Exploratory Analysis of Simple Network Management Protocol Variables in Communications Networks

这个小企业创新研究第一阶段项目将开发一种自动监控SNMP数据的技术。简单网络管理协议（SNMP） Internet管理信息库（MIB I和MIB II）中定义的许多变量的行为概况将被构建并使用统计分析进行评估。在我们的探索性分析中，将每隔15秒收集连接到大学局域网的设备的选定变量的数据，并下载以供以后分析。将采用方法来处理变量的噪声或突发性质，并处理动态和非平稳方面。我们开发的技术可能适用于许多其他行业的诊断问题，这些行业面临着可以监控的数据中非平稳、高度复杂的行为。即使研究不成功，Intellinet, Inc.也无法确定具有足够稳定预测能力的概况或多变量测量，这项工作的结果和他们检验的方法将使研究界受益。此外，他们收集的关于在正常和故障条件下广泛的SNMP变量的行为以及这些变量的相对预测能力的数据在未来可能对网络运营商和设备制造商有价值。

This Small Business Innovation Research Phase I project will develop a technique for automated monitoring of SNMP data. Profiles of behavior for many variables defined in the Simple Network Management Protocol (SNMP) Internet Management Information Base (MIB I and II) will be constructed and evaluated using statistical analysis. In our exploratory analysis, data for selected variables will be collected at 15 second intervals for devices attached to a university LAN and downloaded for later analysis. Methods will be employed to handle the noisy or bursty nature of the variables and to deal with the dynamic and non-stationary aspects. The techniques we develop are likely to be applicable to diagnostic problems in many other industries that face non-stationary, highly complex behaviors in the data that can be monitored. Even if the research is not successful and Intellinet, Inc., is unable to identify profiles or multivariate measures with sufficiently stable predictive power, the results of this work and the methodologies they examine will benefit the research community. Moreover, data that they collect concerning the behavior of a broad class of SNMP variables in both normal and faulty conditions and the relative predictive power of these variables are likely to be of value to network operators and device manufacturers in the future.