The Challenge

Digital experience is a key competitive differentiator for those in media and publishing vertical, as they increasingly depend on digital channels to boost it. With a user visit of 100k in 24hrs, our client - Hindawi - aimed at becoming more responsive to its customer’s needs in order to drive more submissions, attract more readers and engagement

The client was also seeking to rebrand its online presence and merge two of its web properties into one. They wanted to establish a scalable platform for a low-latency delivery.

Solving with AWS

To provide the scholars and readers with seamless user experience, Srijan redefined the site performance to manage many requests at the same time on the scalable AWS cloud infrastructure.

The approach was divided into two phases:

Phase 1 - Migrating the data to AWS cloud and merging web properties

Merging the data of Hindawi’s two web properties into one. The web application built on a decoupled architecture - to leverage on the platform and open scope for CaaS.

DynmoDB, an isolated service, was leveraged to help to get the published article meta.

The web properties built on Drupal were then connected by Lambda, which acted as a bridge between DynamoDB data and sharing the data to Drupal backend. It helped to push the Data on SQS so that SQS could capture any modification on DynamoDB for a longer time.

Phase 2 - Enhance the user experience with a scalable and multi-utility cloud platform

Once done with migration the project was focused to improve the performance and do away with low-value monotonous work. Cloudfront, which can manage many requests at the same time, was deployed to make the site high performance. RDS, another high availability and high performance, a database which could create replicas was deployed to assist with website performance.

With AWS Cloud solution, the company works in a flexible, open-source development framework while also addressing customer needs for security-rich data hosting.

Architecture Diagram

When an event is triggered by Hindawi MTS (where the data is stored) the data is updated on Amazon DynamoDB, which stores and pushes the data to Amazon SQS via AWS Lambda. After queuing the data in SQS it is shared with Drupal (which is mostly responsible for the business logic along with the frontend) via another Lambda and Amazon S3 bucket which stores the article artefacts. Drupal collects the information from Amazon DynamoDB and S3 bucket and SQS via AWS Lambda.

Once on Drupal, the information is cached in MySQL RDS, Elastic Cache, and Elastic Search. The frontend renders this information through AWS CloudFront and Drupal via an API middleware.

The complete solution was made possible by Drupal 8, React/Next js.

Besides AWS Cloud, other AWS solutions used were:

• Amazon S3: Working as a DAM for Drupal, Drupal stores its assets on Amazon S3. It is also being used to store the artefacts (HTML of an article, images/figures/SVG) of Articles
  
  
• Amazon EKS: Hindawi is using Amazon EKS build site system with Kubernetes, with EKS being used cluster and managing pods inside it
  
  
• Amazon EC2: We have multiple Amazon EC2 nodes which are managed by Amazon EKS. We also used EC2 heavily for distributed performance testing.
  
  
• Amazon CloudWatch: Debugging and managing lambda functions logs.
  

Business Benefits

• Website load time decreased by 50%
• Sitemap optimized for easy navigation and structure
• Content authoring became easy, user-friendly
• Easy promotion (through syndication) Control the content delivery