EverC: Regaining Control of Data Platform Costs on AWS with EMR on EKS

EverC, a technology company operating in highly regulated environments, set out to regain control over growing data platform costs without compromising analyst productivity or operational flexibility.

Working with BigData Boutique, EverC evaluated its existing Spark workloads and redesigned its data processing platform around Amazon EMR on EKS. The result was a more cost-efficient, AWS-native architecture that preserved familiar workflows while restoring predictability and control over compute economics.

EverC, now part of G2 Risk Solutions, is a pioneer in AI-driven risk intelligence dedicated to making the internet a safer and more transparent place for global ecommerce. The company provides advanced risk and compliance technology that empowers banks, payment providers, and marketplaces to manage risk throughout the entire merchant lifecycle.

The EverC platform relies heavily on large-scale data processing to support monitoring, analysis, and decision-making across customer datasets. As with many data-driven companies, EverC's engineering and analytics teams depend on Apache Spark for both batch processing and interactive analysis. Over time, these workloads became central to the business — and so did the infrastructure costs associated with running them.

EverC's Spark workloads provided a strong developer and analyst experience, but as usage grew, so did costs. Over time, the pricing premium became increasingly difficult to justify, especially for workloads that did not require all of the infrastructure's higher-level abstractions.

The challenge was not limited to simply reducing spend. EverC also aimed to:

• Continue supporting both interactive analyst workflows and ETL workloads
• Maintain a familiar notebook-driven experience for data teams
• Stay fully aligned with AWS-native infrastructure and security controls
• Improve cost predictability and reduce platform overhead

This was a strategic decision rather than a crisis response. The current system worked, but it was no longer economically aligned with EverC's long-term needs.

BigData Boutique partnered with EverC to evaluate alternatives and design a data platform that balanced cost efficiency, flexibility, and usability. Rather than recommending a one-size-fits-all replacement, the engagement focused on understanding how EverC actually used Spark in production.

Evaluating EMR Deployment Models

Together, the teams assessed several AWS-native options:

• EMR Serverless, for simplified operations and bursty workloads
• EMR on EC2, for traditional yarn-based execution
• EMR on EKS, for deeper integration with Kubernetes-based infrastructure

Given EverC's existing investment in Kubernetes and its desire to reuse compute resources efficiently, EMR on EKS emerged as the best fit.

Why EMR on EKS

Running EMR on EKS allowed EverC to:

• Reuse existing Kubernetes clusters instead of provisioning dedicated Spark infrastructure
• Take advantage of fine-grained scheduling and scaling
• Improve utilization of spot instances and reduce idle compute
• Eliminate the platform premium associated with managed Spark services

Crucially, this approach gave EverC more direct control over how and when compute resources were consumed — without forcing a radical change in how teams worked.

Preserving the Analyst Experience

A key requirement was ensuring that analysts could continue working productively. By incorporating EMR Studio, EverC retained a notebook-based interface that felt familiar to the previous infrastructure they were accustomed to. This minimized disruption while enabling the underlying platform shift.

The redesigned platform delivered meaningful improvements across several dimensions:

• Reduced platform costs by eliminating unnecessary service premiums
• Improved cost predictability, with clearer visibility into how workloads consume compute
• Operational flexibility, allowing different workload types to coexist on shared infrastructure
• Continuity for data teams, who were able to maintain established workflows

Beyond the initial migration, EverC now has a foundation that supports ongoing optimization — scaling resources up or down as needs change, without being locked into a single execution model.

EverC selected BigData Boutique for its deep specialization in large-scale data platforms and its long-standing expertise in designing and operating Spark-based workloads on AWS. As an AWS Partner, BigData Boutique brings hands-on experience working with Amazon EMR, Kubernetes-based architectures, and cost-optimized cloud-native data platforms in regulated, production environments.

Rather than promoting a single service or tool, BigData Boutique approached the engagement from an architecture-first perspective — helping EverC reason through trade-offs, evaluate AWS-native options, and design a platform aligned with real workload patterns, security requirements, and cost constraints.

The collaboration extended beyond initial implementation. BigData Boutique continues to support EverC with ongoing optimization and architectural guidance, ensuring the data platform remains efficient, flexible, and aligned with EverC's evolving business needs.

With EMR on EKS in place, EverC is well positioned to continue scaling its data processing capabilities while keeping costs under control. The platform provides the flexibility to support future workloads and the discipline to ensure infrastructure decisions remain aligned with business priorities.

BigData Boutique remains a trusted partner as EverC continues to refine and optimize its data architecture on AWS.