This is an old revision of the document!
====== Proposed Methodology ======
This page describes the proposed methodology in the Yardstick Project.
The methodology consists in decomposing typical VNF work-load performance metrics into a number of characteristics/performance vectors, which each of them can be represented by distinct test-cases.
The metrics shall be divided into areas:
Add your metric suggestion to the tables below.
Way of Working: Once the team decides to implement a specific metric, a link to the corresponding Jira issue will be added.
Compute Area
Metric | Tool | Test Case | Notes |
---|---|---|---|
<Metric> | <Use this tool to measure> | <test case that deliveres the metric> | <Notes, e.g. Tool already in framework?, Dependencies? > |
# cores | |||
# threads | |||
# virtual machines | |||
processor utilization (%) | |||
processing speed (instructions/s) | |||
memory consumption | |||
cpu load (max, average, std deviation) | Linux Kernel | ||
cache performance | Linux Kernel | ||
response time | |||
throughput | |||
bandwidth | |||
availability | |||
QoS |
Storage Area
Metric | Tool | Test Case | Notes |
---|---|---|---|
<Metric> | <Use this tool to measure> | <test case that deliveres the metric> | <Notes, e.g. Tool already in framework?, Dependencies? > |
# read/write | |||
storage size | |||
capacity allocation (object-based, block-based) | |||
block size | |||
usage | |||
QoS |
Networking Area
Metric | Tool | Test Case | Notes |
---|---|---|---|
<Metric> | <Use this tool to measure> | <test case that deliveres the metric> | <Notes, e.g. Tool already in framework?, Dependencies? > |
# packets sent | |||
# packets received | |||
# packets lost (ppm) | |||
# flows | |||
latency per packet | |||
jitter | |||
network bandwidth | |||
link availability | |||
QoS |