BitShares Forum
Main => General Discussion => Topic started by: questionsquestions on September 22, 2014, 02:51:19 pm
-
I've been doing more digging around the wiki and I've come up with some questions I cannot find satisfactory answers to. Could someone outline and honest and realistic appraisal of the limitations of Bitshares vs Bitcoin? Specifically;
- Scalability: Bitcoin suffers from a maximum theoretical limit of 7.2 transactions-per-second (without a hard fork (http://"https://en.bitcoin.it/wiki/Hardfork")). While the current traffic hasn't reached anywhere near this number, if Bitshares were to take off it is entirely possible that the number of transactions would grow rapidly. So;
- What's the maximum number of transactions Bitshares can accommodate in a block?
- Does Bitshares have the ability to scale the size of blocks to accommodate growth
- How large is the blockchain in Bitshares likely to be vs Bitcoin? Bear in mind that Bitcoins full blockchain is running at ~30 GB currently
- How will a large blockchain affect new clients entering the eco-system? Bitcoin offers work-arounds such as SPV and server based offering such as Electrum - how does Bitshares accomodate this?
-
- Fees: The fee structure appears to be a mandatory minimum 0.5 BTSX / operation. Fees do not appear to be refunded in the event of a cancellation
- Can fees be varied if the value of BTSX grows? Bitcoin suffers from its own success in this regard. Transactions fees are generally 0.0001 BTC, which when Bitcoin was circa 1000 USD was the equivilent 10 cents just to send any value through the system. This made micro-payment scenarios uneconomical vs incumbent solutions like credit cards. NB: Floating fees aimed to address this, but Gavin Andressen indicated that in some cases the transaction fees would actually go up rather than down.
- Through the distributed exchange mechanism - is it possible to receive a rebate for a cancelled trade? For example; a limit order isn't filled
-
- Distributed Exchange:
- Do the market orders pay the exact price offered by a BUYer, or do they cross over the order-book ala Bitstamp? For example; I want to buy 10 BITUSD, there are 5 on offer at 1 BTSX each and 5 on offer at 2 BTSX each. If I make a limit order and offer 15 BTSX, do I end up with 10 BITUSD (5 BITUSD @1 BTSX each and 5 BITUSD @2 BTSX each) or 7.5 BITUSD - meaning I pay 2 BTSX each for each BITUSD, until all 15 of my BTSX are consumed?
- Is it possible to create a market order to buy or sell assets? For your average consumer who just wants to move wealth in and out (akin to normal forex transfers), a limit order would mean waiting around for it to fill or paying above market price?
Anything else that paint a true picture of the limitations of Bitshares.
Thanks for all the responses.
-
I've been doing more digging around the wiki and I've come up with some questions I cannot find satisfactory answers to. Could someone outline and honest and realistic appraisal of the limitations of Bitshares vs Bitcoin? Specifically;
- Scalability: Bitcoin suffers from a maximum theoretical limit of 7.2 transactions-per-second (without a hard fork (http://"https://en.bitcoin.it/wiki/Hardfork")). While the current traffic hasn't reached anywhere near this number, if Bitshares were to take off it is entirely possible that the number of transactions would grow rapidly. So;
- What's the maximum number of transactions Bitshares can accommodate in a block?
We have already tested 10 tps. Theoretically DPOS can scale to the size of visa and cost the network a few million dollars as opposed to a few hundred million dollars with Bitcoin and a few billion dollars in the case of visa.
- Does Bitshares have the ability to scale the size of blocks to accommodate growth
Yes
- How large is the blockchain in Bitshares likely to be vs Bitcoin? Bear in mind that Bitcoins full blockchain is running at ~30 GB currently
It is currently ~500mb. The blockchain size is not really a problem. The average user will not have to download the whole blockchain in the future.
- How will a large blockchain affect new clients entering the eco-system? Bitcoin offers work-arounds such as SPV and server based offering such as Electrum - how does Bitshares accomodate this?
Expect an SPV over the course of the next 2-3 months.
[/li][/list]
-
- Fees: The fee structure appears to be a mandatory minimum 0.5 BTSX / operation. Fees do not appear to be refunded in the event of a cancellation
The mandatory transaction relay free is 0.1 BTSX not 0.5 BTSX. 0.5 BTSX is the default setting in the wallet.
- Can fees be varied if the value of BTSX grows? Bitcoin suffers from its own success in this regard. Transactions fees are generally 0.0001 BTC, which when Bitcoin was circa 1000 USD was the equivilent 10 cents just to send any value through the system. This made micro-payment scenarios uneconomical vs incumbent solutions like credit cards. NB: Floating fees aimed to address this, but Gavin Andressen indicated that in some cases the transaction fees would actually go up rather than down.
Delegates can set fees. The fees in all of these systems can be changed.
- Through the distributed exchange mechanism - is it possible to receive a rebate for a cancelled trade? For example; a limit order isn't filled
I suppose this is possible but it is counterproductive. When you make a transactions you are assessed a fee to pay for the cost of broadcasting and storing that transaction. The exchange doesn't take a percentage from each trade like other exchanges, so you are saving a lot more money than you realize even if you place an order and have to cancel it.
[/li][/list]
-
- Distributed Exchange:
- Do the market orders pay the exact price offered by a BUYer, or do they cross over the order-book ala Bitstamp? For example; I want to buy 10 BITUSD, there are 5 on offer at 1 BTSX each and 5 on offer at 2 BTSX each. If I make a limit order and offer 15 BTSX, do I end up with 10 BITUSD (5 BITUSD @1 BTSX each and 5 BITUSD @2 BTSX each) or 7.5 BITUSD - meaning I pay 2 BTSX each for each BITUSD, until all 15 of my BTSX are consumed?
You get exactly what you ask for. So if you ask for 10 BitUSD @ 1BTSX/bitUSD (cost of 10 btsx), then the market buys 5 bitUSD @ 2btsx/bitusd (cost of 2.5 btsx) and 5 bitUSD @ 1btsx/bitusd (cost of 5 btsx), gives you 10 bitusd and pockets the remaining 2.5 btsx as a market fee.
- Is it possible to create a market order to buy or sell assets? For your average consumer who just wants to move wealth in and out (akin to normal forex transfers), a limit order would mean waiting around for it to fill or paying above market price?
A market order is just buying at lowest ask or selling at the highest bid. You can obviously do that now...
[/li]
[/list]
Anything else that paint a true picture of the limitations of Bitshares.
Thanks for all the responses.
Basically BitShares has no limits lol
-
In BTSX exchange you always "get what you ask for" in the exchange... ie: if you are willing to pay X per Y then you will pay X for all Y even if some Y are available for less than X. This prevents front running and encourages traders to trade on value rather than technicals. If you want a "market order" that will have to be done slowly via multi-step process.
-
In BTSX exchange you always "get what you ask for" in the exchange... ie: if you are willing to pay X per Y then you will pay X for all Y even if some Y are available for less than X. This prevents front running and encourages traders to trade on value rather than technicals. If you want a "market order" that will have to be done slowly via multi-step process.
Ok, that makes sense. What about the other questions? I find it slightly hard to believe:
Basically BitShares has no limits lol
however well thought-out and implemented Bitshares may be.
It generally makes it easier to convince people of the merits of a system if you can also indicate its short-comings. Bitcoin has plenty of shortcomings, but it does have the benefit of the network effect.
I'd appreciate if anyone can provide any more detailed answers to the questions I've asked.
Thanks once again.
-
The limits of our "testing" are 10 Transactions per second....
The limits of our protocol which forks weekly at this point in time are "unknown" rather than non existent.
The size of our block-chain will grow at a rate similar to Bitcoin for equal transaction volumes.
Light weight clients are possible.
The question is perhaps poorly phrased: Bitcoin could easily be hard-forked to support 1000 transactions per second, but then it would become fully centralized and everyone would have to use light weight clients.
Bitshares could also support 1000 transactions per second, but the centralization would be tempered by delegate approval.
Our blockchain has a larger "on-disk" and "in memory" footprint that Bitcoin-Core because we have more indexes generated from the blockchain data itself.
-
How would a lightweight client keep track of which delegate signatures are legitimate?
Wait, they could just check if a majority of known-legitimate delegates accept a block produced by a new delegate.
But if the delegate turnover rate ever exceeds fifty percent in one round, then the lightweight clients would be unable to pass that point, and in-fact the kicked-out delegates would be able to form their own fork.
But wait a minute, a fifty-one delegate bloc could make a fork that even heavyweight clients would accept. I guess this objection is irrelevant because you always trust fifty-one delegates, heh.
-
How would a lightweight client keep track of which delegate signatures are legitimate?
Wait, they could just check if a majority of known-legitimate delegates accept a block produced by a new delegate.
But if the delegate turnover rate ever exceeds fifty percent in one round, then the lightweight clients would be unable to pass that point, and in-fact the kicked-out delegates would be able to form their own fork.
But wait a minute, a fifty-one delegate bloc could make a fork that even heavyweight clients would accept. I guess this objection is irrelevant because you always trust fifty-one delegates, heh.
Light weight means there is trust. That trust can be placed in anyone running a full node and can be redundant (checking with more than one source).
BTSX by its nature cannot support light weight client validation (no Proof of Stake coin can). Light weight clients simply trust those with full clients.
If you need "heavy trust" then you can ask the heavy weight client to post a bond and sign under risk of forfeiture of their bond. This will allow you to trust them for amounts up to their bond + value of all future revenue and trust they may have in the industry.
-
Light weight means there is trust. That trust can be placed in anyone running a full node and can be redundant (checking with more than one source).
I really don't like the sound of that. Can't we argue that lightweight clients trust that the delegates have little incentive to falsify blockchains since they will eventually be caught and fired? I see no reason why the lightweight client can't keep proof of transaction existence (Merkle branches and block headers) signed by delegates for all received funds so that if they later find out it was invalidated by a double-spend they can broadcast the proof to the network for all the full clients to vote out the evil delegates (actually the lightweight clients could check the proof and vote out the evil delegates too).
-
Thanks for the answers. This response:
The question is perhaps poorly phrased: Bitcoin could easily be hard-forked to support 1000 transactions per second, but then it would become fully centralized and everyone would have to use light weight clients.
begs one question; does Bitshares have a mechanism for enforcing protocol upgrades? As you mentioned, Bitcoin can be hard-forked - but the inherent risks are significant if a majority of the network does not follow suit. Where the majority does not upgrade, there is a high likelihood of two chains occurring as happened back in 2013 in Bitcoin when an upgrade of the database storage introduced blocks incompatible with legacy clients.
If two or more chains exist for any period of time, there is likely to be a cascade loss of confidence in the system. This, as I understand it, is one of the primary reasons that upgrading Bitcoin is so difficult. So, does Bitshares have some way of managing this?
-
Thanks for the answers. This response:
The question is perhaps poorly phrased: Bitcoin could easily be hard-forked to support 1000 transactions per second, but then it would become fully centralized and everyone would have to use light weight clients.
begs one question; does Bitshares have a mechanism for enforcing protocol upgrades? As you mentioned, Bitcoin can be hard-forked - but the inherent risks are significant if a majority of the network does not follow suit. Where the majority does not upgrade, there is a high likelihood of two chains occurring as happened back in 2013 in Bitcoin when an upgrade of the database storage introduced blocks incompatible with legacy clients.
If two or more chains exist for any period of time, there is likely to be a cascade loss of confidence in the system. This, as I understand it, is one of the primary reasons that upgrading Bitcoin is so difficult. So, does Bitshares have some way of managing this?
This question is very important and must be addressed.
-
Voters elect delegates that upgrade as they wish.
-
Upgrading the protocol is easy if the community understands why it is being done and agrees with it, and/or trusts the lead developers. The lead developers simply update the software and users upgrade. Its no big deal.
Also, updates include checkpoint ids which deal with forks.
-
In BTSX exchange you always "get what you ask for" in the exchange... ie: if you are willing to pay X per Y then you will pay X for all Y even if some Y are available for less than X. This prevents front running and encourages traders to trade on value rather than technicals. If you want a "market order" that will have to be done slowly via multi-step process.
Ok, that makes sense. What about the other questions? I find it slightly hard to believe:
Basically BitShares has no limits lol
however well thought-out and implemented Bitshares may be.
It generally makes it easier to convince people of the merits of a system if you can also indicate its short-comings. Bitcoin has plenty of shortcomings, but it does have the benefit of the network effect.
I'd appreciate if anyone can provide any more detailed answers to the questions I've asked.
Thanks once again.
I provided the answers to your question in my original post in bold.
-
Upgrading the protocol is easy if the community understands why it is being done and agrees with it, and/or trusts the lead developers. The lead developers simply update the software and users upgrade. Its no big deal.
Also, updates include checkpoint ids which deal with forks.
No offense, but that sounds dismissive. An earlier post indicted the problems that occurred with Bitcoin in 2013 when some of the clients didn't upgrade, thus making it difficult to integrate the pre / post fork chains to ensure a trustworthy bitcoin network. It became a big deal for Bitcoin when everyone didn't upgrade. Is that scenario also possible with BitShares and if not why not? How is it different?
-
Upgrading the protocol is easy if the community understands why it is being done and agrees with it, and/or trusts the lead developers. The lead developers simply update the software and users upgrade. Its no big deal.
Also, updates include checkpoint ids which deal with forks.
No offense, but that sounds dismissive. An earlier post indicted the problems that occurred with Bitcoin in 2013 when some of the clients didn't upgrade, thus making it difficult to integrate the pre / post fork chains to ensure a trustworthy bitcoin network. It became a big deal for Bitcoin when everyone didn't upgrade. Is that scenario also possible with BitShares and if not why not? How is it different?
In bitshares client upgrade is irrelevant. As long as some delegates upgrade the network will continue to function normally.
-
Voters elect delegates that upgrade as they wish.
How consolidated are btsx shares at this point?
I remember the idea of a btsx rich list was tossed around but with TITIAN I would imagine it might be difficult. I would be curious to see how much everyone's votes really matter in comparison with the large holders.
-
Voters elect delegates that upgrade as they wish.
How consolidated are btsx shares at this point?
I remember the idea of a btsx rich list was tossed around but with TITIAN I would imagine it might be difficult. I would be curious to see how much everyone's votes really matter in comparison with the large holders.
Why dont you start a poll asking how many BTSX people have.
-
Voters elect delegates that upgrade as they wish.
How consolidated are btsx shares at this point?
I remember the idea of a btsx rich list was tossed around but with TITIAN I would imagine it might be difficult. I would be curious to see how much everyone's votes really matter in comparison with the large holders.
Why dont you start a poll asking how many BTSX people have.
I could but I doubt the big holders will participate, which would make it useless.....
-
I provided the answers to your question in my original post in bold.
I see - I missed them since they were embedded inside the quoted text. Thanks alot!
In response to to this question:
Is it possible to create a market order to buy or sell assets? For your average consumer who just wants to move wealth in and out (akin to normal forex transfers), a limit order would mean waiting around for it to fill or paying above market price?
A market order is just buying at lowest ask or selling at the highest bid. You can obviously do that now...
A market order isn't a straightforward BUY or SELL at the lowest ASK or BID price though. A market order steps up the price paid, or drops down the price offered until either the requested number of assets has been obtained or expended. E.g. Sell 1000 BTSX for BITUSD - at whatever price you can get, and keep selling until all the BTSX have been sold.
Buying at either the lowest ask, or selling to the highest bid would be a limit order. E.g. Sell whatever I can into this BID or ASK, and then just leave the rest on the order book if the full amount was not sold or purchased.
So, the question still stands; does the ability to liqudate or fill a position at any price automatically exist, or does the system only support limit orders?
-
Voters elect delegates that upgrade as they wish.
How consolidated are btsx shares at this point?
I remember the idea of a btsx rich list was tossed around but with TITIAN I would imagine it might be difficult. I would be curious to see how much everyone's votes really matter in comparison with the large holders.
Why dont you start a poll asking how many BTSX people have.
I could but I doubt the big holders will participate, which would make it useless.....
You can try to analyze genesis block: https://raw.githubusercontent.com/dacsunlimited/bitsharesx/master/libraries/blockchain/genesis.json
Probably you understand what even in Bitcoin you can't make exact rich list, you only can see the top of unspend outputs and can't know if few of them belong to same owner.