Amazon Web Services
Influenster encourages consumers to share their product experiences and lets other shoppers learn about the specifics of products and brands. Overall, the platform increases the voice of consumers and helps brands to improve their offerings in line with the consumer perspective and their experience. As today’s consumers embrace online product research and reviews, Influenster becomes much more important for highly educated and engaged consumers. Their active user base continues to grow as online shopping becomes more convenient and users compare product information online. Their growth in the market was exceptional and Influenster was acquired by Bazaarvoice, a global software company that helps brands and retailers collect and leverage consumer-generated brand content on their e-commerce websites. Influenster displays over 42 million product reviews and has 6 million+ members on their platform. With the Bazaarvoice’s global network of 6000 brand and retailer websites, Influenster becomes the first platform that comes to mind when searching for product reviews.
Influenster’s website and mobile applications were backed up with an exceptional number of asynchronous tasks to take specific actions such as sending notifications to consumers, checking user streaks, etc. The application was processing over 500M tasks per month and expected to increase further in the future.
They had been using RabbitMQ queue system to collect and process all these asynchronous tasks that the application requires. In detail, RabbitMQ was running on two large typed EC2 instances to provide high-availability for the applications. However, it started to become an unscalable component on their infrastructure as their active user number and task count grew over time. Emerging scalability and reliability issues caused blocked operations, increasing end-user complaints due to improper task runnings. Influenster teamed up with Sufle to meet their increasing demand and growth opportunities with a proper solution for their infrastructure.
Influenster had been running all production and test environments on AWS for more than 8 years. They had a considerable amount of experience both in their environments on AWS Cloud and related AWS Services. They engaged with the Sufle team considering their expertise and experience on AWS technologies and service implementations. One of the reasons encouraged Influenster to work with Sufle was the Sufle team’s commitment to successful organizational adaptability in their consulting practices. Sufle’s proven experience in Python and PostgreSQL assured the Influenster team that their infrastructure would be assessed and understood to the fullest extent.
Sufle suggested using Amazon SQS to replace RabbitMQ, as a highly available and scalable message queuing service. Amazon SQS was suggested to decrease the operational workload of applications as well. To help the Influenster team effectively leverage the newly implemented service, Sufle team organized training sessions and explained the limits and features of Amazon SQS that differs from RabbitMQ. Throughout the implementation and training, Sufle coordinated with the development team to integrate the application with Amazon SQS.
Amazon Simple Queue Service enabled Influenster to improve its operational efficiency and achieve a reliable environment. Besides its operational advantages and built-in high-availability features, Amazon Simple Queue Service also helped Influenster realize cost efficiencies as they operate. The service is 46% cheaper on current task load and 57% cheaper on future-predicted loads when compared. In this way, Influenster is now able to scale their operations with ease and possible minimum cost. Influenster continues to work with Sufle to receive consulting services on their cloud operations and AWS service implementations.