Technology

Our Technology Stack – What we build on.

Posted by the Millionlights Tech Team:

So we have been asked  by our students and partner universities how do we plan to reach millions of students and teach them in terms of technology and scaling.

This post is about technology and what we are using to build out the platform.

o4QvTHxv2ifp60

The Millionlights Platform – Security and Features – Extending OpenEdx 

1. LMS ( Learning management system)

It is one of the most visible parts of Millionlights with which the students interact. It displays content, run quizzes, discussion forum and interactive apps. LMS also provides instructor dashboards.

It uses a number of data stores. Courses are stored in MongoDB along with the videos served from YouTube and Amazon CDN and the student data is stored in MySQL.

2. CMS( Content management system) or Studio or course authoring tool:

It is the course authoring environment. The Millionlights Studio lets you weave your content together in a way that reinforces learning, insert videos, discussions, and a wide variety of exercises over a friendly graphical user interface. The kinds of exercises or lectures or videos one can create are virtually limitless.

Open edX integrates the LTI components, customized Javascript as well as tools such as Google Instant Hangout and even a Molecule Editor and much more.

Discern allows anyone to use machine-learning-based automated textual classification as an API service.

This is an API wrapper for a service to grade arbitrary free text responses. The goal is to provide a high-performance, scalable solution that can effectively help students learn. Feedback is a major part of this process.

 

  1. Discussion forum

It’s a fairly simple discussion service that interacts with MongoDB. A server-side RoR application that supports the voting, nested comments and instructors’ endorsements.

  1. XQueue Service.

It establishes an interface for the LMS to communicate with the external grader services. For example, when a learner submits a problem in the LMS, it is sent to XQueue in order for it to be processed further.

 

  1. XServer

XServer accepts the student code submissions from the LMS and runs the code in AppArmor/Sandbox. This is to be used with the edX-platform and xqueue.

 

  1. ORA2

It allows for the assessment of open response problems on the Millionlights platform.

 

  1. Discern

Discern allows anyone to use machine-learning-based automated textual classification as an API service.

This is an API wrapper for a service to grade arbitrary free text responses. The goal is to provide a high-performance, scalable solution that can effectively help students learn. Feedback is a major part of this process.

 

  1. EASE

EASE (Enhanced AI Scoring Engine) is a library that allows for machine learning based classification of textual content. This is useful for tasks such as scoring student essays.It provides functions that can score arbitrary free text and numeric predictors. The goal here is to provide a high-performance, scalable solution that can predict targets from the arbitrary values.

 

  1. Edinsights

EASE (Enhanced AI Scoring Engine) is a library that allows for machine learning based classification of textual content. This is useful for tasks such as scoring student essays. It provides functions that can score arbitrary free text and numeric predictors. The goal here is to provide a high-performance, scalable solution that can predict targets from the arbitrary values.

 

  1. Notifier

Notifier sends the daily digests of new content to the subscribed forum’s users, with a goal of eventually supporting the real-time and batched notifications of various types of content across various channels.

11. Azure and AWS – We use Azure for hosting our Millionlights front-end and AWS for the LMS and Studio.

Scaling Millionlights

An overview of how we are planning to scale when we hit 500k users.

 

aws

 

cropped-11.jpg