AVNMP

Active Virtual Network Management Prediction
Chapter 7: Atropos Driving Process

Stephen F. Bush and Amit Kulkarni
(bushsf@crd.ge.com, kulkarni@crd.ge.com)
 
Previous Chapter
Beginning of Tutorial
Next Chapter
This chapter assumes that the previous chapter covering the SNMP instrumentation and connection to Atropos of an active network application in Magician has been read and understood. In this chapter, and throughout the rest of this tutorial, we add incremental features to our sample application known as LoadGen. Recall from the previous chapter that we have enhanced the number of packets transmitted with Atropos by making its SNMP Object Identifier known to Atropos.

Key Concept

The goal in this chapter is to take a more detailed look at the Driving Process and its role in the optimistic simulation based virtual overlay of Atropos. The Driving Process generates virtual messages which represent messages and/or states expected to occur in the future. These message are automatically propagated throughout the system and drive Atropos forward in time and space. A simplified view is shown below. The Driving Process makes a prediction, generally by either encapsulating or sampling the edge of the Atropos-enhanced system. Then the prediction is sent to the destination Logical Process as an active packet. A copy is stored in the Send Queue by sendvm() in case a rollback is necessary: 


Predict -> output -> sendvm()

The following lines of code control the performance of the Driving Process:
 final static double lvm = 0.5;

final static long Dvm = 20000;

The first line is the message generation rate of virtual messages. The second line is the interval of virtual time between virtual messages; i.e. the time step of the simulation.

Running the Complete Atropos Active Application

After reading through this code and understanding what it does, you can compile the code "as-is" using the included Makefile. Run in the Magician execution environment by entering:
 
  • send InjectSnmp AH-1 
  • send LoadGenAppBasic AH-1 
  • send Flood AH-1 
  • send AvnmpSmartDriver AH-1 

  •  
The object identifiers created within the application and the predictions generated by AVNMP should now be queriable via an SNMP client. 
Previous Chapter
Beginning of Tutorial
Next Chapter


 The complete Atropos documentation is found in the AtroposAPI .

This research is funded by DARPA/ITO Contract Number: F30602-98-C-0230 supported by the Air Force Research Laboratory/IF.
The project web site is http://www.crd.ge.com/~bushsf/an


 
SourceForge
Logo