Skip to content

Commit

Permalink
Update to installation documention
Browse files Browse the repository at this point in the history 
Issue-Id: AIMLFW-47

Change-Id: I8d88cbb9c3ce38f5c9f8ffb4e4e22d43c5d09150
Signed-off-by: josephthaliath <[email protected]>
  • Loading branch information
@josephthaliath
josephthaliath committed Jun 8, 2023
1 parent 78f6836 commit 0ea2aae
Show file tree
Hide file tree
Showing 3 changed files with 199 additions and 7 deletions.
195 changes: 191 additions & 4 deletions docs/installation-guide.rst
View file
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@ Installation Guide
Abstract
--------

This document describes how to install AIMLFW, it's dependencies and required system resources.
This document describes how to install AIMLFW, demo scenarios, it's dependencies and required system resources.


Version history
Expand All @@ -26,7 +26,7 @@ Version history
| 2022-11-30 | 0.1.0 | | First draft |
| | | | |
+--------------------+--------------------+--------------------+--------------------+
| | | | |
| 2023-06-06 | 1.0.0 | Joseph Thaliath | H Release |
| | | | |
+--------------------+--------------------+--------------------+--------------------+
| | | | |
Expand All @@ -43,7 +43,7 @@ Introduction
This document describes the supported software and hardware configurations for the reference component as well as providing guidelines on how to install and configure such reference system.

The audience of this document is assumed to have good knowledge in RAN network and Linux system.
The audience of this document is assumed to have good knowledge in AI/ML tools, Kubernetes and Linux system.


Hardware Requirements
Expand All @@ -52,11 +52,13 @@ Hardware Requirements
Below are the minimum requirements for installing the AIMLFW

#. OS: Ubuntu 18.04 server
#. OS: Ubuntu 22.04 server
#. 8 cpu cores
#. 16 GB RAM
#. 60 GB harddisk

.. _reference1:

Software Installation and Deployment
------------------------------------
.. <DESCRIBE THE FULL PROCEDURES FOR THE INSTALLATION OF THE O-RAN COMPONENT INSTALLATION AND DEPLOYMENT>
Expand Down Expand Up @@ -115,6 +117,8 @@ Software Uninstallation
.. _install-influx-db-as-datalake:

.. _reference2:

Install Influx DB as datalake
-----------------------------

Expand Down Expand Up @@ -301,6 +305,7 @@ Check running state of pod using below command
kubectl get pods -n kserve-test
.. _reference4:

Test predictions using model deployed on Kserve
-----------------------------------------------
Expand Down Expand Up @@ -350,3 +355,185 @@ Use command below to trigger predictions
.. code:: bash
source predict.sh
.. _reference3:

Prepare Non-RT RIC DME as data source for AIMLFW
------------------------------------------------

Bring up the RANPM setup by following the steps mentioned in the file install/README.md present in the repository `RANPM repository <https://gerrit.o-ran-sc.org/r/admin/repos/nonrtric/plt/ranpm>`__

Once all the pods are in running state, follow the below steps to prepare ranpm setup for AIMLFW qoe usecase data access

The scripts files are present in the folder demos/hrelease/scripts of repository `AIMLFW repository <https://gerrit.o-ran-sc.org/r/admin/repos/aiml-fw/aimlfw-dep>`__

Note: The following steps need to be performed in the VM where the ranpm setup is installed.

.. code:: bash
git clone "https://gerrit.o-ran-sc.org/r/aiml-fw/aimlfw-dep"
cd aimlfw-dep/demos/hrelease/scripts
./get_access_tokens.sh
Update the RECIPE file (:file:`RECIPE_EXAMPLE/example_recipe_latest_stable.yaml`) of AIMLFW installation with Influx DB details of RANPM setup. The token is displayed when the script :file:`./get_access_tokens.sh` is executed.
Example of updating the RECIPE file is shown below

.. code:: bash
datalake:
influxdb:
host: <IP of RANPM setup>
port: 31812
orgname: est
bucket: pm-bucket
token: <Token shown when ./get_access_tokens.sh is executed>
Execute the below script

.. code:: bash
./prepare_env_aimlfw_access.sh
Add feature group from AIMLFW dashboard, example on how to create a feature group is shown in this demo video: `Feature group creation demo <https://wiki.o-ran-sc.org/download/attachments/71762231/feature_group_create_final_lowres.mp4?api=v2>`__

Execute below script to push qoe data into ranpm setup

.. code:: bash
./push_qoe_data.sh <source name mentioned when creating feature group> <Number of rows> <Cell Identity>
Example for executing above script

.. code:: bash
./push_qoe_data.sh gnb300505 30 c4/B2
Steps to check if data is upload correctly


.. code:: bash
kubectl exec -it influxdb2-0 -n nonrtric -- bash
influx query 'from(bucket: "pm-bucket") |> range(start: -1000000000000000000d)' |grep pdcpBytesDl
Steps to clear the data in InfluxDB

.. code:: bash
kubectl exec -it influxdb2-0 -n nonrtric -- bash
influx delete --bucket pm-bucket --start 1801-01-27T05:00:22.305309038Z --stop 2023-11-14T00:00:00Z
Training job creation with DME as data source
---------------------------------------------

#. AIMLFW should be installed by following steps in section :ref:`Software Installation and Deployment <reference1>`
#. RANPM setup should be installed and configured as per steps mentioned in section :ref:`Prepare Non-RT RIC DME as data source for AIMLFW <reference3>`
#. To create training job, follow the steps in the demo video: `Training Job creation <https://wiki.o-ran-sc.org/download/attachments/71762231/feature_group_create_training_final_lowres.mp4?api=v2>`__
#. After training job is created and executed successfully, model can be deployed using steps mentioned in section :ref:`Deploy trained qoe prediction model on Kserve <reference4>`. Model URL for deployment can be obainted from AIMFW dashboard (Training Jobs-> Training Job status -> Select Info for a training job -> Model URL)

NOTE: Below are some example values to be used for the QoE usecase training job creation

+--------------------+--------------------------------------------------------------+
| **Parameter** | **Value** |
| | |
+--------------------+--------------------------------------------------------------+
| Training Job Name | qoetest |
| | |
+--------------------+--------------------------------------------------------------+
| Training Function | qoe_pipeline_h_release |
| | |
+--------------------+--------------------------------------------------------------+
| Experiment Name | Default |
| | |
| | |
+--------------------+--------------------------------------------------------------+
| Datalake Source | Influx DB |
| | |
| | |
+--------------------+--------------------------------------------------------------+
| _measurement | ManagedElement=nodedntest,GNBDUFunction=1004,NRCellDU=c4_B2 |
| | |
| | |
+--------------------+--------------------------------------------------------------+
| bucket | pm-bucket |
| | |
| | |
+--------------------+--------------------------------------------------------------+
| Feature Name | \* |
| | |
| | |
+--------------------+--------------------------------------------------------------+
| Feature Filter | |
| | |
| | |
+--------------------+--------------------------------------------------------------+
| Hyper Parameters | epochs:1 |
| | |
| | |
+--------------------+--------------------------------------------------------------+
| Description | test |
| | |
| | |
+--------------------+--------------------------------------------------------------+


Training job creation with standalone Influx DB as data source
--------------------------------------------------------------

#. AIMLFW should be installed by following steps in section :ref:`Software Installation and Deployment <reference1>`
#. Standalone Influx DB should be setup and configured as mentioned in section :ref:`Install Influx DB as datalake <reference2>`
#. To create training job, follow the steps in the demo video: `Training Job creation <https://wiki.o-ran-sc.org/download/attachments/71762231/feature_group_create_training_final_lowres.mp4?api=v2>`__
#. After training job is created and executed successfully, model can be deployed using steps mentioned in section :ref:`Deploy trained qoe prediction model on Kserve <reference4>`. Model URL for deployment can be obainted from AIMFW dashboard (Training Jobs-> Training Job status -> Select Info for a training job -> Model URL)

NOTE: Below are some example values to be used for the QoE usecase training job creation

+--------------------+-------------------------+
| **Parameter** | **Value** |
| | |
+--------------------+-------------------------+
| Training Job Name | qoetest |
| | |
+--------------------+-------------------------+
| Training Function | qoe_pipeline_g_release |
| | |
+--------------------+-------------------------+
| Experiment Name | Default |
| | |
| | |
+--------------------+-------------------------+
| Datalake Source | Influx DB |
| | |
| | |
+--------------------+-------------------------+
| _measurement | liveCell |
| | |
| | |
+--------------------+-------------------------+
| bucket | UEData |
| | |
| | |
+--------------------+-------------------------+
| Feature Name | \* |
| | |
| | |
+--------------------+-------------------------+
| Feature Filter | |
| | |
| | |
+--------------------+-------------------------+
| Hyper Parameters | epochs:1 |
| | |
| | |
+--------------------+-------------------------+
| Description | test |
| | |
| | |
+--------------------+-------------------------+







9 changes: 6 additions & 3 deletions docs/release-notes.rst
View file
Original file line number Diff line number Diff line change
Expand Up @@ -23,12 +23,15 @@ Version history
| 2022-12-08 | 1.0.0 | Joseph Thaliath | G release |
| | | | |
+--------------------+--------------------+--------------------+--------------------+
| 2023-06-07 | 2.0.0 | Joseph Thaliath | H release |
| | | | |
+--------------------+--------------------+--------------------+--------------------+


Summary
-------

The G release of AIMLFW provides required installation scripts for deployment of G release version of AIMLFW
The H release of AIMLFW provides required installation scripts for deployment of H release version of AIMLFW


Release Data
Expand All @@ -41,10 +44,10 @@ Release Data
| **Repo/commit-ID** | aiml-fw/aimlfw-dep |
| | |
+--------------------------------------+--------------------------------------+
| **Release designation** | G release |
| **Release designation** | H release |
| | |
+--------------------------------------+--------------------------------------+
| **Release date** | 2022-12-15 |
| **Release date** | 2023-06-07 |
| | |
+--------------------------------------+--------------------------------------+
| **Purpose of the delivery** | AIMLFW Installation and Deployment |
Expand Down
2 changes: 2 additions & 0 deletions tox.ini
View file
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ deps =
sphinxcontrib-httpdomain
recommonmark
lfdocs-conf
urllib3~=1.26.15

commands =
sphinx-build -W -b html -n -d {envtmpdir}/doctrees ./docs/ {toxinidir}/docs/_build/html
Expand All @@ -27,4 +28,5 @@ deps = sphinx
sphinxcontrib-httpdomain
recommonmark
lfdocs-conf
urllib3~=1.26.15
commands = sphinx-build -W -b linkcheck -d {envtmpdir}/doctrees ./docs/ {toxinidir}/docs/_build/linkcheck

0 comments on commit 0ea2aae

Please sign in to comment.