Browse Prior Art Database

Method and Apparatus for Scheduling Monitoring Traffic from Multiple Monitoring Products

IP.com Disclosure Number: IPCOM000180185D
Original Publication Date: 2009-Mar-05
Included in the Prior Art Database: 2009-Mar-05
Document File: 5 page(s) / 34K

Publishing Venue

IBM

Abstract

A number of monitoring products are often deployed in an enterprise infrastructure to monitor its performance and health. While they are running, such monitoring products can introduce substantial extra traffic load to the monitored infrastructure. This could affect its normal operation if the monitoring products are not scheduled appropriately, as they are usually running independently and potentially put burst loads on the infrastructure monitored.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 34% of the total text.

Page 1 of 5

Method and Apparatus for Scheduling Monitoring Traffic from Multiple Monitoring Products

A number of monitoring products are often deployed in an enterprise infrastructure to monitor its performance and health. For example, a network performance monitor can be used to periodically report the performance of a network, an event detector can be used to help the system administrator determine the root causes of any problems, a discovery monitor can be used to discover the topology of a network, and so on. While they are running, such monitoring products can introduce a substantial load of extra traffic to the monitored infrastructure. This could affect its normal operation if the monitoring products are not scheduled appropriately, as they are usually running independently and potentially put burst loads on the infrastructure monitored.

All monitoring products tend to follow a similar design pattern: they incorporate monitoring agents that collect information from the OS, machines, servers, devices, etc; and, they have a processing unit that periodically polls information collected from the monitoring agents, aggregates and processes it and then reports the processing results to the system administrator. The monitoring traffic is bursty in nature due to its intermittency and periodicity. As an extreme example, severely bursty traffic occurs if each monitoring product polls information from its monitoring agents at the same time. This can lead to network congestion, delay,

j

                             itter, etc. Therefore, all the monitoring traffic needs to be coordinated to spread the traffic over time, hence minimizing its impact on the infrastructure being monitored.

There are normally two constraints on a monitoring product, namely information timeliness and resource consumption; and, there is a tradeoff between these two constraints. The system administrator can usually specify one or two of these constraints when determining the monitoring policy. To maintain these constraints and prevent the monitored infrastructure from severely bursty traffic patterns, the traffic from each monitoring product should be interleaved in such a way that the impact of the monitoring traffic on the enterprise infrastructure is minimized. This invention presents two methods for coordination and scheduling of monitoring products deployed in the same infrastructure.

We disclose two methods for scheduling the monitoring traffic generated by multiple monitoring products. The monitoring traffic streams from the multiple monitoring products are scheduled in an interleaved fashion to spread their burstiness over time. Since the periodic bursty monitoring traffic from each of the multiple monitoring products is coordinated without overlap, the impact of the monitoring traffic load on the enterprise infrastructure is reduced.

In outline, the process of the proposed two scheduling methods involves the following steps. The steps 1 through 6 are...