Perion Case study

Perion achieves a 52% reduction in EKS costs and 60% reduction in CPU utilization

Download Case Study

About

Perion, a leading company in the online advertising space, delivers data-driven online advertising and search solutions to brands and publishers. Their high-impact advertising division, Undertone, helps brands capture consumer attention via beautiful creative and proprietary ad formats served across top-tier publishers. Perion’s search division, CodeFuel, enables developers to optimize search traffic and generate incremental revenue. MakeMeReach, social advertising platform, drives marketing campaigns and demand generation at scale across Facebook, Twitter, Instagram and Snapchat.

Even though we were skeptical at first, Granulate’s results were amazing, they helped us achieve significant cost reduction in one of our most expensive Amazon EKS services, we will now work to expand it in our entire infrastructure.
Amir Arama, VP Cloud Operations

The Challenge

Perion is a leading company in the online advertising space. Undertone ad serving solution helps companies serve customer tailored ads across top-tier publishers. The infrastructure needed to provide the ad-serving service must be high-performing and operate and strict performance SLAs.

The online advertising company underwent a significant application modernization project, migrating to Kubernetes on AWS EKS. The Kubernetes migration led to improved agility and better scalability but also to increased compute costs. As a publicly traded company operating in a competitive space, this increased the pressure to achieve better operational efficiency and reduce infrastructure costs.

The Results

Intel Granulate was installed on the EKS cluster using a DaemonSet installation. During the autonomous learning stage, the optimization solution analyzed the application’s resource usage patterns and dataflow and identified resource contentions and performance bottlenecks.

Following 4 days of learning, Intel Granulate was activated on a specific service running on the EKS cluster. Within seconds, Perion was surprised by the performance benefits. Monitoring using their own Datadog solution, Perion witnessed 60% CPU utilization reduction of the service pods, along with 10% reduction in the service response time. These performance benefits automatically led to Kubernetes scale-out, reducing the cluster size by 52%.

60%
Reduced CPU Utilization
52%
Reduced Compute Cost
10%
Reduced Response Time

Operating in a competitive market segment, Perion turned to Intel Granulate to achieve better operational efficiency, reduce infrastructure costs and improve application performance. Using Intel Granulate's solution, Perion saw 52% reduced compute cost, 60% reduced CPU utilization and 10% reduced response time, with 0 R&D efforts.