Scalability experiments Sample Clauses

Scalability experiments. The scalability experiments include ScalingMsgsA, ScalingMsgsP , Scaling- CPUA and ScalingCPUP . These experiments investigate the ability of each monitoring mode to scale in distributed environments. In the ScalingMsgsA and ScalingMsgsP experiments, consumers and pro- viders create agreements that use either only the active monitoring mode with a fixed interval of 30 seconds or only the passive monitoring mode with a fixed interval of 10 seconds. These values were chosen based on the per- formance of previous experiments. One TTP and two TMMs are used. As with the previous set of experiments, each of these experiments is run for 30 minutes and results are averaged. Each experiment is repeated with the number of agents ranging from 2 to 100, in increments of 2. In the ScalingCPUA and ScalingCPUP experiments, a consumer agent and a provider agent are launched simultaneously on 32 separate Locations on 32 separate nodes. A consumer chooses a provider from amongst the Locations using a random function with a uniform distribution. The chosen provider is always located on a different node than the consumer. After choosing a provider, the consumer creates an agreement that uses monitoring. ScalingCPUA uses active monitoring mode with an interval of 10 seconds. ScalingCPUP uses passive mode with an interval of 60 seconds. These values were chosen based on the performance of previous experiments. A new consumer agent and provider agent are launched every 5 seconds until the desired number is reached. For the first run, this number is 64 (2 per node). The number of agents increases by 64 on each consecutive run until reaching 4096 (128 per node). The entire process is repeated 5 times and the CPU load results for all 32 participating nodes are averaged.