Skip to content

Latest commit

 

History

History
executable file
·
383 lines (308 loc) · 11 KB

talk.org

File metadata and controls

executable file
·
383 lines (308 loc) · 11 KB

Deployment of a Virtual Lab

Who we are?

People

Thirumal Ravula,

Anon Ray

at Virtual Labs Engineering and Architecture Division IIIT, Hyderabad

Work

We develop, engineer and maintain the Central Platform of Virtual Labs.

This includes the hosting infrastructure, and a plethora of other related platform services required by a Virtual Lab during its life cycle.

Agenda of this Workshop

What is Virtual Labs?

Virtual Labs is a collection of online labs running on a common platform and acts as a complement to learning in Engineering and Sciences curriculum. All premier Indian technical institutes are contributing their Virtual Labs to aid Indian students who don’t have access to good laboratory infrastructure.

Why Virtual Labs?

  • Virtual Labs is the next technology to deliver educational content in a highly interactive manner.
  • The platform aims to provide high availability and auto-scalability of your lab.
  • The platform provides usage analytics, single sign on etc.
  • The application can choose to be deployed on its preferred platform.

Another View of Virtual Labs

Hosting Infrastructure

Interplay of services while deploying a lab

Auto-Deployment Service (ADS)

What is ADS

Leveraging ADS

Labspec

Detailed discussion regarding various attributes of the lab spec.

  • How to specify dependencies.
  • How to specify pre-install steps.
  • How to specify install steps.
  • How to specify post-install steps.
  • How to specify various meta information regarding the lab.
{
    "lab": {
      "description": {
            "id": "stall",
            "name": "STALL_LAB",
            "discipline": ["SHOWCASE"],
            ...
            "developer": [
                {
                    "contact": {
                        "email": "[email protected]",
                        "alternate_email": "",
                        "mobile_number": "",
                        "office_number": ""
                    },
                    "department": "",
                    "institute": "",
                    "name": "",
                    "organization": "",
                    "role": "",
                    "title": "",
                    "web": ""
                }
            ]
        },

        "build_requirements": {
            "platform": {
                "arch": "",
                "build_steps": {
                    "build": ["make -C ../src"],
                    ...
                },
            }
        },

        "runtime_requirements": {
            "platform": {
                "arch": "x86_64",
                "installer": ["sudo apt-get update", 
                              "sudo apt-get install -y apache2"],
                "lab_actions": {
                    "init": ["mv /var/www/index.html index.html.default",
                             "cp -r ../src/* /var/www/"],
                    "shutdown": ["service apache2 stop"],
                    "start": ["service apache2 start"],
                },
                "memory": {
                    "max_required": "",
                    "min_required": ""
                },
                "storage": {
                    "min_required": ""
                }
            }
        }
    },
    "template": "1.0"
}

Develop and deploy your own virtual lab

Download the sample template of a virtual lab

unzip t4e-sample-lab-master.zip

Edit the source of the virtual lab template

  • let’s look into the sample template
cd t4e-sample-lab-master
  • the `src` directory contains the source code of the lab
  • the `scripts` directory contains the labspec and other scripts
  • let’s make some modifications in the source HTML files
cd src
  • open index.html in an editor and make modifications.

Edit the lab spec of your new virtual lab

Editing the lab spec

  • We already know the `labspec` file which is required to specify the installation and deployment steps of a lab.
  • Lets look at the labspec now which is located inside the `scripts` directory.
cd ../scripts
  • Open the labspec.json in an editor.
  • Let’s look at the following fields that are necessary for a lab specification, one by one to understand them and fill them with appropriate values.

Meta information related to the lab

  • id: any string which can be used as an identifier for the lab. e.g - let’s use three-letter discipline name followed by 2-digit numeric to represent your lab. But this can be anything.
  • name: the name of the lab
  • discipline: to specify the discipline name of the lab
  • description -> developer: the list of the developers for the lab along with their contact details
{
    "id": "cse02",
    "name": "Computer Programming",
    "discipline": [
        "Computer Science & Engineering"
    ],
    "description": {
        "developer": [
            {
                "contact": {
                    "alternate_email": "",
                    "email": "[email protected]",
                    "mobile_number": "",
                    "office_number": ""
                },
                "department": "",
                "institute": "IIIT Hyderabad",
                "name": "Jawahar C.V",
                "organization": "",
                "role": "Lab Developer",
                "title": "",
                "web": ""
            }
        ]
}

Sample labspec to show the lab information related fields

Build requirements of the lab

  • build steps -> build: to specify any build steps that are necessary for the lab before installation
  • os, and osVersion: to mention on what OS and OS version the build steps to be performed.
"build_requirements": {
    "platform": {
        "arch": "i386",
        "build_steps": {
            "build": ["make -C ../src"],
            .
            .
        },
        "os": "ubuntu",
        "osVersion": "12"
        .
        .
    }
    .
    .
}

Sample labspec to show the build requirment related fields of a lab

Runtime requirements:

  • architecture: on what architecture the lab will run (i386 - 32bit or x86_64 - 64bit)
  • hosting: dedicated hosting or shared hosting; default is dedicated.
  • installer: the list of steps that are to be performed for the lab to get installed. This is the field where one specifies the steps to install the dependencies of a lab.
"installer": ["sudo apt-get update", 
             "sudo apt-get install -y apache2"],
  • lab actions -> init: to specify the steps to run after the lab is installed. E.g - the final deployment steps, post-install steps etc.
"init": ["mv /var/www/index.html index.html.default",
         "cp -r ../src/* /var/www/"]
  • lab actions -> start: the steps to start the services of a lab(if required)
"start": ["service apache2 start"],
  • lab actions -> shutdown: the steps to stop the services of a lab(if required)
"shutdown": ["service apache2 stop"],
  • Snippet of sample run time requirements
"runtime_requirements": {
    "platform": {
        "arch": "x86_64",
        "hosting": "dedicated",
        "installer": [
            "sudo apt-get update",
            "sudo apt-get install -y php5 apache2",
            "sudo apt-get install -y mysql memcache"
        ],
        "lab_actions": {
            "init": [
                "mv /var/www/index.html index.html.default",
                "cp -r ../build/* /var/www/"
            ],
            "shutdown": ["service apache2 stop"],
            "start": ["service apache2 start"],
            .
            .
        }
    }
}

Sample labspec to show the runtime requirment related fields of a lab

System requirements of the lab

  • memory: minimum required memory to be filled out. we also have a cap on the maximum memory a lab would need.
  • os, osVersion: to specify OS and OS version on which the lab would run.
  • storage: the minimum required storage for a lab

Initialise your virtual lab to be a git repository

  • now let’s make our virtual lab a git repo as well.
  • run `git init` inside the sample template directory. Make sure you are at the top level.
git init
  • If git in not present, install git.
apt-get install -y git
  • Add the files
git add .
  • Commit the files
git commit -m "Initial commit"

Create an online repository on GitHub

  • go to www.github.com and login to your account. (Create if you do not have one)
  • Click on New repository button on the home page.
  • Provide the Repository name
  • Provide Description
  • Click on Create repository button

Push the virtual lab repository on to GitHub

  • Copy the repository URL
  • Make it the remote of your newly created local git repository.
git remote add origin https://github.com/<user-name>/<repo-name>
  • Push all the contents of the local repository to GitHub
git push origin master

Go to ADS service and Deploy your lab

  • Open browser and navigate to http://<ip-address>:8080. Replace <ip-address> with the ip address given to you.
  • Fill Lab ID and Repo url fields. Lab ID is the id that you have given in the labspec.json file. Repo url is the url of your github repository.

Watch your lab go live!

  • Hit the submit button and sit back, it takes some time to deploy the lab!
  • Access the new deployed lab by navigating to http://<ip-address>