KPI Mappings

The data is generated from deployments of a virtualised media processing function using different allocations of the vCPU and memory. The specific combinations are as follows: | Name | RAM | Disk | Ephemeral | VCPUs | +------------+-------+------+-----------+-------+ | i1.small | 1024 | 80 | 0 | 1 | | i1.medium | 2048 | 80 | 0 | 1 | | i1.large | 4096 | 80 | 0 | 1 | | i1.xlarge | 8192 | 80 | 0 | 1 | | i2.small | 1024 | 80 | 0 | 2 | | i2.medium | 2048 | 80 | 0 | 2 | | i2.large | 4096 | 80 | 0 | 2 | | i2.xlarge | 8192 | 80 | 0 | 2 | | i3.small | 1024 | 80 | 0 | 4 | | i3.medium | 2048 | 80 | 0 | 4 | | i3.large | 4096 | 80 | 0 | 4 | | i3.xlarge | 8192 | 80 | 0 | 4 | The instances of virtualized media processing function are stressed using a workload generator (HAMMER) with the following profile. 5 second steps of 5 incremental users for 50 steps resulting in a final simulated user loads of 250 users. The telemetry data in the dataset is collected using the open source snap telemetry framework for each unique deployment. The dataset also integrates selected metrics as measured by Hammer such as through and latency for each unique deployment. The dataset contains two iterations for each unique deployment of the of the virtualized media processing function This data can used as input into analytics process such as KPI mapping.