Scalability, fault tolerance and fault management for twamp with a large number of test sessions

Abstract

The disclosed methods and systems of using TWAMP measurement architecture for testing a large network include a control-client running on a first network host initializing memory for test session parameters used to originate a test, parsing a configuration file to populate the memory with IP addresses, ports and QoS parameters for control-servers and session-reflectors; and originating test sessions using the test session parameters. The method includes extending to thousands of control-clients, each originating respective test sessions with control-servers in a mesh network using respective test session parameters; and while the test is running, optionally sending an updated configuration file to at least one control-client that introduces a new control-server or replaces a control-server; and the control-client parsing the updated configuration file and updating memory to include the new control-server IP address, port numbers and QoS parameters; and expanding the test and monitoring the running test sessions for results.

We claim as follows: 
     
         1 . A method of enhancing scalability and fault tolerance using a Two-Way Active Measurement Protocol (abbreviated TWAMP) measurement architecture for testing a large network including:
 causing a control-client running on a first network host,
 to initialize an in-memory data store of test session parameters used to originate a test including a set of two-way (abbreviated TW) test sessions originating from the first network host; 
 to parse a configuration file to populate the in-memory data store with destination IP addresses, TCP and UDP transport ports and IP quality of service (abbreviated QoS) parameters for control-servers and session-reflectors; and 
 to originate test sessions with the control-servers and session-reflectors using the test session parameters; 
   extending the causing to initialize the in-memory data store, parse the configuration file, and originate the test sessions to dozens to thousands of control-clients, causing each to originate respective test sessions with control-servers in a mesh network using respective test session parameters;   while the test is running,
 sending an updated configuration file to at least one control-client that introduces a new control-server or replaces a control-server; 
 causing the control-client to parse the updated configuration file and update the in-memory data structure to include the new control-server destination IP address, TCP and UDP transport port numbers and IP QoS parameters; 
 causing the control-client to expand the test to include the new control-server; and 
   monitoring the running test sessions with the control-servers for reports of results.