Due to an aging customer demographics and smartphones offering more features, our client was looking to develop a Mobile Application that would create customized, targeted customer alerts related to functional telemetry data (usage, battery life, storage remaining, and network). The goal was to make it easier for customers to anticipate utilization risks and needs and to enhance the user experienced through persona-based profiling recommendations.

Synechron developed a Mobile application solution for a Japanese client with a highly-elastic and scalable backend platform to allow for low infrastructure maintenance costs and rapid, iterative development and deployment. The app boasted robust authentication mechanisms for mobile clients, data security at rest and in-transit functionality.

Synechron conducted a use case-driven vendor analysis, and Amazon Web Services (AWS) emerged as the top choice for architecting the solution due to its ability to handle the key functionality required for the Mobile Big Data and Analytics Application:

• Scalability - ingest variable volumes of mobile telemetry data across thousands to hundreds of thousands of mobile devices and conduct computational processing and analysis to derive meaningful, timely insights and customized notifications
• Fault tolerance - have appropriate monitoring, logging and alerting capabilities as is expected from enterprise-scale applications
• Data Security - whether in transit or at rest and with the appropriate mechanisms to authenticate and authorize requests coming in from mobile clients
  • PII data (Personally Identifiable Information) which should be secured to local regulatory requirements
• Cost Effectiveness – across hardware and software

The analysis found AWS would provide a combination of auto scaling groups and variable sized EC2 instances to provide push button scalability for a range of thousands to millions of daily requests, a highly-scalable service (Kinesis) to ingest high-volume mobile telemetry data, and elastic Map Reduce cluster deployment in the cloud to process and derive data insights. Services like Simple Queue Service (SQS) and Simple Notification Service (SNS) were a value add to send high volume notifications and SNS connectors with an in-built authentication and authorization mechanism (Cognito). AWS also would provide a highly-scalable caching mechanism (elasticache) which includes Redis and Memcached, a secure deployment within Virtual Private Cloud - protected by Security groups and Access Control Lists, and a highly-scalable elastic object storage service with durability greater than or equal to 99.99%. Both RDBMS and NoSQL solutions could be used for persistence.

Synechron developed the backend technical architecture of the application, which included:

• REST APIs hosted on AWS Elastic Beanstalk Web Installation receiving telemetry and analytics data from devices
• A Virtual Private Cloud housing all AWS components with specific network ingress and egress definitions
• Beanstalk Stack description – Java 1.8, Tomcat servers running 64-bit Linux AMIs – both web installed and worker installed
• Autoscaling based upon CPU utilization and time of day rules
• Cognito stores authentication data
• Telemetry data posted to an EMR cluster hosted in cloud by a Java Lambda function which pulls data from Kinesis streams
• SNS Queues to decouple services within the system
• A Lambda function which pushes analytic events to mixpanel
• Mixpanel to initiate notification campaigns based on user segmentation data
• Redis Elasticache is used to store a One-time password (OTP) generated during user validation
• SNS connector applications for GCM and APNS push notifications as well as SMS delivery