# Daily Note 12/08/2020
###### tags: `Daily Notes` , `Acumos`
## Name : Christofel Rio Goenawan
## University : Bandung Institute of Technology (ITB)

---

## Schedule:
1. Weekly Meeting.
2. Writes Notes about Firewall Configuration to Solve Open Shift Issues.
3. Study Detailed Works of ML lifecycle management in Acumos AI.
4. Continue to try installing Acumos AI in NTUST Server.

## Outcome :
1. Create Notes about Firewall Configuration to Solve Open Shift Issues.
2. Explained Detailed Background and Works of ML lifecycle management in Acumos AI.
3. Couldn't Give Admin Permission to Installer of Acumos AI.

## Further Plan :
- Continue to deploy Acumos AIO in NTUST server
- Study more detailed about AIO Installation in Kubernets

---

## Daily Log

### 1.Weekly Meeting. <mark>(9.00)</mark>
- Study more detail explanation in [Documentation](https://wiki.onap.org/display/DW/High+Volume+VES+Collector ) and other sources.

### 2.Writes Notes about Firewall Configuration to Solve Open Shift Issues. <mark>(11.00)</mark>
- Study more detail explanation in [Documentation](https://github.com/openshift/origin/blob/release-3.11/docs/cluster_up_down.md#linux) and other sources to solve the issues.
- The notes can be seen [here](https://hackmd.io/@christofel04/TEEP_Notes_Solution-to-Open-Shift-Issues).

### 3.Study Detailed Works of ML lifecycle management in Acumos AI. <mark>(12.00)</mark>
- Study more detail explanation in [Hitachi's Presentasion](https://events19.linuxfoundation.org/wp-content/uploads/2018/07/OSSJ2019-Machine-learning-lifecycle-management-with-Acumos-AI-platform-across-multiple-environment-r2.pdf) and other sources.

### 4.Continue to try installing Acumos AI in NTUST Server.<mark>(15.00)</mark>
- Continue to try to deploy Acumos AIO in NTUST server using Prep- Deploy Process based one previous [study notes](https://hackmd.io/@christofel04/TEEP_Daily_Notes_10_7_2020).

---

## Report

### 1. Firewall Configuration to Solve Open Shift Issues.

>In this note Writer use [Documentation](https://github.com/openshift/origin/blob/release-3.11/docs/cluster_up_down.md#linux) as reference.

:::info
**The Notes can be seen [here](https://hackmd.io/@christofel04/TEEP_Notes_Solution-to-Open-Shift-Issues)**
:::

---

### 2. ML lifecycle management in Acumos AI.

>In this note Writer use [Hitachi's Presentasion](https://events19.linuxfoundation.org/wp-content/uploads/2018/07/OSSJ2019-Machine-learning-lifecycle-management-with-Acumos-AI-platform-across-multiple-environment-r2.pdf) as study sources.

#### MLOps Cycle

Machine learning model is **developed, evaluated, and operated** in **DevOps manner (so called “MLOps”)** in nature because **it requires quick iterations of trial and error.**

##### 1. Data Mining Process Model

To gain insight and create dataset , the workflow can be seen as below.

1. **Problem Understanding**
2. **Data Understanding**
3. **Data Preparation**
4. **Modelling**
5. **Evaluation**
6. **Deployment**

##### 2. MLOps Problems

Some main problems with MLOps can be seen as below.
1. **We need many open source tools to develop models**
2. **Process are iterative, with many tools to operate**.
3. **Too complex to handle workflow**.

The simple problem scheme can be seen as below.


#### Acumos AI

**Acumos AI** is **OSS AI platform hosted by The Linux Foundation**.
It aim to **Makes it easy to build, share, deploy AI apps**.

Some features in Acumos AI can be seen as below.
1. Package tool kits (TensorFlow, scikit-learn) and models with a common API
2. Provide marketplace for sharingAI models internally within company and publicly
3. Container based easy deployment to both public cloud and private environmen

The simple works of Acumos AI can be seen as below.


##### Why Acumos AI Platform?

From presenatation , the 4 advantages in Acumos AI Platform can be seen as below.

1. **Acumos can run inside on- premise environment**
- AWS, Azure and GCP are great! But requires users to put data into public cloud.
- Many OT users cannot ship their data, on-premise solution is necessary.

2. **Acumps supports multiple ML Libraries like scikit-learn , TensorFlow, etc**
- Supporting only TensorFlow, or only scikit-learn is not enough.
- We want general and standard solution to support multiple library
3. **Acumos can handle multi0 tenant with authentication and access control**
- MLflow, and other tools can handle “single-user” mode for now.
4. **Acumos doesn't require any infrastructures skills for Student and Data Scientist**
- Example in Kubeflow , Data Scientist should define CRD to rn simple TensorFlow training job.

#### How Acumos AI Support ML Life Cycle ?

##### 1. Process of ML Life Cycle Now


##### 2. Supports that ML Life Cycle Needed


##### 3. Acumos AI Supports for ML Life Ccle

There are 5 main supports from Acumos AI for ML Life Cycle as below.

##### a. Managed Workspacs And Integrated SDK

Acumos manages **set of training resources and editors (like jupyter notebook) for every scientists by grouping them as “Project.”** •
Data scientists **don’t have to construct and maintain their environment**.


##### b. Efficient Resource Sharing with Training Job Scheduler

Usually it like to run model training jobs on **shared GPU resource pool to gain computation capabilites**
Acumos team will **support NiFi to define and execute training pipelines in the Projects environment**.

:::info
**But in the Hitachi's Presentation this supports hasn't tested yet**
:::

##### c. Data and Experiment Management

**Tracking model training history** is crucial in machine learning development.
Currently, there seems no way to **manage and track the parameters, data, and result of experiments using Acumos AI ML workbench**. It woould be better to have experiment management per “Project,” like jupyter notebook.


##### d. Serving Models with Automated Pipeline and API Generation

In Acumos AI , model is **wrapped by executable platform binary, pipeline, and API endpoints to form microservice**.


Then the downloaded solution deployment package can **be exported to other clusters**.
Deployment script **automatically setup required environment on the target platforms**.

:::info
On-premise environment should be able to connect the development environment via network connection
:::


##### e. Monitoring and Tracking the Behaviour of The Deployed Model

In Acumos AI model performance is **recorded in log, and collected by Beats and ELK stack**.
We can analyze the model performance **from collected set of data**.


##### 4. Conclusion Supports of ML Life Cycle in Acumos AI

The conclusion of supports in Acumos AI for ML Life Cycle can be seen as below.


---

### 3. Try Solve Open Shift Problem for Acumos AI

>This deployment is continuation from [yesterday's notes]( https://hackmd.io/@christofel04/TEEP_Daily_Notes_11_8_2020 ).
In this note Writer use [Previous Notes](https://hackmd.io/@christofel04/TEEP_Daily_Notes_10_7_2020) as study sources.

After checking the Open Shift cluster installation , there are issues when setting cluster up in Open Shift.

:::danger
In Writer's installation the setting freeze after installing kubeflow as below.


But after look at [reference](https://github.com/openshift/origin/issues/21253) , it should end up showing Open Shift user account and server as below..

:::

:::warning
After Writer search in internet and from reference it found out this issue **often happens for CentOS with Open Shift veruson 3.11 like Writer's**.
:::

:::info
**Next Writer Will Continue to Solve the Open Shift Issue in NTUST Server**
:::

---

## Reference
1. https://github.com/nokia/ONAP-VESPA
2. https://docs.o-ran-sc.org/projects/o-ran-sc-ric-plt-vespamgr/en/latest/overview.html
3. https://wiki.o-ran-sc.org/display/RICP/RIC+Alarm+System