Performance Optimization of a video
streaming platform
Footters is a streaming platform dedicated to the football world, their services are based on football content provided by their partners and suppliers as well as their affiliated football club.
Footters connect with those who want to know a modest, close, and quarry football. A different, more competitive, and exciting football. A football with a multitude of games to enjoy and great stars to discover. A football that until now was practically invisible.
The main infrastructure that supports requests to access to the streaming videos was deployed on AWS, relying mainly on AWS Elastic Beanstalk services for hosting all the backend and Amazon RDS to manage their database.
In spite of having enough computational resources deployed to sustain double the actual number of transactions, the infrastructure was found unable to withstand the user request workload it was receiving, as well as displaying a significant latency in responses to such received requests.
Our main challenge was to optimize and increase the platform performance through making changes on the configurations as well as the integration of new tools and services to attending to all the workload caused by the user requests, we also had to solve the latency problem and reduce infrastructure costs.
The first this we established with the client was the execution of a deep analysis of the AWS architecture with the implementation of advanced monitoring tools and technologies.
Through this analysis, we were able to identify the problems and improvements that should be implemented based on the best practices of a well-architected infrastructure: performance, costs, and security.
After this analysis a new architecture was designed to meet the real needs of the platform, here is the detail of what we did:
We deployed a new networking ecosystem (VPC), a multi-layer system, guaranteeing the security and high availability of the solution.
The entire Backend was deployed on a custom system based on AWS EC2, eliminating the AWS ElasticBeanstalk solution, this allowed us a much higher level of customization and configuration.
For the database system, we used Amazon Aurora MySQL to get more scalability concerning Reader capacity, as well as the integration of several ElastiCache clusters with REDIS engine to guarantee an improvement of data delivery.
All those changes made us proceed to migrate the entire ecosystem of Footters to the new architecture, this was done on a scheduled basis and designed jointly with the client, to avoid any kind of disruption for the users.
We also assured the full management of the infrastructure with our Managed Cloud service Evolve, monitoring and evolving the platform with our DevOps service