基于同态哈希函数的云数据完整性验证算法

云存储已经成为未来存储发展的一种趋势,但也带来新的安全挑战,如云服务提供商可能出于某种目的篡改数据。为确保云数据的完整性,提出一种基于同态哈希函数的云数据完整性检查算法。该算法在可信第三方的审计下,通过聚合多个RSA签名,对云数据进行完整性验证。为不泄露用户数据信息,采用同态线性认证与随机掩蔽技术,实现隐私保护。分析结果表明,该算法不仅能有效地抵抗服务器的恶意攻击,而且支持数据更新,与现有审计算法相比,该算法在验证过程中能减少计算代价,在批量审计中降低通信量,从而提高验证的效率。     

命名数据网络中基于Hash映射的命名检索

命名数据网络(NDN)是一种以内容为中心的新型网络架构,可有效提高网络资源的共享利用率。但与传统的IPv4、IPv6相比,NDN命名的长度更长且具有可变性,因此实现NDN中命名的快速检索对提高网络性能具有重要作用。为此,提出一种基于Hash映射的分治命名检索方法,将命名分解为组件并进行CRC32映射后分别存储在相应的Hash表中,对Hash表中的数据进行快速排序后使用二分查找定位Hash值,并利用排序后Hash表的递增数据结构进行Hash冲突的快速检测,通过对Hash值添加标志位的方法解决冲突问题。实验结果表明,相比建立命名前缀树的检索方法,该分治命名检索方法可将NDN命名的存储空间压缩近65%,并且大幅提升了检索速度。     

一种面向大规模短特征集的字符串匹配技术

面向大规模特征集的字符串匹配技术在病毒检测、内容过滤等问题上的应用愈加广泛,而短模式串一直是阻碍性能提升的重要瓶颈。针对短模式串进行分析讨论,基于跳跃算法优化,采用了动态块大小和动态Hash处理以及Hash函数设计场景化的策略,同时探讨了多核处理器与多线程设计之间的关系。实验数据证明改进的算法策略具有支撑百万级特征集字符串匹配的能力。     

一种有效的无序多图像分组及其拓扑有序化的算法

针对大规模三维重建应用中多幅无序图像的分组及其有序化问题,提出一种鲁棒的无序图像分组方法。首先,对大量无序图像提取SIFT(scale invariant feature transform)特征,接着采用DBH(dichotomy based hash)算法对图像特征集合进行快速匹配,最后应用一种新的基于图像内容的图像相似度度量准则,将无序图像分组并采用视图生成树拓扑化组内图像。实验结果表明,该算法能快速有效地对无序图像分类并拓扑化。     

混合网中的P2P资源共享机制

为减轻对等网络（P2P）应用造成的IPv4网络带宽紧张问题,提出了一种IPv4/IPv6混合网中基于层次化分布式散列表（DHT）的P2P资源共享机制. 该机制使得IPv6和IPv4节点可查询和获得彼此DHT覆盖网中的P2P资源,实现P2P应用在IPv4/IPv6之间的互操作. 仿真结果表明,该P2P资源共享机制能够适应IPv4/IPv6混合网的特性,并具备很好的扩展性和负载均衡特性.      

A Cellular Automata Based Verifiable Multi-secret Sharing Scheme Without a Trusted Dealer

—We proposed a verifiable multi-secret shar-ing scheme without a dealer. We use cellular automata, having the properties of linear computations and parallel computations, to construct our scheme. The linear com-putational property makes it possible to build a scheme without the assistance of a trusted dealer, and the parallel computational property contributes to the high efficiency of the scheme. Hash function is used to realize the verifica-tion of shares, which also makes our scheme more efficient. The security of the scheme is also analyzed.     

Key processing with control vectors

A method is presented for controlling cryptographic key usage based on control vectors. Each cryptographic key has an associated control vector that defines the permitted uses of the key within the cryptographic system. At key generation, the control vector is cryptographically coupled to the key by way of a special encryption process. Each encrypted key and control vector are stored and distributed within the cryptographic system as a single token. Decryption of a key requires respecification of the control vector. As part of the decryption process, the cryptographic hardware verifies that the requested use of the key is authorized by the control vector. This article focuses mainly on the use of control vectors in cryptosystems based on the Data Encryption Algorithm.     

基于KryptoKnight的移动用户认证协议

本文基于KryptoDnight[6],为个人通信系统设计了一轻量级认证与密钥分配协议．该协议与现有的协议相比，新协议具有很高的安全性和运行效率，特别适合用作PCS移动环境中的低层网络协议     

Hash channels

It is believed that in authentication protocols where no random numbers are involved, it is hard to introduce subliminal channels. In this paper we show that subliminal channels exist in digital signature schemes where no random numbers are used. The source where the subliminal channels exist is widely acceptable hash functions with their outputs being of some randomness. Because the subliminal channels are established by using hash functions, the subliminal channels are called hash channels in this paper. This indicates that there exist subliminal channels in any authentication protocols that use hash functions.