docs/userguide/opnfv_yardstick_tc043.rst

   1 .. This work is licensed under a Creative Commons Attribution 4.0 International
   2 .. License.
   3 .. http://creativecommons.org/licenses/by/4.0
   4 .. (c) OPNFV, Huawei Technologies Co.,Ltd and others.
   5
   6 *************************************
   7 Yardstick Test Case Description TC043
   8 *************************************
   9
  10 .. _cirros-image: https://download.cirros-cloud.net
  11
  12 +-----------------------------------------------------------------------------+
  13 |Network Latency Between NFVI Nodes                                           |
  14 |                                                                             |
  15 +--------------+--------------------------------------------------------------+
  16 |test case id  | OPNFV_YARDSTICK_TC043_Latency_between_NFVI_nodes_            |
  17 |              | measurements                                                 |
  18 |              |                                                              |
  19 +--------------+--------------------------------------------------------------+
  20 |metric        | RTT, Round Trip Time                                         |
  21 |              |                                                              |
  22 +--------------+--------------------------------------------------------------+
  23 |test purpose  | To do a basic verification that network latency is within    |
  24 |              | acceptable boundaries when packets travel between different  |
  25 |              | nodes.                                                       |
  26 |              |                                                              |
  27 +--------------+--------------------------------------------------------------+
  28 |configuration | file: opnfv_yardstick_tc043.yaml                             |
  29 |              |                                                              |
  30 |              | Packet size 100 bytes. Total test duration 600 seconds.      |
  31 |              | One ping each 10 seconds. SLA RTT is set to maximum 10 ms.   |
  32 |              |                                                              |
  33 +--------------+--------------------------------------------------------------+
  34 |test tool     | ping                                                         |
  35 |              |                                                              |
  36 |              | Ping is normally part of any Linux distribution, hence it    |
  37 |              | doesn't need to be installed. It is also part of the         |
  38 |              | Yardstick Docker image.                                      |
  39 |              |                                                              |
  40 +--------------+--------------------------------------------------------------+
  41 |references    | Ping man page                                                |
  42 |              |                                                              |
  43 |              | ETSI-NFV-TST001                                              |
  44 |              |                                                              |
  45 +--------------+--------------------------------------------------------------+
  46 |applicability | Test case can be configured with different packet sizes,     |
  47 |              | burst sizes, ping intervals and test duration.               |
  48 |              | SLA is optional. The SLA in this test case serves as an      |
  49 |              | example. Considerably lower RTT is expected, and             |
  50 |              | also normal to achieve in balanced L2 environments. However, |
  51 |              | to cover most configurations, both bare metal and fully      |
  52 |              | virtualized ones, this value should be possible to achieve   |
  53 |              | and acceptable for black box testing. Many real time         |
  54 |              | applications start to suffer badly if the RTT time is higher |
  55 |              | than this. Some may suffer bad also close to this RTT, while |
  56 |              | others may not suffer at all. It is a compromise that may    |
  57 |              | have to be tuned for different configuration purposes.       |
  58 |              |                                                              |
  59 +--------------+--------------------------------------------------------------+
  60 |pre_test      | Each pod node must have ping included in it.                 |
  61 |conditions    |                                                              |
  62 |              |                                                              |
  63 +--------------+--------------------------------------------------------------+
  64 |test sequence | description and expected result                              |
  65 |              |                                                              |
  66 +--------------+--------------------------------------------------------------+
  67 |step 1        | The pod is available. Two nodes as server and client.        |
  68 |              | Ping is invoked and logs are produced and stored.            |
  69 |              |                                                              |
  70 |              | Result: Logs are stored.                                     |
  71 |              |                                                              |
  72 +--------------+--------------------------------------------------------------+
  73 |test verdict  | Test should not PASS if any RTT is above the optional SLA    |
  74 |              | value, or if there is a test case execution problem.         |
  75 |              |                                                              |
  76 +--------------+--------------------------------------------------------------+