A Novel Proof-of-Reputation Consensus for Storage Allocation in Edge Blockchain Systems 精读笔记（一）

Abstract

Edge computing guides the collaborative work of widely distributed nodes with different sensing, storage, and computing resources. For example, sensor nodes collect data and then store it in storage nodes so that computing nodes can access the data when needed. In this paper, we focus on the quality of service (QoS) in storage allocation in edge networks. We design a reputation mechanism for nodes in edge networks, which enables interactive nodes to evaluate the quality of services for reference. Each node publicly broadcasts a personal reputation list to evaluate all other nodes, and each node can calculate the global reputation of all nodes by aggregating personal reputations. We then propose a storage allocation algorithm that stores data to appropriate locations. The algorithm considers fairness, efficiency, and reliability which is derived from reputations. We build a novel Proof-of-Reputation (PoR) blockchain to support consensus on the reputation mechanism and storage allocation. The PoR blockchain ensures safety performance, saves computing resources, and avoids centralization. Extensive simulation results show our proposed algorithm is fair, efficient, and reliable. The results also show that in the presence of attackers, the success rate of honest nodes accessing data can reach 99.9%.

摘要——边缘计算指导协同工作具有不同传感、存储和计算资源。例如，传感器节点收集数据并然后将其存储在存储节点中，以便计算节点可以访问需要时提供数据。在本文中，我们关注的是质量边缘网络存储分配中的服务（QoS）。我们设计了一个边缘网络中节点的信誉机制，这使得交互节点评估服务质量以供参考。每个节点公开广播个人信誉列表给评估所有其他节点，每个节点都可以计算全局通过聚合个人声誉来获得所有节点的声誉。然后我们提出了一种存储分配算法，能将数据存在适当的位置。该算法考虑了公平性，效率和可靠性源于声誉。我们建立一个新颖的信誉证明（PoR）区块链来支持关于信誉机制和存储分配的共识。PoR区块链确保安全性能，节省计算资源，避免中心化。广泛的模拟结果表明我们提出的算法是公平、高效和可靠的。这结果还表明，在存在攻击者的情况下，成功诚实节点访问数据率可达99.9%。

Summarize

1. Target：质量边缘网络存储分配中的服务（QoS）
2. Work：
   • 信誉机制：每个节点公开广播个人信誉列表给评估所有其他节点，每个节点都可以计算全局通过聚合个人声誉来获得所有节点的声誉。
   • 存储分配算法：将数据存在适当的位置
   • 信誉证明（PoR）区块链：支持关于信誉机制和存储分配的共识
3. Advantage：
   • 公平性，效率和可靠性源于声誉。
   • PoR区块链确保安全性能，节省计算资源，避免中心化。

INTRODUCTION

The exponentially growing edge nodes, such as Internet of- Things (IoT) sensors, smartphones, even vehicles, generate data and create value anytime and anywhere. Under the condition that a single device cannot complete the ever-increasing large-scale tasks, emerging edge computing has promoted the cooperation of edge devices with different capabilities all over the world.

The storage allocation is an important consideration in the research and application of edge computing [1]. Consider the following scenarios that edge devices storage and access data. Many advanced sensors can produce extremely large and continuous data streams [2]. Short video applications such as TikTok allow users to share their daily lives by recording and downloading video clips [3]. Vehicles can access maps and real-time traffic flow through cellular networks to plan driving
routes [4]. These scenarios require servers to store large amounts of data generated in the network and quickly access these data when needed by terminals. To resist disagreements and adversaries, some traditional centralized solutions need basic trust like trusted third-party (e.g. certificate authority) or transitive trust assumption. The limitations of these centralized solutions that rely on trust parties are significant and inevitable,such as efficiency problems caused by disconnection and privacy problems caused by the trust. To avoid the issues caused by centralization, decentralized solutions are introduced to edge computing environments.

As a distributed ledger, blockchain is the most prevalent decentralized system in Peer-to-Peer (P2P) networks. Recently, the blockchain is introduced as a safe and effective technology for storage allocation in edge networks. During the storing process, the edge device obtains a piece of new data and stores it in other edge devices or servers in the edge network, and stores the storage location on the blockchain. During the accessing process, the edge device queries the blockchain to obtain storage locations of the data and directly selects the appropriate location to request. Compare to centralized solutions, the blockchain solution guarantees that all the content can be accessed even when some nodes are offline, and the query will only send to the selected location without revealing privacy.

Based on current solutions to resource allocation in edge network environments, we find that two points are often ignored. First, current blockchain consensus mechanisms have limitations. The classic consensus mechanism of blockchain is Proof-of-Work (PoW), which consumes a large number of computing resources. Since most edge devices have limited computing power, it is impossible for them to consume a lot of computing power to generate blocks. Some work [5], [6] apply Proof-of-Stake (PoS) instead. The basic idea of the PoS is that nodes with more tokens generate blocks with higher priority. A problem with the PoS is that it often leads to unavoidable centralization. The reason is that block generators usually have more tokens, and they receive reward tokens as blockgenerators, which makes them more likely to become block generators in the future, leading to centralization. Therefore, we need to design a consensus mechanism to avoid the limitations of PoW and PoS. Second, network participants have different reliability, and they may not provide the services as expected. Generally speaking, if a node provides more incomplete data, the peer interacting with the node considers the node to be less reliable. A reputation mechanism is an appropriate solution to the reliability of network participants [7]. We customize a reputation mechanism according to our environment. We are facing the following challenges to overcome the storage allocation in edge networks, 1) how to reach a decentralized consensus on the blockchain through limited resources, 2) how to build an appropriate reputation mechanism for nodes to evaluate the trust of others, and 3) how to consider fairness, efficiency, and reliability for storage allocation in edge networks.

In this paper, we first design a reputation mechanism in edge networks. Every node in the network shares the personal reputation of others. To reach a consensus, every node obtains the same global reputation from the same personal reputation. We then propose a storage allocation algorithm that considers fairness, efficiency, and reputation. Once a user generates a piece of data, the user stores the data to locations derived by the algorithm. With the help of our reputation mechanism, we build a novel Proof-of-Reputation (PoR) blockchain to maintain the reputation and storage allocation information. The basic idea of PoR is to let the node with the most increase of the global reputation value in each block be the generator. The PoR blockchain avoids limitations of PoW and PoS mechanisms. Extensive simulations show that our proposed system ensures a high success rate of accessing data when keeping fair and efficient storage allocation.

Our main contributions are summarized as follows.
We design a reputation mechanism for nodes in edge networks to evaluate each other. The personal reputations are generated by each node, and the global reputations are obtained by aggregating personal reputations.
Considering fairness, efficiency, and reliability, we propose a storage allocation algorithm. Nodes select storage locations through the algorithm to store newly generated data.
Based on the reputation mechanism, we build a PoR blockchain to maintain the related information of storage allocation. The node with the most increase of global reputation value in each block is selected as the generator. The PoR blockchain satisfies low cost and decentralization.
We conduct simulations in edge networks to evaluate the performance. The simulation results show that our PoR blockchain ensures decentralization, and most block generators are honest. Compared with previous work, our algorithm improves the success rate of accessing data, reaching close to 99.9% in our simulations. The simulation results also show that our storage allocation algorithm satisfies fairness, efficiency, and reliability requirements. The rest of this paper is organized as follows. Section II briefly reviews the related work. In Section III, we formulate the problem and propose threat models. In Section IV, we introduce the reputation mechanism. In Section V, we propose the storage allocation algorithm. In Section VI, we propose the PoR blockchain structure. Simulation results are presented in Section VII. The last section concludes this paper.

物联网 (IoT) 传感器、智能手机甚至车辆等呈指数级增长的边缘节点随时随地生成数据并创造价值。在单台设备无法完成日益增多的大规模任务的情况下，新兴的边缘计算推动了各地不同能力边缘设备的协同合作世界。

存储分配是边缘计算研究和应用中的一个重要考虑因素[1]。考虑以下边缘设备存储和访问数据的场景。许多先进的传感器可以产生极大且连续的数据流 [2]。 TikTok等短视频应用程序允许用户通过录制和下载视频片段来分享他们的日常生活[3]。车辆可以通过蜂窝网络访问地图和实时交通流来规划驾驶路线 [4]。这些场景需要服务器存储网络中产生的大量数据，并在终端需要时快速访问这些数据。为了抵制分歧和对手，一些传统的集中式解决方案需要基本信任，例如受信任的第三方（例如证书颁发机构）或传递信任假设。这些依赖信任方的集中式解决方案的局限性是显着且不可避免的，比如断线带来的效率问题和信任带来的隐私问题。为了避免中心化带来的问题，边缘计算环境引入了去中心化的解决方案。

作为分布式账本，区块链是点对点（P2P）网络中最流行的去中心化系统。最近，区块链作为一种安全有效的边缘网络存储分配技术被引入。在存储过程中，边缘设备获取一条新数据并存储在边缘网络中的其他边缘设备或服务器中，并将存储位置存储在区块链上。在访问过程中，边缘设备查询区块链以获取数据的存储位置，并直接选择合适的位置进行请求。与中心化解决方案相比，区块链解决方案保证即使在某些节点离线的情况下也可以访问所有内容，并且查询只会发送到选定的位置而不会泄露隐私。

基于当前边缘网络环境中资源分配的解决方案，我们发现有两点经常被忽略。首先，当前的区块链共识机制存在局限性。区块链经典共识机制是工作量证明（PoW），它消耗大量的计算资源。由于大多数边缘设备的计算能力有限，它们不可能消耗大量的计算能力来生成块。一些工作 [5]、[6] 应用了权益证明 (PoS)。 PoS 的基本思想是拥有更多代币的节点生成具有更高优先级的块。 PoS 的一个问题是它经常导致不可避免的中心化。原因是区块生成者通常拥有更多的代币，并且他们作为区块生成者获得奖励代币，这使得它们在未来更有可能成为区块生成者，从而导致中心化。因此，我们需要设计一种共识机制来规避 PoW 和 PoS 的限制。其次，网络参与者的可靠性不同，他们可能无法按预期提供服务。一般来说，如果一个节点提供的数据不完整，则与该节点交互的对等方认为该节点的可靠性较低。信誉机制是网络参与者可靠性的适当解决方案[7]。我们根据我们的环境定制一个声誉机制。我们面临以下挑战来克服边缘网络中的存储分配，1）如何通过有限的资源在区块链上达成去中心化共识，2）如何为节点建立适当的信誉机制来评估他人的信任，以及 3）如何考虑边缘网络中存储分配的公平性、效率和可靠性。

在本文中，我们首先设计了边缘网络中的信誉机制。网络中的每个节点都共享他人的个人声誉。为了达成共识，每个节点都从相同的个人声誉中获得相同的全球声誉。然后，我们提出了一种考虑公平、效率和声誉的存储分配算法。一旦用户生成了一条数据，用户将数据存储到由算法决定的地方。在我们的信誉机制的帮助下，我们构建了一个新颖的信誉证明（PoR）区块链来维护信誉和存储分配信息。 PoR 的基本思想是让每个区块中全局信誉值增加最多的节点作为生成者。 PoR 区块链避免了 PoW 和 PoS 机制的限制。广泛的模拟表明，我们提出的系统在保持公平和高效的存储分配的同时确保了访问数据的高成功率。

我们的主要贡献总结如下。

• 我们为边缘网络中的节点设计了一种信誉机制来相互评估。个人声誉由每个节点生成，全局声誉通过个人声誉聚合得到。
• 考虑到公平性、效率和可靠性，我们提出了一种存储分配算法。节点通过算法选择存储位置来存储新生成的数据。
• 基于信誉机制，我们构建了一个PoR区块链来维护存储分配的相关信息。选择每个区块中全局信誉值增加最多的节点作为生成器。 PoR 区块链满足低成本和去中心化的要求。
• 我们在边缘网络中进行模拟以评估性能。模拟结果表明，我们的 PoR 区块链确保了去中心化，并且大多数区块生成者都是诚实的。与之前的工作相比，我们的算法提高了访问数据的成功率，在我们的模拟中达到了接近 99.9%。仿真结果还表明，我们的存储分配算法满足公平性、效率和可靠性要求。

本文的其余部分安排如下。第二节简要回顾了相关工作。在第三节中，我们制定了问题并提出了威胁模型。在第四节中，我们介绍了信誉机制。在第五节中，我们提出了存储分配算法。在第六节中，我们提出了 PoR 区块链结构。模拟结果在第七节中给出。最后一节是本文的总结。

Summarize

存储分配是边缘计算研究和应用中的一个重要考虑因素

传统的集中式解决方案需要基本信任，例如受信任的第三方（例如证书颁发机构）或传递信任假设。这些依赖信任方的集中式解决方案的局限性是显着且不可避免的，比如断线带来的效率问题和信任带来的隐私问题。

为了避免中心化带来的问题，边缘计算环境引入了去中心化的解决方案。与中心化解决方案相比，区块链解决方案保证即使在某些节点离线的情况下也可以访问所有内容，并且查询只会发送到选定的位置而不会泄露隐私。

基于当前边缘网络环境中资源分配的解决方案，两点经常被忽略。

1. 当前的区块链共识机制存在局限性。

   • 区块链经典共识机制是工作量证明（PoW），它消耗大量的计算资源。
   • PoS 的基本思想是拥有更多代币的节点生成具有更高优先级的块。 PoS 的一个问题是它经常导致不可避免的中心化。

2. 其次，网络参与者的可靠性不同，他们可能无法按预期提供服务。

主要贡献总结。

• 为边缘网络中的节点设计了一种信誉机制来相互评估。个人声誉由每个节点生成，全局声誉通过个人声誉聚合得到。
• 考虑到公平性、效率和可靠性，提出了一种存储分配算法。节点通过算法选择存储位置来存储新生成的数据。
• 基于信誉机制，构建了一个PoR区块链来维护存储分配的相关信息。选择每个区块中全局信誉值增加最多的节点作为生成器。 PoR 区块链满足低成本（PoW缺陷）和去中心化（PoS缺陷）的要求。
• 在边缘网络中进行模拟以评估性能。模拟结果表明，PoR确保了去中心化，大多数区块生成者都是诚实的。提高了访问数据的成功率，满足公平性、效率和可靠性要求。

RELATED WORK

Bitcoin was invented in 2008 by Satoshi Nakamoto [8]and has been one of the most popular P2P applications. The related technology blockchain is composed of a chain of blocks, and the blockchain records transactions generated by P2P network participants as a distributed consensus ledger. According to whether a node can join the blockchain freely, blockchains can be in two forms, permissionless blockchain and permissioned blockchain. Swanson et al. [9] discussed the main differences between permissioned and permissionless blockchains. Briefly speaking, blockchain was first designed for anonymous cryptocurrency transactions in untrusted environments, and nodes can join a permissionless network freely. There are various permissionless blockchain systems, such as Bitcoin and Ethereum [10]. Permissioned blockchains have authorized identities, and different participants have different access-control authorizations [11], thus a node cannot join a permissioned blockchains freely since a new node is unauthorized. Hyperledger Fabric [12] and Corda [13] are examples of permissioned blockchain applications.

比特币由中本聪 [8] 于 2008 年发明，一直是最受欢迎的 P2P 应用程序之一。相关技术区块链由区块链组成，区块链将P2P网络参与者产生的交易记录为分布式共识账本。根据节点是否可以自由加入区块链，区块链可以分为无许可区块链和许可区块链两种形式。斯旺森等人。 [9] 讨论了许可和无许可区块链之间的主要区别。简而言之，区块链最初是为不受信任环境中的匿名加密货币交易而设计的，节点可以自由加入一个无许可的网络。有各种无需许可的区块链系统，例如比特币和以太坊 [10]。许可区块链具有授权身份，不同的参与者具有不同的访问控制权限[11]，因此一个节点不能自由加入许可区块链，因为新节点是未经授权的。 Hyperledger Fabric [12] 和 Corda [13] 是许可区块链应用程序的示例。

The advantages of PoR, including low energy consumption and safety performance, make it an emerging blockchain consensus mechanism. PoR blockchain applications covering a variety of scenarios are rising. Gai et al. [14] presented PoR, the reputation serves as the incentive for both good behavior and block publication. Based on PoR, Yu et al. invented RepuCoin [15] with better attack resistance and higher throughput compared to PoW blockchains. Oliveira et al. [16] developed PoR and proposed an advanced consensus for private blockchain systems. Wang et al. [17] implemented a reputation incentive scheme for blockchain consensus of Industrial Internet of Things (IIoT).

PoR 的低能耗和安全性能等优势使其成为新兴的区块链共识机制。 覆盖多种场景的 PoR 区块链应用正在兴起。

• Gai等人。 [14] 提出 PoR，声誉作为良好行为和阻止发布的激励。
• 基于 PoR，Yu 等人。 发明了 RepuCoin [15]，与 PoW 区块链相比，具有更好的抗攻击性和更高的吞吐量。
• 奥利维拉等人。 [16] 开发了 PoR 并提出了私有区块链系统的高级共识。
• 王等人。 [17]为工业物联网（IIoT）的区块链共识实施了声誉激励计划。

Blockchain techniques provide secure and decentralized applications in edge computing environments. To empower resource trading in mobile edge computing networks, Qiao
et al. [18] applied blockchain to manage resource trading and task assignment. It combines a third-party trust center server and blockchain ledger to manage activities reliably. Huang
et al. [19] proposed a consensus resource allocation system in pervasive edge computing environments. This work fairly and efficiently allocates storage resources and applies PoS to save energy consumption. As practical applications, edge computing and blockchain have become common tools for Internet-of-Vehicles (IoV) [20], [21].

区块链技术在边缘计算环境中提供安全和分散的应用程序。 为了赋能移动边缘计算网络中的资源交

• 易，乔等。 [18] 应用区块链管理资源交易和任务分配。 它结合了第三方信任中心服务器和区块链账本来可靠地管理活动。
• 黄等。 [19] 在普遍的边缘计算环境中提出了一种共识资源分配系统。 这项工作公平有效地分配存储资源并应用 PoS 来节省能源消耗。

作为实际应用，边缘计算和区块链已成为车联网 (IoV) [20]、[21] 的常用工具。

Reputation management in different edge computing scenarios draws attention from researchers in recent years. Some work that applies reputation mechanisms to blockchain is studied based on various application scenarios. Huang et al. [22] discussed reputation management in vehicular edge computing and networks. Liu et al. [23] optimized resource allocation in blockchain-based video streaming systems with mobile edge computing. With the help of reinforcement learning, Xiao et al. [24] utilized a blockchain-based trust mechanism to resist attacks in edge networks. In edge computing and networks, the trust environment, conditional restrictions, and environmental requirements are highly customized, thus we can see a variety of different reputation mechanisms and blockchain designs.

近年来，不同边缘计算场景中的声誉管理引起了研究人员的关注。 基于各种应用场景研究了一些将信誉机制应用于区块链的工作。

• 黄等人。 [22] 讨论了车辆边缘计算和网络中的声誉管理。

• 刘等人。 [23] 使用移动边缘计算优化基于区块链的视频流系统中的资源分配。

• 在强化学习的帮助下，肖等人。 [24] 利用基于区块链的信任机制来抵抗边缘网络中的攻击。

在边缘计算和网络中，信任环境、条件限制和环境要求是高度定制的，因此我们可以看到各种不同的信誉机制和区块链设计。

A. Problem Statement
In order to characterize the storage allocation model, we describe the following two operations that nodes in the network can perform. First, a node generates a new piece of data, then stores it in other nodes. Second, a node requests a specific piece of data from another node that owns the data

Our purpose is to allocate storage resources reasonably to fulfill fairness, efficiency, and reliability requirements. The fairness requirement is to balance the proportion of space
consumed by each node. The efficiency requirement is to store each piece of data in the network so that all nodes can efficiently request the data. The reliability requirement is to store data in reliable nodes. The fairness and efficiency requirements are straightforward, and we explain the reliability requirement here. Generally speaking, complete data is considered reliable, and partially or completely missing data is considered unreliable. Multiple reasons may lower the reliability of data, such as network fluctuation or malicious behavior. When accessing data, nodes that provide complete data are more reliable with a higher probability. To help nodes quantify the reliability of others, we design a reputation mechanism that enables nodes to evaluate the reputations of others. We design a structure to maintain the storage allocation and the reputation mechanism so that nodes in edge networks can query data storage and historical reputation records. We mainly focus on the following three issues.

A. 问题陈述
为了表征存储分配模型，我们描述了网络中节点可以执行的以下两种操作。首先，一个节点生成一条新数据，然后将其存储在其他节点中。其次，一个节点从拥有数据的另一个节点请求特定的数据

我们的目的是合理分配存储资源，满足公平、高效、可靠的要求。公平的要求是平衡空间的比例，每个节点消费。效率要求是将每条数据存储在网络中，以便所有节点都可以高效地请求数据。可靠性要求是将数据存储在可靠节点中。公平和效率要求很简单，我们在这里解释可靠性要求。一般来说，完整的数据被认为是可靠的，部分或完全缺失的数据被认为是不可靠的。多种原因可能会降低数据的可靠性，例如网络波动或恶意行为。

在访问数据时，提供完整数据的节点更可靠，概率更高。为了帮助节点量化他人的可靠性，我们设计了一种声誉机制，使节点能够评估他人的声誉。我们设计了一种结构来维护存储分配和信誉机制，以便边缘网络中的节点可以查询数据存储和历史信誉记录。我们主要关注以下三个问题。

1. We design a reputation mechanism for the nodes in the edge network, which enables the interacting nodes to evaluate each other for reference.
2. We propose a storage allocation system that considers fairness, efficiency, and reliability.
3. We build a structure to maintain and support consensus on the storage allocation and reputation mechanism.

1）我们为边缘网络中的节点设计了一种信誉机制，使交互的节点能够相互评估以供参考。
2）我们提出了一个考虑公平、效率和可靠性的存储分配系统。
3）我们建立了一个结构来维护和支持存储分配和声誉机制的共识。

B. Network Assumptions
The network consists of nodes with different storage resources. A node can generate data, select some nodes to store data, and request data from other nodes. We assume that all nodes have the basic computing power to generate blocks of the blockchain. We assume that nodes pay providers revenue for requested data. This not only prevents nodes from sending a large number of malicious data requests but also incentivizes nodes to provide storage resource services. The specific payment method is beyond the scope of this paper.

B. 网络假设
该网络由具有不同存储资源的节点组成。 一个节点可以生成数据，选择一些节点存储数据，向其他节点请求数据。

我们假设所有节点都具备生成区块链区块的基本计算能力。 我们假设节点向提供者 支付 请求数据的收入。 这不仅可以防止节点发送大量恶意数据请求，还可以激励节点提供存储资源服务。 具体的支付方式超出了本文的范围。

C. Threat Model
We list several possible attacks on the mechanism by malicious attackers to discuss the security performance of our system. We assume that attackers are all rational, which means they only focus on maximizing their profits, and do not attack without profit.

C、 威胁模型

我们列出了恶意攻击者可能对该机制进行的几种攻击，以讨论我们系统的安全性能。我们假设攻击者都是理性的，这意味着他们只关注利润最大化，而不会无利可图地攻击。

1. Bad-mouthing attack: Attackers use false reputation feedback to cause deviations in the reputation evaluation mechanism. Generally speaking, attackers will give high reputation ratings to their own nodes, while giving other nodes low reputation ratings.
2. Denial-of-service attack: In a general Denial-of-Service (DoS) attack, attackers dilute regular data by sending a large amount of false and interfering data. In our environment, attackers may send a large amount of feedback to interfere with reputations.
3. Sybil attack: Sybil attackers register multiple different identities and use expanded capabilities of multiple different identities to attack. In our environment, attackers may create multiple identities to participate in the reputation system to obtain profits.

1） 恶意攻击：攻击者使用虚假声誉反馈导致声誉评估机制出现偏差。一般来说，攻击者会给自己的节点高信誉评级，而给其他节点低信誉评级。

2） 拒绝服务攻击：在一般拒绝服务（DoS）攻击中，攻击者通过发送大量虚假和干扰数据来稀释常规数据。在我们的环境中，攻击者可能会发送大量反馈以干扰声誉。

3） Sybil攻击：Sybil攻击者注册多个不同身份，并使用多个不同身份的扩展功能进行攻击。在我们的环境中，攻击者可能会创建多个身份来参与声誉系统以获取利润。

For the sake of simplicity, in our following discussion, nodes are divided into honest nodes and malicious nodes. Honest nodes reply to all requested data and make true evaluations. Malicious nodes may not reply to requested data and make false evaluations. /bf In order to study the worst-case performance of our system, we assume that all malicious nodes belong to the same attacker, which means that the attacker can control all malicious nodes to maximize their total profits.

为了简单起见，在下面的讨论中，节点分为诚实节点和恶意节点。诚实节点回复所有请求的数据并进行真实的评估。恶意节点可能不会回复请求的数据并进行错误的评估。/bf为了研究我们系统的最坏情况性能，我们假设所有恶意节点都属于同一个攻击者，这意味着攻击者可以控制所有恶意节点以使其总利润最大化。

Summarize

区块链可以分为无许可区块链和许可区块链两种形式。

• 无需许可的区块链系统，例如比特币和以太坊 [10]。

• Hyperledger Fabric [12] 和 Corda [13] 是许可区块链应用程序的示例。

1. 目的：合理分配存储资源，满足公平、高效、可靠的要求。

2. 方法：为了帮助节点量化他人的可靠性，设计了一种声誉机制，使节点能够评估他人的声誉。我们设计了一种结构来维护存储分配和信誉机制，以便边缘网络中的节点可以查询数据存储和历史信誉记录。

3. 假设：

   • 所有节点都具备生成区块链区块的基本计算能力。 我们假设节点向提供者 支付 请求数据的收入。 这不仅可以防止节点发送大量恶意数据请求，还可以激励节点提供存储资源服务。
   • 假设所有恶意节点都属于同一个攻击者

4. 威胁模型：

   • 攻击者会给自己的节点高信誉评级**，**而给其他节点低信誉评级。
   • 攻击者可能会发送大量反馈以干扰声誉。
   • 攻击者可能会创建多个身份来参与声誉系统以获取利润