A Foundation for Generalized Network Fault Management

随着计算机通信网络从提供尽力而为的服务发展到提供具有严格的服务质量(Qos)保证的实时服务(例如视频)，对改进的故障管理技术的需求增加。网络故障管理是一个持续增加复杂性的开放问题。这个问题复杂性增加的一个主要原因是网络的异构性。当今的网络由多种技术(例如无线、ATM、以太网)组成，这些技术有望支持各种联网应用(文件传输、WWW、视频会议)。对于不同的技术，故障管理问题已经得到了不同程度的研究，但仍有许多未解决的问题29、40，我们不知道有任何方法可以有效地跨技术应用。目前的方法在很大程度上是临时的，并且依赖于人类网络管理员的专业知识。即使在单一技术网络(例如，以太局域网)中，为一个网络开发的策略也不能轻易推广到其他网络。此外，网络内部的变化速度使网络管理员很难保持足够高的专业水平来开发和维护有效的故障管理策略。需要能够跨技术推广和应用的新技术，以高效地确保网络可靠性。这项研究的长期目标是实现尽可能多的网络管理自动化。在网络管理中，我们将首先关注故障管理。沿着这些思路，计划开发故障管理方法，这些方法(1)可以跨不同的技术应用，(2)可以在网络之间推广，以及(3)可以适应网络中的变化。先前的研究12、13、14已经证明自适应统计技术可用于网络故障检测。这些类型的技术使用网络测量来学习网络的行为，因此可以推广。此规划期的目标是确定故障管理中的一组问题，这些问题可以使用严格的、可推广的方法来解决。故障管理包含大量复杂的问题，这些问题尚未得到很好的研究或定义。在评估方法之前，必须确定要解决的问题。将从故障管理的角度研究每种网络技术，并提取所有或大多数技术中常见的一组问题。对于确定的每个问题，将评估使用自动化技术和推广的可行性。将使用TCP/IP测试网络获得初步结果。规划活动的结果将是一个研究方向，重点是一组广泛适用于网络技术的可行问题。将为故障管理研究规划项目创建一个网页。该网页将载有对各项活动和取得的成果的说明。该网页的URL将为http://www.nml.csam.iit.edu/~planning/.

As computer communication networks evolve from offering best-effort service to offering real-time services with stringent Quality of Service (QoS) guarantees (e.g. video), the need for improved fault management techniques increases. Network fault management is an open problem that continues to increase in complexity. A major cause of the increase in complexity of this problem is the heterogeneity of networks. Networks of today are comprised of a variety of technologies (e.g. Wireless, ATM, Ethernet), that are expected to support a diverse set of networked applications (file transfer, WWW, video conferencing). The fault management problem has been studied in varying degrees for different technologies, but there are many open issues 29 , 40 , and we are unaware of any methods that can effectively be applied across technologies. Current methods are largely ad hoc and rely on the expertise of a human network manager. Even within single-technology networks (e.g. Ethernet LANs), strategies developed for one network, cannot be easily generalized to other networks. In addition, the rate of change within a network makes it difficult for the network manager to maintain a high enough level of expertise to develop and maintain effective fault management strategies. New techniques that can be generalized and applied across technologies are needed to ensure network reliability in an efficient manner. The long-term goal of this research is to automate as much of network management as possible. Within network management, we will first focus on fault management. Along these lines, the plan is to develop methods for fault management that (1) can be applied across different technologies, (2) can generalize from network to network, and (3) can adapt to changes in the network. Previous research 12 , 13 , 14 has demonstrated that adaptive statistical techniques can be used for network fault detection. These types of techniques use the network measur ements to learn the behavior of the network and can thus be generalized. The goal for this planning period is to identify a set of problems within fault management that can be addressed using rigorous, generalizable methods. Fault management encompasses a large, complex set of problems that have not been well studied or defined. Before methods are evaluated, the problems to be addressed must be identified. Each network technology will be studied in terms of fault management, and a set of issues that are common across all or most of the technologies will be extracted. For each of the problems that are identified, the feasibility of using automated techniques and generalization will be evaluated. Preliminary results will be obtained using the TCP/IP test network. The result of the planning activities will be a research direction that is focused on a set of feasible problems that are widely applicable across network technologies. A web page will be created for the fault management research planning project. The web page will contain a description of the activities and results obtained. The URL of the web page will be http://www.nml.csam.iit.edu/~planning/.