一种基于联盟管理的高效分布式域名系统

在域名解析系统中,下级域名的命脉被上级域名所掌握,这种中心化的管理为域名解析带来了巨大的风险。以比特币等加密货币为代表的区块链则具有去中心化的特性。随着namecoin的提出,区块链开始被应用在命名系统和域名解析的领域,之后的Blockstack和ENS都提出了去中心化命名系统的解决方案。其中,Namecoin和Blockstack采用了完全去中心化的命名管理方式,产生了域名抢占问题。因此,我们将目光转向了采用小群组投票决定域名增加和删除的联盟化管理方案。在联盟化管理方案中,比如ENS的和超级账本的均存在交易空间太大的问题,在区块链本身存储代价大的背景下,存储效率将变得低下。因此,DNS系统和区块链的结合难度很大,不仅需要保证在多变的域名存储信息中保证存储总量较小,同时还需要针对域名解析实现高效的联盟化管理,这使得至今仍未有一个令人满意的去中心化域名解析系统的解决方案。为此,我们提出了ECMDNS——一个高效的基于联盟化管理的域名解析系统,既考虑到DNS区域文件具有存储量大且变换频繁的特点,又能在完全中心化和完全去中心化之间采取折衷方案,并拥有较高的时空效率及较小的存储总量。我们通过区分链上链下的存储保证在多变的域名信息中保证存储总量较小;通过群组决策投票的方式实现联盟化管理,同时优化了Hyperledger Fabric提出的混合复制模型,将空间存储效率优化到原本的1/16。并且仅需要花费1次分布式副本同步,就可以完成一项由n名成员背书对同一域名背书的事务,并在联盟化管理的基础上实现区块链交易空间性能的高效性,从而实现整体存储效率的高效性。     

A Distributed π-Calculus with Local Areas of Communication

This paper introduces a process calculus designed to capture the phenomenon of names which are known universally but always refer to local information. Our system extends the π-calculus so that a channel name can have within its scope several disjoint local areas. Such a channel name may be used for communication within an area, it may be sent between areas, but it cannot itself be used to transmit information from one area to another. Areas are arranged in a hierarchy of levels, distinguishing for example between a single application, a machine, or a whole network. We give an operational semantics for the calculus, and develop a type system that guarantees the proper use of channels within their local areas. We illustrate with models of an internet service protocol and a pair of distributed agents.     

一种基于分布对象技术的异构网管系统集成框架

实现异构网管系统的集成化管理是当前网管领域的难点,分布计算正日益成为网络管理的一种热点技术。本文以电信网络管理的应用为背景,提出了一个基于CORBA的异构网管系统的集成框架INMS（Integrated Network Management System,简称INMS）。该结构结合了OSI系统管理对象的表达与多对象访问方面的优点和CORBA的易于编程、多个编程语言的绑定、应用的可移植性及分布性等优点,解决了多个异构网络管理系统的集成问题。     

分布式测控系统任务分类命名与定位研究

任务的命名与定位是分布式测控系统中的一项关键技术。本文基于面向对象技术，提出了一种新的任务命名法——类对象命名(Class object naming．CON)。这是一种全局的、与位置无关的命名法,有效地解决了分布式测控系统应用级的位置透明性问题。文中从任务的分类命名、任务的迁移和迁移任务的访问等三个方面详细介绍了分布式测控系统任务分类命名与定位的具体实现过程和相关算法。     

Naming in the Distributed Operating System ZGL

In this paper,the naming scheme used in the heterogeneous distributed operating system ZGL is described and some of the representative techniques utilized in current distributed operating systems are examined.It is believed that the partitioning of the name space into manyn local name spaces and one global shared name space allows the ZGL system to satisfy each workstation‘s demand for local autonomy and still be able to facilitate transparent resource sharing.By the division of the system into clusters and the use of a combined centralized-distributed naming mechanism,the system is able to avoid both the bottleneck problem caused by a single centralized name server for the whole system and the performance degradation due to a full distributed scheme.     

Scalable networked information processing environment (SNIPE)

Scalable Networked Information Processing Environment (SNIPE) is a metacomputing system that aims to provide a reliable, secure, fault-tolerant environment for long-term distributed computing applications and data stores across the global Internet. This system combines global naming and replication of both processing and data to support large-scale information processing applications leading to better availability and reliability than currently available with typical cluster computing and/or distributed computer environments. To facilitate this the system supports: distributed data collection, distributed computation, distributed control and resource management, distributed output and process migration. The underlying system supports multiple communication paths, media and routing methods to aid performance and robustness across both local and global networks. This paper details the goals, design and an initial implementation of SNIPE, and then demonstrates its usefulness in supporting a middleware project. Initial communications performance is also presented.     

Distribution and synchronisation of engineering information using active database technology

Today, work collaboration is normal practice in developing modern products. Engineering collaborative work involves a number of team members that need to share and exchange design ideas while working with engineering analysis tools such as mechanical computer aided engineering systems. This work presents the M-Sync prototype system that uses an active database approach to enable exchange of engineering information among distributed team members in a timely manner. The distributed data is fully accessible by the local member and is automatically synchronised between different places using a database management system that support event-condition-action (ECA) database rules. Only updates introduced at one location are distributed to other locations, thereby minimizing information transfer and enhancing performance. Members working at different locations can therefore work in a peer-to-peer (P2P) manner and interactively manipulate the same set of information at the same time.     

综合网络管理系统信息模型的设计

通过对一个综合性通信网的分析和研究,设计了一种实用的基于TMN标准的综合通信网管信息模型,并给出若干实际应用的范例。最后论述了应用范围和局限性。     

A security framework for protecting traffic between collaborative domains

In this paper, we propose a novel Secure Name Service (SNS) framework for enhancing the service availability between collaborative domains (e.g. extranets). The key idea is to enforce packet authentication through resource virtualization and utilize dynamic name binding to protect servers from unauthorized accesses, denial of service (DOS) and other attacks. Different from traditional static network security schemes such as VPN, the dynamic name binding of SNS allows us to actively protect critical resources through distributed filtering mechanisms built in collaborative domains. In this paper, we present the architecture of the SNS framework, the design of SNS naming scheme, and the design of authenticated packet forwarding. We have implemented the prototype of authenticated packet forwarding mechanism on Linux platforms. Our experimental results demonstrate that regular Linux platforms are sufficient to support the SNS authenticated packet forwarding for 100 Mbps and 1 Gbps Ethernet LANs. To further improve the performance and scalability, we have also designed and implemented unique two-layer fast name lookup schemes.     

基于图卷积神经网络的函数自动命名

函数自动命名技术旨在为输入的源代码自动生成目标函数名,增强程序代码的可读性以及加速软件开发进程,是软件工程领域中一项重要的研究任务.现有基于机器学习的技术主要是通过序列模型对源代码进行编码,进而自动生成函数名,但存在长程依赖问题和代码结构编码问题.为了更好的提取程序中的结构信息和语义信息,本文提出了一个基于图卷积(Graph Convolutional Network,GCN)的神经网络模型—TrGCN(a Transformer and GCN based automatic method naming).TrGCN利用了Transformer中的自注意力机制来缓解长程依赖问题,同时采用Character-word注意力机制提取代码的语义信息.TrGCN引入了一种基于图卷积的AST Encoder结构,丰富了AST节点特征向量的信息,可以很好地对源代码结构信息进行建模.在实证研究中,使用了3个不同规模的数据集来评估TrGCN的有效性,实验结果表明TrGCN比当前广泛使用的模型code2seq和Sequence-GNNs能更好的自动生成函数名,其中F1分数分别提高了平均5.2％、2.1％.     

基于Web的电信INMS开发方法研究

针对ＩＮＭＳＡ的开发,提出了用ＵＭＬ进行信息建模、用浏览器做人机界面、用分布式组件技术开发管理应用软件的一整套方法。     

网上科技信息深度分析方法及应用

对网上科技信息进行专业化聚合,实现定制分享和态势分析,营造有利于知识发现的研究环境,已成为知识创新和科技管理所面对的重要课题。因此,在分析科技工作者对科技领域公共信息搜索需求的基础上,借鉴行业搜索模式,设计运用垂直搜索引擎等技术手段聚合网上科技信息资源,构建科技信息资源库和共享平台。探讨为科技工作者提供网上科技资源搜索以及科技信息深度分析服务的方法,为知识创新服务。     

集成网络管理系统的责任分层模型

本文提出了一种新型的按责任划分的集成网络管理系统(INMS)的分层模型。利用这个模型,可以实现INMS的系统管理功能与针对不同设备的具体管理功能的分离,从而可以支持“独立开发、集中管理”的系统集成方案。     

基于SOAP的分布式网管系统的设计与实现

讨论了基于SOAP的分布式网络管理系统的优点，提出了一个在大型网络环境下基于SOAP的分布式网络管理体系结构，分析并设计了一个基于该模型的综合网络管理系统INMS的体系结构，解决了网元服务器的远程管理问题，根据综合网络管理系统的开发实践给出了实现方案。     

多智能体在元搜索引擎中的应用研究

随着网络信息的指数增长和用户对检索质量要求的日益提高,传统搜索引擎的查全率和准确率都开始下降,已无法满足用户信息检索的需求。该文从搜索引擎技术的概念入手,分析了元搜索引擎和多智能体的基本原理,并利用多Agent系统自治智能和分布协同的特性．提出了基于多智能体的元搜索引擎。     

支持多领域动态数据集成的数据库网格系统

随着公有数据库资源的丰富,广泛分布的用户希望能够按需地、透明地访问和使用这些丰富的数据资源.DS_Grid(database grid)是一个采用SOA(service-oriented architecture)思想、支持多应用领域数据共享的数据库网格系统.系统采用一种P2P(peer-to-peer)多Chord(MultiChord)网格体系结构,实现数据资源的分布存储、查询处理和动态数据集成;基于文本相似性,可分领域地注册数据资源,实现资源的快速发现;根据领域本体知识和推理规则,实现基于语义的智能查询;采用多根节点多点维护的数据资源副本管理机制,提高系统可靠性;基于关键字过滤的数据集成策略,减少通信代价;采用分布式聚类技术,实现大数据量信息的概要显示.通过实验验证了DS_Grid中所采用的关键技术的可行性和有效性.     

Information mediation in cyberspace: Scalable methods for declarative information networks

An end-to-end discussion, from logical architecture to implementation, of issues and design decisions in declarative information networks is presented. A declarative information network is defined to be a dynamic and decentralized structure where value-added services are declared and applied as mediators in a scalable and controlled manner. A primary result is the need to adopt dynamically linked ontologies as the semantic basis for knowledge sharing in scalable networks. It is shown that data mining techniques provide a promising basis upon which to explore and develop this result. Our prototype system, entitled Mystique, is described in terms of KQML, distributed object management, and distributed agent execution. An example shows how we map our architecture into the World Wide Web (WWW) and transform the appearance of the WWW into an intelligently integrated and multi-subject distributed information network.     

基于云计算的景区数据仓库应用研究

云计算、物联网、大数据等新兴信息技术的发展与应用在提高景区信息化服务水平的同时,也对景区海量信息资源的有效利用提出了严峻挑战。面对超大规模、非结构化的海量数据,传统的基于关系型数据库的数据仓库已很难有效支持景区的数据存储与分析工作。基于此文中提出了一种基于云计算技术的景区数据仓库,通过采用HDFS对数据进行分布式存储管理,利用MapReduce设计海量数据的分析模式,使用HiveQL语言实现数据仓库与前端表现层的交互,能够有效解决景区海量数据的数据管理问题。以黄山风景区为实际背景的实验结果表明了该数据仓库的正确性和有效性。     

基于．NET的水资源管理信息系统设计

针对水资源管理信息系统的界面局限性、DLL黑洞和语言平台兼容不足等问题,设计和实现了基于,NET技术的水资源管理信息系统WRMIS。以四川省水资源管理信息系统为例,增加数据自动采集和网络信息发布功能以完善系统．