Download case study

Get Started

Fill in your details and one of our experts will reach out shortly.

Thank you

Thank you for reaching out.
We will be in touch soon.

How Fyber slashed AWS costs by 20% with no code changes in only one week

Download Case Study
20%
Reduced Compute Costs
17%
Reduced Response Time
15%
Increased Throughput

Fyber turned to Granulate's real-time continuous optimization solution to help them reduce their compute costs and increase their application performance. Granulate agents were installed on Fyber’s AWS EC2 cluster using a Chef cookbook. Following the agent's activation, Fyber experienced significant performance gains. Fyber saw a17% reduction in the response time and 18% CPU utilization all while dealing with 15% more throughput.

About

Fyber is a global technology company, developing a next-generation monetization platform for mobile publishers. Fyber combines proprietary technologies and expertise in mediation, RTB, video and audience segmentation to create holistic solutions that shape the future of the app economy.

Granulate’s real-time continuous optimization is simple, frictionless and delivers great value. It’s all you’d ask for and need for performance and cost optimization
Lior Gildar, Server Team Lead

The Challenge

Fyber processes over billions of events per day in their scalable AWS based RTB service. By integrating with other parts of the marketing technology stack such as ad-servers, demand side platforms (DSPs) and content management platforms, Fyber turns those events into insights and actionable data points. Fyber’s RTB AWS infrastructure receives more than 23 million requests per second, needing hun-dreds of C5.2XL machines to be provisioned in order to adhere to their strict QOS and SLA goals. For Fyber, the company’s success and growth also lead to rapidly increasing compute costs, which in turn impact their operational margin, in a competitive market segment.

The Results

Granulate agents were installed on Fyber’s AWS EC2 cluster using a Chef cookbook. Following several days of autonomous learning and analysis of the resource usage patterns and dataflow, the agents were ready to be activated to start optimizing performance. Following the agent’s activation, Fyber experienced significant performance gains. Fyber saw a 17%reduction in the response time and 18% CPU utilization all while dealing with 15% more throughput. These performance benefits allow Fyber to generate more revenue out of each in-stance, improving their operational margins, in addition to a significant reduction in compute costs.

Figure 1:  17% Response Time Reduction

Figure 2: 15% Throughput Increase

Figure 3: 18% CPU Utilization Reduction

Scroll to top
Skip to content