100k tps looks way off (on your side) if you define transaction correctly, i will not anticipate that amount of volume…
I’m connected to all 48 cryptoccy exchanges and collecting from them all types of data, i will give you some numbers, but lets define whats what.
Transaction => Match between 2 orders of opposite side in exchange orderbook; actual transfer of assets between market participants
Match => Same as transaction
Order => Single operation on order book (insert, update, remove)
Right now (weekend) averaged data from last 5 minutes from all symbols (6804) listed on cryptoccy exchanges:
65 Transactions/s
1519 Orders/s
Weekends are always less traded, lets assume high activity and multiple this by 10,
also assume that you will take 50% of total cryptoccy volume,
and that crypto markets will be in one year 10 times bigger.
So at max you will need to process:
4k Transactions/s
76k Orders/s
Your estimates are 100k Transactions/s, so it’s not an issue here if we are agreeing on definitions.
More important questions for you guys:
How much orders per second you will process
How much time will need to pass between I put order in book and other market participants will see that order
Will validator see my order in book in same time as other market participants
What’s will be minimum time after what I will can cancel my order and avoid transaction match by validator (How much time will be required to cancel my order)
After Transaction/Match, how much time will need to pass to get my assets (for example i got a match on my limit order, 10 BTC/LTC), how much time will pass between match operation and moment when I will able to use my 10 BTC.
I will able to place orders for assets that i don’t currently have, but they will be in time of validator match operation?
Whats are limit related to order book operations, eg. how much orders can be placed at once, limit on total opened orders in book, mininum order size etc
@tonbi - we need what you’re building. I wish it were ready already. It’s a pain in the arse having to mess around with stuff like centralised exchanges, slippage, fees, waiting times, verification, DDOS, hacking, fraud, pftt.
I just want to swap a little bit of bitcoin for a little bit of MAID at a reasonable price!
B.T.W. it looks like Changelly doesn’t do MAID anymore, and neither does ShapeShift…
We agree on the deffinitions, however 100k tps won’t be enough on the next 3 years, i prefer to be as large as possible. No order issue here as it is not a website.
1- It doesn’t really apply as it is not a website, the informations are coming in a different way.
2- It will depend on your connection speed, if the connections are fluid, it will be near instant for everyone.
3- Yes. Except if a participant have low bandwidth.
4- Near instant, if witout matching.
5- Once there is a matching as of the actual system, we are around 0.7 sec, and can go up to 100k tps, once we explore enough 0-conf it could be instant.
6- I don’t understand here, could you please give me details ?
7- It won’t be limited, and should take about 0.2 seconds to the validator to check for the funds, and 0.1 second to deploy a 2-OF-2 (which is approximative).
Using the JAXX private key, we will provide different guides and support from slack. You are welcome.
Can someone at Maidsafe confirm this claim made earlier today by “Lauda” on the bitcointalk NVO thread?
“Every single investment is escrowed, and there are 4 independent escrows. One of which is the CEO of Maidsafe.”
Can you elaborate that point why it’s not apply? I don’t mention website any kind, just talking about changes of the book. If you match orders → make transactions, you will need maintain orderbook with updates to know which orders will need to be filled in what order, assuming price, time priority maching engine. There is a computing limit of orders after what your maching engine will be behind in time and process transactions what should happen for example few minutes ago.
Don’t have any website in mind, just exchange internal architecture 101.
Because of the crowdsale we talked about veserus, some guys wanted to send coins and use it, this is why i decided to put it down, refund people and explain them that it is still in development.
The validator processing will update the order book and then process the transaction as at this point, the order is either processed or cancelled on the orderbook side. On the user side, the transactions are either sent or cancelled.
The delay could be due to propagation issue, or update issue, even there, if there is any problem the validators will do their work and cancel the orders.
@all
I am sorry for today, i’ve been slow to answer you, finally i got tired. need some rest.
Like @BIGbtc and others I am also skeptical about the technical details of NVO’s design. RoboTeddy’s questions in the Reddit thread cut to the core of the issue, and I don’t think their questions are answered by either the whitepaper or the comments. Specifically: if the validator doesn’t hold a private key and just publishes transactions, how do you prevent a race condition? I can easily publish a conflicting transaction that gets confirmed before the validator’s transaction that takes my money. Alternatively, if the validator is to sign transactions itself as part of a multi-sig setup, then where is the validator’s private key stored? It can’t be stored on the safe network or a blockchain, because then it can be accessed by anyone (assuming the validator accesses it autonomously). Solving these problems would be a major breakthrough, and I doubt this can be pulled off in a mere 6 months.
Note that I’m speaking here as an individual, not a MaidSafe contractor. My opinions do not necessarily reflect those of MaidSafe and I make no claim to represent anyone other than myself.
If anybody can access the content i put on safeNet, there is no real usage for safenet, except file sharing, streaming and things like that as megaupload was.
What i mean, is that a separated private access for each validator instance on safenet. We count on NFS to manage this kind of access, or will it be modified too ? i would really appreciate to get that information as you work with maidsafe apparently.
No wories, even if you talked on behalf of maidsafe.
It’s not true in general that anyone can access what you put on the SAFE network – people’s private files are encrypted with a passphrase that only they know. My point is that if you have a validator deployed as a smart contract, it would need to hold a decryption passphrase (or account credentials or whatever) in order to decrypt the data it has stored… and it’s an infinite regress problem: if you try to store the credentials encrypted on the network, then you need somewhere to store the decryption key, and if you try to store that decryption key, then you need to store its decryption key and so on… The easiest way to break the chain is to have a human hold a master password (in their head or locally on their computer), but that isn’t decentralised.
To be clear, the recommended usage of the SAFE network is exactly as I’ve described: each user knows their account secret and passphrase and can use these to access their data.
Yeah thanks for saying that, I never felt like those questions in the reddit thread got answered.
And it would be an amazing breakthrough, but it sounds like this group doesn’t have the answers for it. Not to blame them, as it would be a crazy feat for any person / group. But then there’s the issue of everyone’s money…