阅读此篇论文受益匪浅，特此共享。接上一部分 坚定选择BCH的原因-2
Investigation of the potential for using the bitcoin blockchain as the world’s primary infrastructure for Internet commerce
调查将比特币区块链作为全球互联网商务主要基础设施的潜力
五. Results and discussion
五.结果和讨论
5.1. Cryptographic functions benchmarking
密码功能基准测试
The four panels in Fig. 2 show sample simulation data for arbitrarily chosen scales.
图2中的四个面板显示了任意选定比例的样本模拟数据。
The data represent one configuration of the processors and are intended to demonstrate the relationship between the block size (in Mb), the CPU processing power (in TFlops), and the amount of ECDSA processes to be processed (in number of processes per transaction).
这些数据表示处理器的一种配置，旨在演示块大小（以Mb为单位），CPU处理能力（以TFlops为单位）与要处理的ECDSA进程数量之间的关系（每个事务处理的数量） 。
These graphs show that the relationships closely approximate linearity across the scales likely to be experienced in practice.
这些图显示，这些关系在实践中可能经历的尺度上非常接近线性。
Different simulation settings were trialed for a single node with a combination of NVidia, Xeon Phi such that the smaller sections of code (smaller registers) are routed to NVidia, the large code registers are route to Xeon Phi, and the remaining codes balanced between the two.
使用NVidia，Xeon Phi的组合来对单个节点进行不同的仿真设置，以使较小的代码段（较小的寄存器）被路由到NVidia，大型代码寄存器路由到Xeon Phi，余下的代码在二。
Of note is that the validation of the cryptographic functions including hashes and ECDSA signature verifications consume more time than the network transaction propagation latency.
值得注意的是，包括散列和ECDSA签名验证的密码函数的验证消耗比网络事务传播延迟更多的时间。
The simulation of tokenized9 processes requires more processing than the validation of signatures for standard transactions:
标记化9进程的模拟比标准事务的签名验证需要更多的处理：
the introduction of non-standard transactions requires additional signature checks.
引入非标准交易需要额外的签名检查。
A single node dedicated to ECDSA running with 20 E5 Intel cores10 can scale to be able to handle a maximum of 36-Mb blocks at a total average processing speed of 345 TPS (Fig. 2(b)).
运行20个E5英特尔cores10的ECDSA单个节点可以扩展到345TPS（图2（b））的总平均处理速度，最多可以处理36Mb的数据块。
Using either Xeon Phi or NVidia clusters we have been able to significantly increase this rate.
使用Xeon Phi或NVidia集群，我们已经能够显着提高这个速度。
We have noted that the Xeon Phi coprocessor cards handle the validation of ECDSA signature checking in a manner that is far superior to the NVidia GPU-based CUDA system.
我们注意到Xeon Phi协处理器卡处理ECDSA签名检查的验证的方式远远优于基于NVidia GPU的CUDA系统。
Moreover, NVidia or even the use of systems in conjunction with specialized ASIC systems is able to validate and process hash-based calculations much more efficiently than the Xeon Phi.
而且，NVidia甚至与专用ASIC系统结合使用的系统能够比Xeon Phi更有效地验证和处理基于散列的计算。
Therefore, it is unlikely that we will see a node based on a single architecture being developed in future.
因此，我们不太可能看到将来基于单个架构开发的节点。
Rather, we expect to see a clustered implementation of both Xeon Phi and NVidia machines in a clustered infrastructure.
相反，我们希望在集群基础架构中看到Xeon Phi和NVidia机器的集群实施。
This is to achieve the best return of power consumed to the number of Flops processed, thereby suggesting that we need to move towards co-processor-based systems.
这是为了达到处理Flops数量所消耗的最大功率，从而表明我们需要转向基于协处理器的系统。
Fig. 2. Four different scales of sample data showing relationship between block size, CPU processing power, and number of ECDSA signatures to be processed
图2.四个不同尺度的样本数据，显示了块大小，CPU处理能力和要处理的ECDSA签名数量之间的关系
We see from the data presented in Fig. 2(a) that nodes scale with processing power and that as we increase the number of transactions being processed in the block, that the number of ECDSA operations that are processed within a transaction matters less.
我们从图2（a）中提供的数据看到，节点随处理能力而变化，并且随着我们增加块中正在处理的事务的数量，在事务中处理的ECDSA操作的数量更少。
The simulations proved that it is possible to scale the Bitcoin protocol to support a block size exceeding 300 GB.
模拟证明，可以扩展比特币协议来支持超过300 GB的块大小。
This can be achieved by combining parallel processing and the validation of transactions in the software and the use of a high-powered system.
这可以通过结合并行处理和软件中的事务验证以及使用高性能系统来实现。
A 3-petaFlop HPC system can process a transaction stream 300,000 times greater than the network currently supports.
3-petaFlop HPC系统可以处理比网络当前支持的交易流量大30万倍的交易流量。
This is more than 1,000 times the processing capacity of the existing VISA network.
这是现有VISA网络处理能力的1000倍以上。
In the existing Bitcoin-QT system (Fig. 2(c)), the number of ECDSA processes within each transaction makes a significant difference to the amount of processing power utilized by the system.
在现有的比特币QT系统（图2（c））中，每次交易中ECDSA进程的数量与系统使用的处理能力的数量有很大的差别。
By parallelizing these processes, we significantly increase the ability of the system to scale (Fig. 2(b)).
通过并行化这些过程，我们显着提高了系统的扩展能力（图2（b））。
In an optimized form (Fig. 2(d)) that has been developed to process and validate all transactions in under 500 ms, the use of CUDA-optimized code allows for a system running dual NVidia K80 GPU cards to process and validate an increased block size of over 300 Mb on a single hardware node.
在一个优化的形式（图2（d）），已经开发处理和验证在500毫秒以下的所有交易，使用CUDA优化代码允许系统运行双NVidia K80 GPU卡来处理和验证增加在单个硬件节点上的块大小超过300 Mb。
Fig. 3. Relationship between block size, CPU processing power, and number of ECDSA signatures to be processed on a single-node-based system
图3.块大小，CPU处理能力和要在基于单节点的系统上处理的ECDSA签名的数量之间的关系
A single-node-based system using two current generation Intel Xeon CPUs can expect to return between 160 and 200 GFlops (LINPAC) processing power.
使用两个当代Intel Xeon CPU的基于单节点的系统可能会返回160到200 GFlops（LINPAC）处理能力。
Consequently, it is not difficult to determine that software optimized to run in parallel using a GPU (CUDA) and co-processor cards that can individually deliver up to 8 TFlops can scale.
因此，确定使用GPU（CUDA）和能够单独提供高达8 TFlops的协处理器卡进行并行运行的软件可以进行扩展并不困难。
A cluster of the systems used for Fig. 2(d) can scale to allow unlimited block size.
图2（d）所用的系统集群可以扩展以允许无限的块大小。
5.2. Block size increases and storage
5.2。块大小增加和存储
The resulting storage problem will be addressed in a complimentary paper detailing a form of validated storage node.
由此产生的存储问题将在免费论文中详细说明一个经过验证的存储节点的形式。
By assigning special tasks to individual functions of the Bitcoin nodes, we can allow for the creation of a market-based solution to the issue of transaction costs and spam.
通过将特殊任务分配给比特币节点的各个功能，我们可以允许创建基于市场的解决方案来处理交易成本和垃圾邮件。
The maximum storage required for the Blockchain can be calculated and is a linear relationship over time.
区块链所需的最大存储空间可以计算出来，并且随着时间的推移呈线性关系。
Fig. 4 displays the required storage for a 10-year period.
图4显示了10年期间所需的存储空间。
Here we see that even using a 1,000 MB block size maximum limit, the maximum limit for the storage required is calculated to be 120 TB in 2025.
在这里我们看到，即使使用1,000 MB块大小的最大限制，在2025年所需的存储的最大限制也被计算为120TB。
The rate at which the size of hard drives has been increasing has remained steady and is predicted to grow at a similar rate for at least the next 10 years11.
硬盘大小的增长速度一直保持稳定，预计至少在未来10年内将以类似速度增长。
The primary misunderstanding of the growth capabilities in the Blockchain does not originate from scientific analysis, but from the failure to understand exponential growth.
对区块链增长能力的主要误解不是来源于科学分析，而是来自对指数增长理解的失败。
The Blockchain is a system that, when limited, can grow at a maximum linear rate. Hard drives and systems processing grow exponentially.
区块链是一个有限的系统，可以以最大的线性速度增长。硬盘和系统处理量呈指数增长。
The consequence is that systems that are considered to be large now quickly become the norm.
其结果是被认为是大的系统现在很快成为常态。
Fig. 4. Data storage size as a function of block size
图4.数据存储大小作为块大小的函数
The annual storage requirements for a block size of up to 100 Mb (Fig. 5) are capped at 1.4 TB.
一个块大小为100Mb（图5）的年度存储要求被限制在1.4TB。
This is well within the limits a modern computer system is able to contain.
这完全在现代计算机系统能够包含的范围之内。
Fig. 5. Data storage size as a function of block size
图5.数据存储大小作为块大小的函数
Given that systems growth is exponential (Fig. 6) and the Bitcoin system is limited to a linear growth rate, it is simple to see that the system can grow at a far higher rate than has been proposed.
鉴于系统增长呈指数级增长（图6），而比特币系统仅限于线性增长率，很容易看出系统的增长速度远远高于已经提出的速度。
Fig. 6. Data storage size as a function of block size
图6.数据存储大小作为块大小的函数
The result is that drive space is not a limiting factor for increasing the size of the Blockchain.
结果是驱动器空间不是增加区块链大小的限制因素。
Even with a sustained block size of 5,000 Mb, the total storage capacity of the Blockchain would not exceed the predicted storage capacity of a single server.
即使拥有5,000 Mb的持续块大小，区块链的总存储容量也不会超过单个服务器的预测存储容量。
With the introduction of a dedicated storage node, the growth of the system can be extended to a 300-GB block size and beyond.
随着专用存储节点的引入，系统的增长可以扩展到300GB以上的块大小。
This eventuality was foreseen by Nakamoto (2008), as mentioned in section 7 of his paper.
Nakamoto（2008）已经预见到了这种可能性，正如他在论文第七部分提到的那样。
Here, most nodes would operate with a limited subset of the Blockchain.
在这里，大多数节点将使用区块链的有限子集进行操作。
In this scenario, specialized nodes that act as a store of the complete Blockchain can be referenced with only the Merkle hash necessary for the payment nodes.
在这种情况下，充当完整区块链存储的专用节点只能使用支付节点所需的Merkle散列进行引用。
5.3 Fast payment network
5.3快速支付网络
Fig. 7. Latency against bandwidth, existing network
图7.现有网络对带宽的延迟
Fig. 7 shows the latency measured against the bandwidth for the existing network.
图7显示了针对现有网络的带宽测量的延迟。
Fig. 8 shows a plot of the latency against the continuous bandwidth needed for increased transaction sizes.
图8显示了对于增加事务大小所需的连续带宽的等待时间图。
In this plot, we have also incorporated the error range for each of the measurements.
在这个图中，我们也包含了每个测量的误差范围。
Fig. 9 plots the expected propagation delay in TPS against the existing network.
图9绘制了TPS对现有网络的预期传播延迟。
This system has been modeled utilizing standard commercial hardware.
这个系统已经使用标准的商业硬件来建模。
The system used is Intel i7 with 32 GB of RAM and an SSD consisting of 100 IOPs.
使用的系统是带有32 GB RAM的Intel i7和由100 IOP组成的SSD。
Fig. 10 shows the propagation delays on the simulations using specialized nodes.
图10显示了使用专用节点的仿真的传播延迟。
We can clearly see that dividing the node functionality increases the effectiveness of the network significantly.
我们可以清楚地看到，划分节点功能显着增加了网络的有效性。
Fig. 8. Latency against bandwidth for the proposed network of specialized nodes
图8.建议的专用节点网络对带宽的延迟
Fig. 9. Propagation delay as a function of transactions per second (TPS)for the existing network
图9.传播延迟作为现有网络每秒事务数（TPS）的函数
Fig. 10. Propagation delay as a function of transactions per second (TPS) for the proposed network of specialized nodes
图10.传播延迟作为建议的专用节点网络的每秒事务数（TPS）的函数
For an attack to be successful an attacker would have to connect to several nodes in a network and broadcast the double-spend transaction claiming the outputs quickly such that the miner is more likely to receive this transaction ahead of the legitimate one and process it.
要使攻击成功，攻击者必须连接到网络中的多个节点，并快速广播声明输出的双重支出事务，以便矿工更有可能在合法的事务之前接收此事务并处理它。
The attacker cannot inhibit the communication between nodes and will not be able to identify the neighbors of adjoining nodes.
攻击者不能阻止节点之间的通信，不能识别邻接节点的邻居。
The attacker would not be able to utilize the paid merchant transmission nodes and if an attacker was to succeed their information could be recorded, in which case such a network would be able to instantiate an assurance protocol.
攻击者将无法利用付费商户传输节点，如果攻击者能够成功记录他们的信息，在这种情况下，这样的网络将能够实例化保证协议。
This means that anyone propagating over the network would have to present an assurance contract of Bitcoin such that any double spend or attack would violate the conditions allowing for the payment of an assurance contract exceeding the covering of a double spend or another such attack.
这意味着任何通过网络传播的人都必须提交一份比特币的保证合同，以便任何双重支出或攻击都会违反允许支付保证合同超过双重支付或其他此类攻击的条件。
In our simulations of FPNs, we included nodes that incorporate a monitoring function for double-spend attacks that would alert individual merchants to attacks of this nature.
在我们对FPNs的模拟中，我们包含了包含双重攻击的监控功能的节点，这个功能会提醒个体商户这种攻击。
We showed that such a network could interact within milliseconds allowing a propagation that has occurred across the network to be broadcast faster than the gossip network broadcast between miners and other instantiated instances of the Bitcoin network.
我们展示了这样一个网络可以在几毫秒内相互作用，从而允许在网络上传播的广播比在矿工和其他实例化的比特币网络之间传播的八卦网络更快地传播。
This distributed autonomous corporation (DAC) would then be able to alert merchants of a double spend that had occurred anywhere on the Bitcoin network and allow the merchant to reject the transaction.
然后，这个分布式自治公司（DAC）将能够向商家发出在比特币网络上任何地方发生的双重花费的提醒，并允许商家拒绝交易。
This rejection could happen within seconds, and we show that the creation of a suitable merchant network for nodes would enable a fast spend network that could detect double spending and reject this in a period of less than 5 s.
这种拒绝可能会在几秒钟内发生，而且我们表明，为节点创建合适的商户网络将使快速支出网络能够检测到双倍支出，并在不到5秒的时间内拒绝。
In such a system, we believe that it would be possible to reject small transaction double spends quickly enough that even where a vending machine system has been deployed, goods could be re-routed rather than going to a customer system if a double spend had occurred.
在这样的系统中，我们认为有可能快速地拒绝小额交易双倍花费，即使在已经部署了自动售货机系统的情况下，如果发生双重花费，货物可以被重新路由而不是转到客户系统。
This leads to a truly extensive payment system that could be used globally.
这导致了可以在全球范围内使用的真正广泛的支付系统。
Even with pseudonymous accounts it is possible to utilize the identification of master accounts, thereby allowing merchants to incorporate knowledge of clients that would allow them to verify information over time and quickly transfer value to trusted individuals.
即使使用假名账户，也可以利用主账户的识别，从而允许商家将客户的知识结合起来，使他们能够验证信息，并迅速将价值转移给受信任的个人。
Even without this trust, the integration of propagating nodes would allow clients and merchants to quickly propagate information to miners in a manner that would allow payment with a low probability of double spend.
即使没有这种信任，传播节点的集成将允许客户和商人以允许以低花费的低概率支付的方式将信息快速传播给矿工。
This probability could be insured and assured with successful node integration by providing a network of payment authorities that could be integrated into a DAC.
通过提供一个可以集成到DAC中的支付机构网络，成功的节点集成可以保证这个可能性。
In this system, a user selecting an item would, after payment, wait for a small amount of time such as 4-5 s, and, if no double spend is detected in that period, the goods would be delivered.
在这个系统中，用户选择一个项目后，等待一段时间，如4-5秒，如果在这段时间内没有发现双重消费，货物将被交付。
In this manner, a merchant accepting fast payment can act without incurring the risk that a transaction will not be confirmed by the network and can reduce the risk of an economic loss.
以这种方式，快速付款的商户可以在不发生交易不被网络确认的风险的情况下行事，并且可以降低经济损失的风险。
In case an economic loss does occur this can be incorporated into the price of a sale over time and the loss will in any event be minimized with information about the user being recorded.
如果发生经济损失，可以将其纳入销售价格中，在任何情况下都可以通过记录用户信息来减少损失。
An FPN should be secured on a risk basis: the likelihood of smaller payments being sent quickly without the risk of double spend should be higher due to the limited loss that can be associated with each of these transactions.
FPN应以风险为基础进行担保：由于可能与这些交易相关的损失有限，因此较小的付款被快速发送而没有双重支出风险的可能性应该较高。
We showed that a variety of different nodes (verification node, a transmission node, a node that stores the full Blockchain, and a propagation node) can be integrated to extend the current node system without modification of the Bitcoin protocol.
我们展示了可以集成各种不同的节点（验证节点，传输节点，存储完整区块链的节点以及传播节点）来扩展当前节点系统，而无需修改比特币协议。
Merchants selectively using nodes could use the core Bitcoin protocol with the addNode function.
有选择地使用节点的商家可以使用具有addNode功能的核心Bitcoin协议。
A future implementation of this system would enable merchants to pay for preferred nodes that could propagate information around the network with a guaranteed propagation time.
这个系统的未来实现将使商家能够支付能够以确保的传播时间在网络周围传播信息的优选节点。
In this system, information eclipsing is minimized where the merchant forwards selectively to neighboring nodes that will accept their transaction rather than any attack transaction, i.e., when a node has received a double spend from a trusted merchant it will drop their information and alert the merchant to the scenario.
在这个系统中，当商家有选择性地转发到相邻节点而不是任何攻击事务，即当一个节点从一个信任的商家那里收到双倍支出时，它将放弃他们的信息并提醒商家到场景。
In general information eclipsing, when the merchant forwards Transaction A to its validated nodes these will propagate between the network quickly over a fast back channel which will then propagate through the remainder of the Bitcoin network (Decker and Wattenhofer, 2013).
一般情况下，当商家将交易A转交给经过验证的节点时，这些信息将迅速通过快速返回信道在网络之间传播，然后通过比特币网络的其余部分（Decker and Wattenhofer，2013）进行传播。
This implementation of a modified SIS epidemic network infrastructure (Pastor-Satorras and Vespignani, 2001) allows a merchant to quickly forward their transaction as it is received at many nodes.
经过修改的SIS流行网络基础设施（Pastor-Satorras and Vespignani，2001）的实施允许商家在多个节点处接收它们时快速地转发它们的交易。
In this instantiation nodes can be autonomously verified and checked to ensure that any information sent to them will be expected to respond quickly and securely around the network.
在这个实例中，节点可以被自主地验证和检查，以确保发送给它们的任何信息将被期望在网络周围快速和安全地作出响应。
Random checks by third parties should be conducted on any such network to ensure the honesty of any DAC in this system.
应该在任何这样的网络上进行第三方的随机检查，以确保该系统中的任何DAC的诚实。
This would be quickly and easily verified and it would be expected that any DAC participating in such a system would lose merchants and clients extremely quickly if it did not propagate in the manner advertised.
这将被快速而容易地验证，任何参与这种系统的发展援助委员会如果不以宣传的方式进行宣传，就会非常迅速地失去商人和客户。
In this scenario, when the merchant transmits Transaction A to the connected nodes these connect across to the neighboring nodes in such a manner that they can quickly propagate a transaction around the entire Bitcoin network in a maximum of 3 and generally 2 hops.
在这种情况下，当商家将交易A传送到连接的节点时，这些连接以这样的方式连接到相邻节点，使得它们能够在整个比特币网络周围迅速传播交易，最多3次，一般2次。
This would be one hop into the merchant connected node that could be conducted in less than 3 ms in many instances and the merchant could pay for a service that provides a connectivity of less than one millisecond, the connection between verified nodes would enable the transmission of even large packets to occur within less than 5 ms to the primary wired nodes.
这将是商户连接节点的一跳，在许多情况下可以在小于3ms的时间内进行，并且商家可以支付提供小于1毫秒的连接的服务，验证的节点之间的连接将使得传输即使是在主要有线节点少于5毫秒内发生的大数据包也是如此。
These wired nodes could propagate up to 10,000 nodes within 7-8 ms and in an ideal fast network scenario this could be conducted globally in fewer than 4 ms in urban centers.
这些有线节点可以在7-8毫秒内传播多达10,000个节点，并且在理想的快速网络情况下，在城市中心可以在不到4毫秒的时间内在全球进行。
In this scenario, the attacker sending Transaction B would be expected to require their transaction to be sent to many miners faster than the merchant network.
在这种情况下，发送交易B的攻击者预计将要求他们的交易比商家网络更快地发送给许多矿工。
Here, the verified node would monitor for the merchant any conflicting transactions such as Transaction B and where any transaction such as Transaction B occurred they would know of the double-spending attempt.
在此，经验证的节点将监视商家的任何冲突交易，例如交易B，以及诸如交易B的交易发生在哪里，他们将知道双重消费尝试。
The modification does not require any change to the existing Bitcoin protocol and can be implemented by nodes with slight protocol extensions without any other node being aware that these are merchant nodes.
该修改不需要对现有比特币协议进行任何改变，并且可以由具有轻微协议扩展的节点来实现，而没有任何其他节点知道这些是商家节点。
Using this system, the merchant nodes could implement a distributed update system in which they could capture at least 90-95% of nodes connected to the network at any time and could potentially monitor 100% of nodes connected to the Bitcoin network.
使用这个系统，商家节点可以实现一个分布式更新系统，在这个系统中，他们可以随时捕获连接到网络的至少90-95％的节点，并且可能监视连接到比特币网络的100％的节点。
In doing so, any attacker sending an invalid transaction would reach the merchant within 2 hops, any transaction received by a node would propagate into the merchant network in 1 hop, and the merchant network would instantaneously respond by transmitting a message to the merchant.
在这种情况下，发送无效事务的任何攻击者将在2跳内到达商户，由节点接收的任何事务将以1跳传播到商户网络，并且商家网络将通过向商家发送消息而立即作出响应。
In this scenario, the merchant would not accept any transactions and would invalidate Transaction A, thereby stopping the transmission of the goods and record information about the transaction.
在这种情况下，商户不会接受任何交易，并会使交易A失效，从而停止货物的传输并记录交易信息。
Transaction B would be broadcast to the merchant network, the merchant network would then record information about the double spend ensuring that the information on each of the double-spend attacks was recorded and monitored over time and that information about the double spend could be used to build up a profile of any such attackers.
交易B将被广播给商家网络，然后商家网络将记录关于双重支出的信息，确保每次双重支出攻击的信息随时间被记录和监控，并且可以使用关于双重支出的信息建立任何这样的攻击者的个人资料。
Further, if no nodes have been selectively forwarding double-spend attacks or known IPs or other such source details have been recorded, it would be possible for information to be collected and stored about known attackers that could be used to profile them over time by rejecting any further information.
此外，如果没有节点已经有选择地转发双花攻击或已知IP或已经记录了其他这样的源细节，则有可能收集和存储关于已知的攻击者的信息，这些已知的攻击者可以通过拒绝更多的信息。
Merchant nodes could instantly block any known attacker by updating information such as IP table filters to reject IP addresses or other identifying information in such a manner that known attackers would be unlikely to propagate information across the network.
商家节点可以通过更新诸如IP表格过滤器之类的信息来立即阻止任何已知的攻击者，从而以这样的方式拒绝IP地址或其他识别信息，即已知的攻击者不可能通过网络传播信息。
The ability to randomize merchant connectivity in a manner similar to that over our CORE simulations means that it is possible to quickly simulate alternative network configurations using a virtualized node system that would thwart any attempt by an attacker to block such information.
以类似于我们的CORE仿真的方式随机化商户连接的能力意味着可以使用虚拟化节点系统来快速模拟替代的网络配置，这将阻止攻击者阻止此类信息的任何企图。
The periodic randomization of the merchant network in terms of reconstructing information, IP addresses, and node connectivity while updating this automatically between the nodes would make it computationally difficult for the attacker to propagate any information and would minimize the risk of such an attack.
在重建信息，IP地址和节点连接性方面，商家网络的周期性随机化，同时在节点之间自动更新这将使得攻击者在计算上难以传播任何信息，并且将这种攻击的风险最小化。
The ability to randomize connectivity and map over 90% and sometimes 100% of the node network allows the merchant network to reduce timing attacks and monitor for attacks efficiently.
随机化连接的能力和映射超过90％甚至100％的节点网络的能力允许商家网络减少计时攻击并有效监视攻击
The merchant is thus able to examine a propagation depth that ensures up to a desired level of risk is mitigated.
商家因此能够检查传播深度，确保达到期望的风险水平被减轻。
For instance, in our propagation testing we set a Six Sigma level of risk mitigation meaning that in either 2 or 3 node hops we are able to distribute sufficient data to the Bitcoin network to ensure that any transactions received by the merchant have either not been alerted to the merchant within seconds or that the instance of propagation would require more network nodes than currently exists for the attacker to be successful.
例如，在我们的传播测试中，我们设置了六西格玛风险缓解水平，意味着在2或3个节点跳数中，我们能够将足够的数据分配给比特币网络，以确保商家收到的任何交易未被提醒在数秒内传送给商家，或者传播实例将需要比当前存在的网络节点更多的网络节点以使攻击者成功。
In a scenario such as this the Six Sigma level of risk could be taken by the merchant meaning that fewer than one in a million attacks would succeed and the merchant could minimize their risk by maintaining the level of loss against that propagation rate.
在这种情况下，商户可以采取六西格玛风险水平，即百万次攻击中有不到一次会成功，商家可以通过维持与传播速度相对的损失水平来降低风险。
It would be infeasible for the attacker to maintain knowledge of a suitable level of merchant connections to block these and yet propagate across a suitable number of other nodes in the Bitcoin network.
攻击者不可能保持适当水平的商家连接的知识来阻止这些连接，而是通过比特币网络中适当数量的其他节点传播。
Here only a small minority of nodes would receive and broadcast Transaction B into the mining network.
这里只有少数节点将接收并将事务B广播到采矿网络中。
Given this scenario, Transaction B would be received only by a small number of miners who would be less likely to put this into a block.
考虑到这种情况，只有少数矿工才能收到交易B，而这些矿工不太可能把这个交易放在一个街区。
This scenario could be further extended if mining entities were paid to reject known double spends and an assurance level was paid to protect the client from any adverse inference.
如果支付采矿实体拒绝已知的双重支出，并且支付保证水平以保护客户免受任何不利推论，那么这种情况可以进一步延长。
The implementation of an accounting system that monitors for any double spends would allow the merchant to ensure that where double spends have occurred the customer is not able to validly also seek a refund or return payment.
监控任何双重花费的会计系统的实施将允许商家确保发生双倍花费的客户不能有效地寻求退款或退还支付。
The primary aspect of this system would be chaining the payment and associated goods together.
这个系统的主要方面是将支付和相关货物连在一起。
In this way, the exchange of a tokenized item or quantity on the Blockchain network would be chained to a single transaction involving the payment from the client. In cases where sizable items or individually recorded items or even digital goods are sold, the transaction involving the transfer of the rights to the goods would be tied to the payment to the merchant, in which case a double-spend attack would lead to the rejection of the transfer of the item.
通过这种方式，区块链网络上的标记化物品或数量的交换将被链接到涉及来自客户端的支付的单个交易。如果出售大件物品或单独记录的物品甚至数字货物，涉及货物权利转移的交易将与支付给商人的交易相联系，在这种情况下，双重支出攻击会导致拒绝该项目的转移。
A simple system would involve separating payment and provision of the goods or reversing payment of the goods. Physical delivery of the goods could be quickly separated or the rights to access digital goods would be unavailable.
一个简单的系统将涉及分离付款和提供货物或逆转货物付款。货物的实物交付可能很快分开，或者访问数字货物的权利将不可用。
Where physical goods are tied to a real-world implementation such as a digital signature that unlocks a car, access to the car would not be available even on a hire contract when the transaction paying for the contract had not validly been submitted and processed over the Blockchain network.
如果实体商品与解锁车辆的数字签名等现实世界的实现相关联，则即使签订合同的交易未被有效地提交和处理，也不能在租用合同上访问汽车区块链网络。
In creating such a system, we have taken the information collected over our CORE12 network on Tulip and extended this to simulate the injection of double spends by attackers and have simulated the creation of a merchant node network.
在创建这样一个系统时，我们已经把在我们的CORE12网络上收集的有关Tulip信息的信息扩展到模拟注入攻击者的双重花费，并模拟了商家节点网络的创建。
Each merchant network would compete to gain favor with merchants selectively sending to known trusted merchant networks, and this would be conducted using information such as the IP address, Alternatively, it could be conducted using a fast flux network implemented from tax scenarios that have been noticed over the malware industry in a manner that would enable the IP address and other information to be hidden but still propagated to the merchant.
每个商家网络将竞争获得有选择地发送给已知的可信赖的商家网络的商家，并且这将使用诸如IP地址之类的信息来进行。或者，可以使用从已经注意到的税收情景实现的快速通量网络在恶意软件行业，以使IP地址和其他信息被隐藏但仍然传播给商家的方式。
The node network with the addition of a fast flux network could cycle through IP addresses in a manner not allowing merchants to determine the known addresses of merchants and simultaneously propagating across the Bitcoin network in a manner that does not cause instability to the gossip protocol.
添加快速通量网络的节点网络可以通过IP地址循环，不允许商家确定商家的已知地址，同时以不会导致八卦协议不稳定的方式在比特币网络上传播。
Merchants would be able to verify the availability and rate of distribution from the merchant node network and could selectively verify the existence of the complete Blockchain through archive nodes.
商家将能够从商家节点网络验证分配的可用性和速率，并且可以通过存档节点选择性地验证完整区块链的存在。
Merchants could pay for an enhanced service allowing both the use of merchant nodes for free commerce at a limited rate or for complete merchant propagation at a high rate for a pre-determined fee.
商家可以支付增强的服务，允许以有限的速率使用商家节点进行自由商业，或者以高的费率完成商家传播，达到预定的费用。
Contracts could be built into the Bitcoin protocol allowing the merchant to both pay and check using a smart contract that pays for the utilization of the network ensuring the security of the node network and the propagation for merchants.
合同可以嵌入到比特币协议中，允许商家使用支付网络使用的智能合同进行支付和检查，确保节点网络的安全性和商家的传播。
Such a network would ensure that nodes were maintained and available and the localization of selective node networks could be integrated into collections of global node networks allowing for both the propagation of distributed node organizations and for the reporting and availability of all this information.
这样的网络将确保节点维护和可用，并且选择性节点网络的本地化可以被集成到全局节点网络的集合中，从而允许分布式节点组织的传播以及所有这些信息的报告和可用性。
Each node in the merchant node system would need to have sufficient bandwidth to receive connections from random inputs as well as connections from the paid merchant system.
商户节点系统中的每个节点将需要足够的带宽来接收来自随机输入的连接以及来自付费商家系统的连接。
Although merchant payment transactions could be promoted to the network faster, the node would be required to also receive transactions randomly through the rest of the network.
虽然商家支付交易可以更快地提升到网络，但是节点也需要通过网络的其余部分随机接收交易。
The node could propagate its transaction policy such that any other wallet on the network would be able to selectively add or reject merchants, but the merchant system would be able to supply bandwidth at a rate to the merchant ensuring a set level and standard of service.
该节点可以传播其交易策略，使得网络上的任何其他钱包将能够选择性地添加或拒绝商家，但商家系统将能够以一定的速率向商户提供带宽，以确保设定的水平和服务标准。
In this manner, as the double-spend attack occurs, the two conflicting transactions transmitted into the network would be quickly detected.
通过这种方式，当双重攻击发生时，传输到网络中的两个冲突交易将被快速检测到。
Large packets could be propagated through the network within seconds and detection of any double-spend transactions would happen at the same time.
大型数据包可以在几秒钟内通过网络传播，同时检测到任何双重支出交易。
The forward belief network was calculated using a Bayesian probability matrix for which we used the previous values and, given the movement over time with various scenarios, extrapolated using a back-propagation network and deep learning algorithms to reproduce the likely scenario given each of the proposed sizes of blocks etc.
前向信念网络是使用贝叶斯概率矩阵计算的，我们使用了以前的值，并且考虑到随着时间在各种情况下的移动，使用反向传播网络和深度学习算法来推断给出每个提议块的大小等
This required us to model the network sizes for each of the nodes and extrapolate the loss of nodes and the change in node types that would be necessary for a sustainable network.
这要求我们对每个节点的网络规模进行建模，并推断节点的损失和节点类型的变化，这是可持续网络所必需的。
Over time this sustainable network would change with the commercialization and professionalization of node management.
随着时间的推移，这个可持续网络将随着节点管理的商业化和专业化而改变。
The current scenario involves many amateur networks, which incorporate transient nodes that join and leave the network.
目前的情况涉及许多业余网络，其中包括加入和离开网络的瞬态节点。
These networks are less secure and flapping or transient responses can occur.
这些网络不太安全，可能会发生震荡或瞬态响应。
Each of the simulations was conducted by utilizing a reconstructed network based on the responses of the remaining nodes and an extrapolated response by an increased number of commercial nodes.
通过基于剩余节点的响应的重构网络和通过增加数量的商业节点的外推响应来进行每个模拟。