Application level active networking

In this paper we describe and discuss an Application Level Active Network system. This system provides the benefits of proposed Active Networks, including rapid and transparent deployment of new network services. However our system is also relatively free of the problems of router-level Active Network deployment, such as concerns over safety and resource management. We describe our overall architecture and its components. We then describe and discuss an implementation of the architecture in Java. We present a number of applications that have been implemented on the architecture, and indicate the benefits of our approach.     

Configuring policies in public health applications

Public health is a complex practice due to the requirements of different jurisdictions. These requirements present a challenging environment in which to develop public health applications; software must be flexible in order to adapt to the complexities of different jurisdictions. One approach is to integrate policy management. Policies that define the rules governing an application can be created, modified, or deleted based on the deployment of that application. This paper describes a software architecture and expert system implementation of a policy manager designed to address jurisdictional requirements in public health applications. We define our policy requirements and policy model, the components of the architecture, and how the architecture has been used to implement our policy manager. Finally, we present examples of how the policy manager has configured policies used in three public health applications.     

Compiler Support for Array Distribution on NUMA Shared Memory Multiprocessors

Management of program data to improve data locality and reduce false sharing is critical for scaling performance on NUMA shared memory multiprocessors. We use HPF-like data decomposition directives to partition and place arrays in data-parallel applications on Hector, a shared-memory NUMA multiprocessor. We describe a compiler system for automating the partitioning and placement of arrays. The compiler exploits Hectors shared memory architecture to efficiently implement distributed arrays. Experimental results from a prototype implementation demonstrate the effectiveness of these techniques. They also demonstrate the magnitude of the performance improvement attainable when our compiler-based data management schemes are used instead of operating system data management policies; performance improves by up to a factor of 5.     

Configuration Maintenance for Distributed Applications Management

The MANDAS project has defined a layered architecture for the management of distributed applications. In this paper we examine a vertical slice of this architecture, namely the management applications and services related to configuration management. We introduce an information model which captures the configuration information for distributed applications and discuss a repository service based on the model. We define a set of services and management applications to support maintenance of configuration information, and describe how the different types of configuration information are collected. Finally, we present two management applications that use configuration information.     

Management Architecture for ATM Over ADSL-Based Access Networks

ADSL (Asymmetric Digital Subscriber Line) is a new key technology for offering broadband services to home users and small businesses. ADSL technology is ready to take off. Management discussions aim at achieving an ADSL broadband service architecture are still taking place, however. In order to support multiple services over widely spread networks incorporating various transmission techniques, a cost-effective network operation requires an end-to-end network management architecture. We describe here how to integrate ADSL equipment into the existing telecommunications management network. Alternative protocols and information models are also discussed. The overall network management architecture gives the motivation for the management concept applied to our Digital Subscriber Line Access Multiplexer (DSLAM). The design of such a network element covers all management functionality that is required to multiplex ATM connections being transferred over the ADSL system and to integrate the network element into any telecommunications network. We have chosen a flexible approach to adapt the DSLAM management to different usage environments.     

On Developing Distributed Middleware Services for QoS- and Criticality-Based Resource Negotiation and Adaptation

The Global Resource Management System (GRMS) provides middleware services for QoS- and criticality-based resource negotiation and adaptation across heterogeneous computing nodes and communication networks. This paper presents GRMS's design, prototyping, and performance evaluation. We introduce GRMS design principles and two key concepts unified resource model and ripple scheduling--and describe our architectural design based on these concepts. Further, we present a decentralized end-to-end two-phase negotiation and adaptation protocol with the functionality of distributed, dynamic QoS adjustment and stream preemption. We discuss GRMS's system prototyping and lessons learned and report our experimentation and simulation results, providing insights into design and implementation of a middleware-based distributed resource management system.     

An integration experience of a software architecture and a monitoring infrastructure to deploy applications with non‐functional requirements in computing grids

Resource management is an important aspect to consider regarding applications that might have different non‐functional or operational requirements, when running in distributed and heterogeneous environments. In this context, it is necessary to provide the means to specify the required resource constraints and an infrastructure that can adapt the applications in light of the changes in resource availability. We adopted a contract‐based approach to describe and maintain parallel applications that have non‐functional requirements in a Computing Grid context, called ZeliGrid. To form the supporting infrastructure we have designed a software architecture that integrates some of the Globus services, the LDAP and the NWS monitoring services. Some modules that map the contract approach into software artifacts were also integrated to this architecture. This paper addresses the architecture and integration issues of our approach, as well as how we put the pieces together highlighting deployment and implementation details, which have to consider diverse aspects such as monitoring, security and dynamic reconfiguration. Copyright © 2010 John Wiley & Sons, Ltd.     

A middleware architecture for distributed systems management

This paper presents a middleware solution for global management of any kind of distributed system, such as networks of PCs/workstations, clusters or server farms. Our approach lies in an object-oriented software architecture that models all kind of management information using the common information model (CIM) developed by the Distributed Management Task Force (DMTF). The classes and attributes obtained after the modeling process are mapped to a Lightweight Directory Access Protocol (LDAP) repository. This paper discusses the key features of our middleware that allows that any element (physical, logical, device, user or system) can be managed using a network-oriented and topology-independent approach. A representative example of management domain illustrates the procedures followed to build management applications using our middleware architecture.     

一种改进的家庭网络管理协议--ExHNMP

分析了家庭网络管理协议HNMP在服务管理、联网家电控制和兼容性等方面的局限性。针对这些局限性提出一个HNMP协议的改进和扩展：ExHNMP.ExHNMP是一个分布式的家庭网络管理协议，联网家电可以对等的直接控制网络上的其它联网家电。ExHNMP还引入了简单的服务管理，是联网家电可以根据任务自动选择适当的服务。最后，给出了一个ExHNMP的实现框架，讨论了ExHNMP和其它家庭网络管理协议的互联互通。     

A Software Agent Architecture for Network Management: Case Studies and Experience Gained

Current Network Management paradigms are rigid and lack flexibility. This makes the task of managing a highly evolving and dynamic network difficult to cope with. This paper presents the results of our work on Agent technology as a new paradigm for developing Network Management applications. First, we present our agent architecture that is built in a way that allows the agent to acquire new capabilities at runtime. Second, we present two case studies implemented with a prototype of this agent architecture. The first case study consists of an agent system in which faulty agents are automatically detected, their tasks then being reallocated to other agents, thus providing a fault-tolerant management system. The second case study deals with the configuration of heterogeneous ATM networks to establish end-to-end permanent virtual channels. Finally, we evaluate our agent architecture and the agent paradigm in general when applied to Network Management.     

Active Grid Information Server for grid computing

We describe the Active Grid Information Server for performing expressive resource discovery searches and resource management in a grid environment. We represent server meta-data, such as their CPU power, storage capacity and bandwidth as points in a multi-dimensional space and then express queries as predicates over these points. The design of the system is driven by its application as a part of the information infrastructure for computational grids. Such grids provide an infrastructure for sharing computing resources; an information infrastructure is their inherent part which collects resource data and provides search functionality. Our approach complements current solutions such as MDS by adding Event Condition Action rules and an ability to efficiently handle dynamic attributes. An Event Condition Action rule-based system can support ad hoc, adaptive, flexible, and dynamic schedulers that are modifiable at runtime. We evaluate our Active Grid Information Server by evaluating Event Condition Action rules for termination, confluence, and conflict.     

A framework for distributed management with mobile components

The increasing importance of networks and the growing numbers of devices and services that run on them necessitate effective network and systems management. The traditional centralized management paradigm alone is no longer sufficient for effective management solutions, primarily as it does not scale well. Distribution of management tasks is a promising approach. The distributed management framework (DMF) presented in this paper provides an environment that allows a broad range of management tasks to move and run anywhere within the managed system. In our approach, management tasks are lightweight applications that can be configured and downloaded dynamically as required, reducing the load on managed resources and simplifying the problem of management software updates. We present an object-oriented, Java-based implementation of the DMF and describe applications developed on this platform.     

Component Management in a Dynamic Architecture

Modern distributed programming environments commonly restrict programmers to a single form of inter-component configuration of an application. However, these distributed applications are prone to evolve drastically during their lifetime, to take into account the new users requirements, the technological evolution, or to improve the performance of the application. Evolution involves changes related to the application architecture, the software implementation, or the physical location of software components in a distributed environment during the application execution. In this paper, we describe an original approach for the dynamic application management which fulfills the above requirements with minimal perturbation. In order to prove the advantage of our approach, we have implemented a prototype of this service by extending a CORBA environment.     

Design,implementation, and performance of an automatic configuration service for distributed component systems

Component technology promotes code reuse by enabling the construction of complex applications by assembling off‐the‐shelf components. However, components depend on certain characteristics of the environment in which they execute. They depend on other software components and on hardware resources. In existing component architectures, the application developer is left with the task of resolving those dependencies, i.e. making sure that each component has access to all the resources it needs and that all the required components are loaded. Nevertheless, according to encapsulation principles, developers should not be aware of the component internals. Thus, it may be difficult to find out what a component really needs. In complex systems, such as the ones found in modern distributed environments, this manual approach to dependency management can lead to disastrous results. Current systems rely heavily on manual configuration by users and system administrators. This is tolerable now, when users have to manage a few computers. But, in the near future, people will have to deal with thousands of computing devices and it will no longer be acceptable to require the user to configure each of them. This paper presents the results of our 6 year research (from 1998 to 2003) in the area of automatic configuration, describing an integrated architecture for managing dependencies in distributed component‐based systems. The architecture supports automatic configuration and dynamic resource management in distributed heterogeneous environments. We describe a concrete implementation of this architecture, present experimental results, and compare our approach to other works in the area. Copyright © 2005 John Wiley & Sons, Ltd.     

管理信息系统的开发途径——一个大型系统BD-MIS的研制

本文简单介绍了北京电视机厂计算机辅助企业管理系统(BD—MIS)的系统结构和功能特点;阐述了管理信息系统开发中的数据库设计方法和增量开发方法,以及用户参与系统开发的重要性。     

基于Web/CORBA的网管关键技术的研究

本文在分析了传统的基于Ｗｅｂ的网络管理系统的结构及ＣＯＲＢＡ技术的基础之上,提出了将ＣＯＲＢＡ技术应用于基于Ｗｅｂ的网管系统中,文中论述了这种方式的可行性,并讨论了用分布对象技术建立网管信息模型、Ｗｅｂ对象计算技术在ＷＢＭ系统中的应用及ＣＯＲＢＡ的事件服务在网管的事件处理模块中的应用等关键技术问题。     

Policy‐directed certificate retrieval

Any large scale security architecture that uses certificates to provide security in a distributed system will need some automated support for moving certificates around in the network. We believe that for efficiency, this automated support should be tied closely to the consumer of the certificates: the policy verifier. As a proof of concept, we have built QCM, a prototype policy language and verifier that can direct a retrieval mechanism to obtain certificates from the network. Like previous verifiers, QCM takes a policy and certificates supplied by a requester and determines whether the policy is satisfied. Unlike previous verifiers, QCM can take further action if the policy is not satisfied: QCM can examine the policy to decide what certificates might help satisfy it and obtain them from remote servers on behalf of the requester. This takes place automatically, without intervention by the requester; there is no additional burden placed on the requester or the policy writer for the retrieval service we provide. We present examples that show how our technique greatly simplifies certificate‐based secure applications ranging from key distribution to ratings systems, and that QCM policies are simple to write. We describe our implementation, and illustrate the operation of the prototype. Copyright © 2000 John Wiley & Sons, Ltd.     

Designing peer-to-peer distributed user interfaces: Case studies on building distributed applications

Building a distributed user interface (DUI) application should ideally not require any additional effort beyond that necessary to build a non-distributed interface. In practice, however, DUI development is fraught with several technical challenges such as synchronization, resource management, and data transfer. In this paper, we present three case studies on building distributed user interface applications: a distributed media player for multiple displays and controls, a collaborative search system integrating a tabletop and mobile devices, and a multiplayer Tetris game for multi-surface use. While there exist several possible network architectures for such applications, our particular approach focuses on peer-to-peer (P2P) architectures. This focus leads to a number of challenges and opportunities. Drawing from these studies, we derive general challenges for P2P DUI development in terms of design, architecture, and implementation. We conclude with some general guidelines for practical DUI application development using peer-to-peer architectures.