McGraw-Hill Education

McGraw-Hill Education continues to innovate, transforming education from a textbook-based model to a next-generation, fully-digital, interactive and self-adapting learning portal model. Learn how McGraw-Hill Education combines Couchbase Server and Elasticsearch to build rich applications and learning aids for professors and students.

Traditional vs. digital media

McGraw-Hill Education (MHE is a traditional media publishing company looking for its new role in the digital media world. The Internet has changed everything, an MHE Labs, the R&D arm of MHE, began taking lessons from advertising and social gaming companies and applying it to educational media. Their goal was to personalize the learning experience, by professionally curating content, delivering it digitally, and in the end making individual users feel like someone understands what they are trying to learn and cares about enhancing that learning experience.

The problem

As learning moves online in ever growing numbers – for example, through massively online courses like Coursera, online university courses, online lectures, and so on – it create a difficult challenge technical challenge. MHE needed to build an interactive learning environment (and underlying technology platform) that:

  • scales to millions of users (learners)
  • supports content from MHE along with content (or metadata) from third parties that enriches it
  • open content (e.g.., Wikipedia)
  • supports the integration of interactive apps
  • self-adapts based on usage data (and captures all the data that needs to be analyzed

The technology approach - and challenge

MHE believed that the combination of front-end HTML5 and the cloud on the back end was the right, dynamic approach to building their platform. The back-end platform, then,  is an interactive content delivery cloud that must:

  • scale, especially to support usage spikes
  • have the ability to catalog and deliver content from many sources
  • provide low latency for metadata and stats access
  • support full-text search
  • offer tunable content ranking and recommendation functions

MHE experimented with a number of technologies including XML databases, SQL, in-memory data grids, and enterprise search servers. However, the problem was that they needed large scale – actually Internet scale – solution that any of these traditional enterprise-scale offerings could provide. They identified Internet-based use cases that pattern-matched what they were trying to achieve: content caching (scale); social gaming (user stats); ad targeting (smarts).

The technologies

MHE Labs explored a number of Internet technologies, and found the right combination in a stack that featured:

  • Backend: Couchbase NoSQL database and Elasticsearch JSON-REST search engine
  • Middleware: Ziniki Java-based integration and security framework
  • Front end: ember, Javascript MVC and ORM framework

The Learning Portal 

Using these technologies, MHE had the vision to take a text book, tear it into “pieces” (similar to an index card), and render each piece in HTML so it looks like a web page. Each of these “cards” is a unit with which they assemble the learning environment. For example, every online course contains multiple cards, which are actually JSON documents. Because they are JSON documents, they can be arranged in different combinations and tailored to meet different learning needs. This vision took form with MHE's Learning Portal project, which:

  • store full-text articles as well as document metadata for images, video, and text content in Couchbase
  • logs user behavior to calculate user preferences (e.g.., user prefers video to text)
  • continuously accepts updates from Couchbase with new content and stats
  • combines user preferences statistics with customer relevance scoring to provide personalized search results

Watch the video to see more details on the learning portal, including a walk-through demo.

NoSQL database technology, and in particular Couchbase, fits well with our vision. As a document database, Couchbase Server 2.0 provides a flexible data model – including indexing and querying, full-text search integration and incremental Map Reduce – on which we can develop content-rich, highly dynamic, interactive applications that will transform how people teach and learn.

– Christopher Tse, Director MHE Labs

McGraw Hill Education is committed to building innovative digital products and services for the education market. Using new approaches to technology to pionee new ways of teaching and learning with digital technologies.