AppCritical Standalone Sequencer Overview

The AppCritical network analysis product employs a very unique methodology for testing and analyzing IP networks. This analysis is accomplished by 'transmitting' small groups of ICMP Echo Request and UDP packets from one workstation/server ("transmitter") through the network in question to a second workstation/server ("receiver") and gathering the timing information for each of the packets within the group. This information is then analyzed by the AppCritical Network Intelligence System (NIS) which will generate a mathematical model of the "apparent network", or the network as it appears to an application. This allows it to determine a complete set of performance metrics as well as provide extensive diagnostic information. The software used to transmit the packets and gather the timing information is a standalone executable called the StandAlone Sequencer or SAS.

The SAS is a very compact software component; less than one megabyte for most platforms, and is capable of being executed without involving any additional libraries or external dependencies. When running it will typically consume less than 4MB of memory, and less than 0.1% of CPU resources. The primary purpose of the SAS is to place packets on a network in an extremely efficient manner. We do not need to place a lot of traffic on a network to achieve accurate measurements. What is important is that these packets are put on the network in a very controlled manner so that we can achieve extremely accurate measurements of how the packets were affected by the devices that 'handled' them during their trip between the servers.

In a typical test run between two systems the SAS (using default values) will transmit approximately 20 to 40 packets per second for each hop on a 100Mbps Ethernet network (less for slower or more congested links.) The packets will typically vary in size from 46 byte to 1500 byte payloads. On a 100 Mbps network this adds much less than 1% to the entire network load. For our own purposes it is important that we do not contribute significantly to the network load because we do not want to influence our own test results. This load is so light that it can be used with no fear of triggering Denial of Service (DoS) alarms or Intrusion Detection Systems. The SAS also has a self tuning capability; it will adjust the rate that it transmits according to the speed of the network it is testing. On slower networks it will use a much larger delay between the packet groups; this will take longer to perform a test, but will prevent any noticeable loading of the network.

For our analysis we do not require any proprietary or possibly confidential information. The SAS does not collect any RMON, SNMP or other system information. It will perform a very basic DNS lookup of any device names to match these names to IP addresses. It will then proceed to test and verify the Path MTU settings of each hop through the usage of varying packet sizes. Once this has been completed it will proceed to transmit the test packet groups and collect the packet timing information.

When a test has completed, the SAS will write its output to a file in the XML format. This file is 'human readable' and can be checked to verify that no proprietary or secure information has been collected. Once this file is returned, it will be uploaded by the NIS and a full analysis can be performed. This analysis can be summarized and output as a report in PDF format for distribution.

Back