"Bitshares 2.0 - Over 100K Ts/sec"

https://bitshares.org/technology/industrial-performance-and-scalability

They use “Delegated Proof Of Stake” and claim the ability to do 100.000 transactions a second with Bitshares 2.0. The technology is partly based on LMAX and asks miners to use a great amount of RAM to speed things up. The blocktime could be under a second.

From their website:

"Overview
To achieve this industry-leading performance, BitShares has borrowed lessons learned from the LMAX Exchange, which is able to process 6 million transactions per second. Among these lessons are the following key points:

Keep everything in memory.
Keep the core business logic in a single thread.
Keep cryptographic operations (hashes and signatures) out of the core business logic.
Divide validation into state-dependent and state-independent checks.
Use an object oriented data model.
By following these simple rules, BitShares is able to process 100,000 transactions per second without any significant effort devoted to optimization. Future optimizations are expected to bring the performance of BitShares to levels similar to LMAX."

More on LMAX in this video.

5 Likes

A really good read thank you for sharing. A side note, BTS is up 55%+ today; weird?

That’s what caused my attention, I saw them higher and was looking for the news behind it. I didn’t found any, but I noticed they’re working on a 2.0 version. I have no idea whenever it will be out, but I think they’s doing a very nice job from a blockchain point of view.

This news is couple of months old (I listened to the interview with the chief developer back in June or May).

Truth is, there’s nothing that needs to be traded at that rate, and if there was, it wouldn’t be traded on a PoS coin.

Do you think BTS is integrated as well as it is for safecoin? Will transaction fees be enough to reward new development? I have been considering spreading my eggs a little more and Ethereum’s approach of flashy up front, partnering with corporations and difficulty knowing how it will be self-sustainable?

Well, there might be. Let’s look at Ethereum for a second, their system can handle between 7 to 100 Ts/sec. What if people start to play Satoshi Dice on each block? With a blocktime of 12 seconds, this means that with 120 Ts/block the system starts to show delays. So when someone builds an App to gamble on the outcome of then next block, at 120 users, the system might start to crack. Not to mention Augur that wants a few thousand people sending the price of oil to a block. What are the Devs gonna do? Ask people politely to use another App that doesn’t interact with the blockchain each block? Block the App? It’s a decentralized system, quite hard to do.

They won’t block the app, the cost of transaction will go up so that you won’t be as eager to use the system.

By the way, you may want to take a look at various payment channel innovations that will allow similar transaction rates on Bitcoin.

(Another factoid that came to light today is that BitUSD (see on coinmarketcap.com) is way over $1. Totally unreliable and ludicrous for anyone who may think these “instruments” have any practical purpose for people with serious amounts of money).

2 Likes

Ha, yeah, I never understand why people are so eager to create “new ways” to peg a cryptocurrency to common one. I also believe all these will fail. That’s already fixed on the normal markets. These are called futures and CFD’s and they work very well. So all you have to do is create a smart contract that does just that, combine a currency with a CFD and you’re done. Only tricky part is the datasource. Augur might have a solution to that. Vitalik also offered a solution called Shelling coin.

2 Likes

If you dare to peg to the dollar to order to have same value, you’re gonna have a bad time.

bitshares is a banking system. Nobody should be touching it with 100 foot pole.

It’s actually true. Bitshares can do at a minimum 100,000 transactions per second. The architecture they are using is actually capable of 6 million (LMAX) transactions per second.

The reason Bitcoin cannot do this is specifically because it is not able to scale and keep it’s Proof of Work consensus algorithm. On the other hand the evidence is that Delegated Proof of Stake is adequately secure for these purposes. The reason Bitshares can do this is because it’s not a general purpose monolithic design like Bitcoin or Ethereum but it is instead a highly specialized design.

For people who know about CPUs the Bitshares architecture is like RISC while Bitcoin and Ethereum are more like CISC. Bitcoin has to be backward compatible with Satoshi’s intentions which restricts it from being able to achieve an optimal evolutionary trajectory. Ethereum is restricted by the hardcoded decisions of it’s developers, one of which may have been a mistake is Turing completeness.

I’ve bet on Bitshares being capable of doing at least 100,000 transactions per second and now my money and reputation are on the line.

How do you expect people to accept cryptocurrency if they have to worry about volatility? The Bitshares peg does hold and I’ve tested it. The purpose of Bitshares is to allow people to trade virtual USD without having to trust a counter party or centralized exchange. The point is to be a decentralized exchange and if you want to go the centralized route then go ahead and pay the millions of dollars in regulatory fees, then trust some company like BitReserve.

If you don’t want that then there is Bitshares and technologically speaking no other architecture can compete with it at this moment in time. It has a niche and that is to be the best decentralized exchange.

If they only had someone who wants to transact.

I said at the very top that latest advances in payment channels technology allow Bitcoin to achieve very high transaction rates, but who cares when you can make unsubstantiated claims while ignoring all other comments.

:smile: What reputation? Do you mean on this forum?

1 Like

Bitcoin can never achieve that high rate without forcing you to trust centralized entities. Payment channels or off-chain transactions involve trusting centralized entities. Show me otherwise.

If Bitcoin is to remain decentralized by design then it will never be able to scale. Maybe it’s more important for Bitcoin to be useful to activists than to have it be useful to Wall Street so it might not even be important that Bitcoin try to do that.

My reputation on this forum and in any other forum matters to me. If I say that Bitshares can do over 100k transactions per second then if it turns out to be true it boosts my reputation. It means I told it as I saw it instead of adjusting the facts to fit political sensibilities.

Bitshares as I see it spec for spec is the best decentralized exchange technology. BitUSD does hold the peg. Bitcoin will never be able to scale to do what Bitshares can do and still be Bitcoin as we know it. The Lightening network is not Bitcoin and while you can add that to it, or raise the max block size, it’s not necessarily important for Bitcoin to be a decentralized exchange with 100,000 transactions per second.

Bitcoin doesn’t have the governance structure to decide certain things. Delegated Proof of Stake gives you a governance structure to decide how the DAC should evolve. BItcoin isn’t sure if it’s a DAC or a Dapp or what because the core developers are in a feud. SAFE Network seems to want to be private, anonymous, file storage and decentralized computing eventually.

SAFE Network will probably never be able to do 100k transactions per second but it will never need that. They are on different evolutionary trajectories, with different design decisions.

Bitshares is a PoS coin where you trust Bitshares holders to do the right thing, which by design is proven to be impossible over a long term because it is in the holders interest to provide their own version of history (as opposed to the bitcoin blockchain, which is everyone’s version of history).

Payment channels implementations have a number of mechanisms to mitigate or completely avoid centralization.
For example - and this is not the best implementation - one can use several decentralized trusted oracles (https://github.com/orisi/wiki/wiki/Orisi-White-Paper). There are also models for completely trustless payment channels, I just can’t find the PDF link now.

That depends on your definition of “scale”. We already know that the maximum block size will be raised to at least 2MB, and by next year we may have first trustless off-chain transactions (although I would argue that off-chain micro-transactions don’t need to be trustless - personally I don’t see a situation in which I would have more than 1 micro transaction going on at any given time, so if I lost $3.99 (in BTC), I couldn’t care less).

But in any case, a combination of trustless payment channels and a larger block size will be sufficient.

I know that NYSE will use Bitcoin (colored coins) and several other implementations are currently being evaluated, while a bitcoin-based bond was issued by Overstock. Maybe Bitshares is targeting high frequency trading, but why use Bitshares for that when stocks are bought and sold in microseconds and the system works fine as-is?
The main problem is to turn T+3 into T+0 and Bitcoin will be successful in that, and especially in bonds.

There isn’t a hard limit on the number of Bitshares holders. All information security platforms work on “minimal” trust. There is no such thing as “trustless”. In Bitcoin you’re trusting the core developers and the mining elites which is actually less people than you have to trust with Bitshares. Also Bitcoin cannot be reconfigured on the fly so with Bitshares you can adjust the number of delegates or witnesses or people who have to be in different positions, and you can choose your people or kick them out.

Show me more about payment channels. If the transactions take place off chain then it’s centralized by my understanding. So if it’s not off chain then it must be on chain?

I wouldn’t choose to use Bitcoin. I think that may have been a mistake on the part of NYSE but that might also be part of why there is an intensity in the max block size debate. One faction wants to bring Bitcoin mainstream while another faction wants to keep Bitcoin an activist oriented currency. We don’t know which faction is going to win, we don’t have a process for governance, and that is part of the problem I see with Bitcoin.

There has to be a process for deciding the evolutionary trajectory of the project. Bitcoin is at a crossroad but in my opinion we should not assume it will ever be adopted by the mainstream. Suppose for example the core developers choose to make it a more appealing currency for activists by making the currency more anonymous? In doing that it will make it less attractive to NYSE and to the mainstream audiences.

There’s a natural tendency of such platforms to centralize (maybe for PoW as well, we’ll see).
Is there a graph or other info that shows the number of Bitshares stake-holders over time?

I put a link in my previous comments, it’s a WP.

Also see https://en.bitcoin.it/wiki/Contract#Trust_minimization:_challenges.

I agree with this but at least with Bitshares if it does centralized we can do something about it. We can vote them out of their positions.

Bitshares is not perfectly decentralized, it’s just more decentralized than Bitcoin.

Bitshares is currently centralized around it’s core developers just like with Bitcoin. This is because very few people have the knowledge to actually build and improve on Bitshares. So it’s knowledge centralization and centralization of development resources.

Over time that might be less of a problem and Bitshares can continue on with new developers if necessary.