrong, but I can’t find the evidence to back up my claim, so I’m going to give this one up. I have errred - my apologies.
Happens to the best.
@Seneca you’re right.
I just wanted to put this into perspective here of how devistating the increased blockchain size would be… I only had to time to type up that little blurb up there before I had to leave this morning. With a blockchain transaction limit of 20MB blocks every 10 minutes… 2.8GB a day. If the blockchain actually increased by this much every day, within a year the blockchain would be almost 1 TB in size!
2.8 GB a day * 365 days in a year = 1022 GB
With a blockchain that big… I couldn’t even run the bitcoin core on my computer even if I wanted to! My hard drive only has 300GB of storage… and even less on my laptop (with an SSD). If the blockchain didn’t reach this size in 1 year, what about in 2 years? or 5?
This is one thing that sets Maidsafe apart from bitcoin. With Maidsafe, everyone can contribute. Not just people with 20-30 GB of storage space. The thing I realized after being around bitcoin for a few years is… bitcoin is filled with arbitrary code. Bitcoin tends to create it’s own problems. Why wait 10 minutes for a block of transactions? Why set a block size limit of 1MB? Why the transaction fees??? Now I know there are reasons behind each of those, but we have the ability to build a system that is not unreasonable, or based on an individual’s wishes. When Maidsafe comes out, it will prove what decentralized technology really can do!
Lightning network is a proposed solution that afaik, all of the core devs are supportive of. It requires patching the malleability issue in bitcoin in a soft fork.
Gavin’s proposal is certainly a short-term solution that doesn’t really do much good in the short term.