Driving Visibility in Business Travel Operations with Real-Time Analytics

The Increasing Need for the Analytic Data Capacity and Scalability

TripActions was looking to design and build a new data streaming solution, because:

• The increasing amount of historical indexed data overstrained the company’s infrastructure and primary storage solution slowing down its performance and causing an ever-increasing cost of ownership
• The historical data was never cleaned while analytical data was stored across various databases in different formats, which created multiple data silos and made data unavailable for analytics and machine learning
• The company’s existing data solution failed in terms of analytic capacity and scalability, which increased operational costs, slowed down onboarding of new clients, and stifled business growth

TripActions’ initial architecture and data solution were based on Amazon Elasticsearch, which proved to be inefficient and expensive when data volumes increased.

Data was schemaless, and there was no mechanism to join data from different databases. Partial data in Amazon S3 was stored in JSON format and synced with one-day lag, with no partitioning, which delayed TripActions’ reaction to issues or changes in data. Some data shapes were stored in Redshift, others in MySQL. MySQL intermediate instances and ad-hoc scripts were used as intermediate ETL storage.

TripActions reached out to Provectus to implement a new data streaming solution that could eliminate the abovementioned issues and accelerate the company’s global expansion.

Building a Real-Time Streaming Solution for Data Processing and Transformation

Provectus delivered a real-time streaming solution on top of Apache Kafka for data ingestion, processing, enrichment and transformation capable of accommodating the growing volumes of historical and analytical data on the platform.

Data streams were consumed by real-time reporting and machine learning applications. All the data streams were designed in a way that they could be automatically stored in the data lake on top of Amazon S3 and AWS Glue. They were optimized for sub-second queries from Amazon Athena.

The historical data from the existing platform was migrated from ElasticSearch, Redshift, RDS to a separate data lake to ensure data consistency and availability for analytics and machine learning.

All Data Team-related services were moved to a separate Data VPC to improve the team’s productivity and gain better visibility into data.

Accelerating Business Growth and Reducing TCO on a New, NextGen Data Platform

TripActions received a new real-time data streaming solution, which allowed the company to accelerate it month over month business growth by more than 30%. The company saw an 80% reduction in total cost of ownership. It was able to shorten release cycles for its data analytics project by 12x and to provide data analytics capabilities to its internal users and customers.

More than 200 data products power TripActions day-to-day decisions and operations. Personalized traveler experience, proactive 24/7 customer support, and enterprise reporting are key features of the TripActions platform, which are deployed on the streaming data platform and that contribute to TripActions’ business growth.

Moving Forward

1. Learn more about the Provectus NextGen Data Platform and AWS Cost Optimization
2. Explore more customer success stories covering data platforms: Swiftmile, Second Genome, IMVU, Appen
3. Apply for the Provectus NextGen Data Platform Acceleration Program to get started