Right!
I referred to this:
Btw. time to split the topic? This is no longer dealing with hardware, but with design and possible improvements of BCE.
It shouldn’t be burrowed in a supposedly hardware related thread.
1 Like
I haven’t done any measurement but I guess it’s indeed the SHA256.
It’s essentially a centralized solution and also a rather trivial one. In blockchains you need to deal with forks, which are semantically valid records of history but only one can be chosen. Triple receipts are still a semantic thing and doesn’t help you choose between forks. If you were not talking to me, I rest my case.
My hopefully useful takeaways and considerations.
My takeaway from the hardware configuration discussion is that reliability comes over security. It is important to have those keys available for signing in time. The risk that those keys are compromised are relatively low due to multi-sig. Having a few thousand dollars worth of funds in a single-sig account on a home-pc is likely to result in a higher risk profile given the attack vectors than some multi-sig keys on a VPS. So a VPS appears ok to me.
Tor could decrease the traceability assuming the exit node is not compromised. Did someone consider Cloudflare as a way of obfuscating the IP address of the B&C signing wallet? Should be able to use port 80 for the traffic, but I haven’t tried it yet.
Re availability, I think it is important to have recent copies of blockchains readily available in case something corrupts and a time consuming resync from scratch is required. Corruption can be caused by the media or the network. I have only seen that issue by protocol updates myself but older SSD media is also vulnerable as it wears out I’ve heard from others. This might not apply to the latest generation or professional grade SSDs.
Still making up my mind which way to go. Some experimenting is likely to be required. I might just start with 2 or 3 VMs on my highly spec’ed home server as it has plenty of diskspace which comes at a relative premium on a VPS.
Thanks! And as @masterOfDisaster said CPU occupation varies from 80-95% when Rasp minting, I believe this is single thread occupation.
Here is the SHA256 performance for Cortex A7（1.2Ghz） and Apple A9（1.85Ghz） comparison.
http://browser.primatelabs.com/geekbench3/compare/5762981?baseline=5764201
I believe single thread SHA256 comparison of 0.9Ghz Cortex A7 vs 1.85Ghz Apple A9 is around 500 vs 4000, which means a iPhone Nu app minting will lead to 10% occupation of iPhone6/6s/se’ CPU. And the data upload/download per month is around 200-300MB.
@Eleven, in your quick wallet, the instant transactions between accounts demand your server’s computation, what’s kind of it? Also SHA256? How many instant transactions per second can be performed by Intel i5/Xeon 2.4Ghz single core?
I wonder whether our B&C’s one dozen of signers can perform 50,000x per second for global customers, with i5 CPU and 10MB/s network. That is VISA/MASTER level. 
Last but not least, B & C don’t need to apply a payment license, LOL.
1 Like
I reported this CPU occupation for blockchain syncing.
It’s multithreaded, because the CPU load can exceed 100%, which is only possible, if more than one CPU is used by nud.
That looks like a prime condition for double spending on the same address. How does subsequent B&C tx know there is enough fund on a user’s address if previous tx to the address haven’t been confirmed? The fund could be transfered away by a tx outside B&C.
Because the user’s address is in the BCE context a multisig address controlled by reputed signers?
This can only happen, if reputed signers sign that tx, right?
Wen a user place an order, the fund moves from his normal address to a signer multisig address. When the oder is filled, his counterparty’s fund moves from the signer’s address to the user’s normal counter-currency address. It’s the users normal addresses I talked about.
I thought we were talking about 0-conf tx within BCE context (multisig addresses of reputed signers).
Funds that get moved from/to user’s singlesig address need to be treated differently.
I thought every trade is a pair of txs between normal-adress/signer-address? There is no fund movement within B&C i.e. not involving a normal address. no?
How do you interpret the paragraph that starts at end of page 7?
Can you clarify?
When I run “top” command, cpu load sometimes exceeds 100%, so that’s multiple program based on single core/thread occupation load calculation.
On the right top corner of raspberry, the cpu load indicated as 27%, that’s same load, but based on multi core calculation.
4 cores’ 27% = one core’ 110%
But the Nu/B&C is mainly single thread program although occasional exceeds one core’s capability.
Last question, the more NSR I have, the more burden on my raspberry’s cpu in minting?
Huh? I thought all funds at B&C would be held in multisig, the single sig addresses are only there to account for what B&C account makes which deposit right? So after the funds are deposited they are “credited” to that person’s B&C account and moved to “storage” aka multisig.
So when trading happens all funds stay within B&C multisig and don’t move at all just “messages” on the B&C blockchain as to which funds belong to which trader. How does it work otherwise? Transferring funds on their respective chains after every trade would take way too much time + tx costs.
1 Like
“Message”… You remind me of the Peermessage project, how is it going?
It seems that “Open transaction” uses Bitmessage for communication among trading parties. Then we clone Peermessage into B&Cmessage service, Can we use it as secure information tunnel among signers?
Hi Sabreiib, Actually we do quickwallet for our exchange Huobi. If a user used quickwallet to transfer his asset to huobi, huobi don’t need wait for the transaction confirmation and Huobi would immediately add the corresponding amount to the user’s account. Because Huobi believe our quickwallet don’t let users double-spending. As our quickwallet is 2-2 multi-sig.
Does Huobi have a 3rd party payment lisence in China?
In fact, if we build up a B&C account instant payment system, we can competite with Alipay/Applepay/paypal/VISA.
Can someone clarify? I think this is a very important question.
1 Like
Is this the “escrow system” called by @JordanLee?
unfortubately, the fund is insufficient to develope such off-chain transaction.