Skai Cuts Hardware Needs by 50% and Improves Real-Time Insights with SingleStore on AWS

Brands and agencies use Skai, an omnichannel demand-side advertising platform, to manage campaigns across search engines, social media, retail media networks, and mobile apps—even in walled garden environments. Clients can oversee multiple campaigns, analyze performance in real time, and make fast, data-driven decisions to maximize their marketing impact. As a result, data plays a central role in the Skai platform and requires instant and reliable delivery. To enhance query speeds and responsiveness, Skai adopted SingleStore’s high-performance database, which significantly improved real-time analytics for its growing client base. However, as Skai expanded further, the limitations of its on-premises infrastructure became clear. The solution, which relied heavily on in-memory features and an automated backup strategy for disaster recovery, became increasingly expensive to maintain.

Skai’s systems struggled to keep up with the platform’s expanding requirements. Scaling hardware required extensive advanced planning, often resulting in costly over-provisioning or delays in feature rollouts. According to Ziv Fridfertig, data infrastructure manager at Skai, managing this setup was both costly and time-consuming: “Our team was spending up to 30 percent of its time just managing infrastructure. We needed a way to focus on innovation rather than hardware cycles.” Performance limitations also threatened the seamless, responsive experience clients expected, while demand for real-time data processing grew. To keep up with these demands, Skai needed a scalable, high-performance data platform that could support rapid scaling, real-time analytics, and ease of maintenance—without the need for constant hardware upgrades.

To deliver faster insights, scale operations seamlessly, and reduce maintenance overhead, Skai migrated to a cloud-based, multi-tenant architecture built on SingleStore and AWS. Skai selected SingleStore—which handles SQL, JSON, full-text, and vector workloads in one unified platform—to support its need for high-speed data processing. As a high-performance, distributed SQL database, SingleStore rapidly aggregates and analyzes millions of metadata rows and billions of event records, delivering real-time insights for clients. Because SingleStore is MySQL compatible, Skai established a hybrid setup: MySQL handled write operations as the primary source of truth, while SingleStore powered read-intensive, client-facing activities. To maintain near-real-time data flow between these databases, Skai implemented a streaming pipeline built on Kafka. Each partition in this pipeline processes thousands of rows per second, ensuring seamless data integration and rapid access for end users.

AWS provided the reliable foundational infrastructure to support Skai’s cloud migration. To optimize storage and compute resources, Skai uses Amazon Elastic Compute Cloud (Amazon EC2) instances for cost-effective performance of memory-intensive workloads. For storage, Skai implemented high-throughput Amazon Elastic Block Storage (Amazon EBS) to improve data access speeds. Skai also configured a highly available disaster recovery solution with AWS multi-Availability Zones that included leaf replicas across different Availability Zones. This cloud-based configuration replaced Skai’s on-premises disaster recovery system, which relied on automated backups using Amazon Simple Storage Service (Amazon S3).

To enhance monitoring, SingleStore and Skai, together with one of their technology partners, integrated a comprehensive solution with custom metrics to proactively identify and address resource needs. Additionally, restructured data pipelines and simplified online cluster management help handle client migrations and adjust resource allocation with minimal operational disruptions. Overall, said Fridfertig, “With the integration of SingleStore’s high-speed data processing and the robust, scalable infrastructure of AWS, we have a resilient, efficient platform designed to handle the demands of rapidly growing client data.”

By migrating to AWS with SingleStore as its primary database, Skai transformed its infrastructure to achieve significant improvements in cost efficiency, performance, and client satisfaction. “One of the most impactful changes has been our ability to cut overall hardware requirements by 50 percent, which has translated into substantial savings,” said Fridfertig. “Prior to migrating to AWS, we operated 450 servers across 50 clusters. Today, we have reduced this to 210 servers and 25 clusters. This shift has allowed us to focus more on strategic projects and innovation, rather than routine maintenance.” With high-performance Amazon EBS storage, Skai achieved faster data access speeds for 45 percent of its workloads. In addition, it cut in-memory requirements by 80 percent, down from 12 TB to just 2 TB. As a result, queries achieved even better performance on AWS, with the support of SingleStore’s columnar engine and high-performance storage, compared to the previous on-site in-memory engine. Additionally, high availability across multiple AWS Availability Zones minimized disruptions during traffic spikes or outages for added resilience and reliability.

Operational efficiency also improved considerably. Freed from frequent maintenance demands, Skai’s team could prioritize strategic projects, accelerate feature deployment, and streamline internal processes. “This operational shift contributed directly to our growth,” Fridfertig noted. “Over the past four years, we’ve achieved a significant increase in growth and delivery, driven by our ability to innovate and respond quickly to client needs.” Skai’s clients quickly noticed the difference, expressing increased satisfaction with the platform’s responsiveness and data flexibility. Looking ahead, Skai plans to deepen its partnership with AWS and SingleStore by exploring advanced AI capabilities with vector databases and expanding batch processing on SingleStore. These initiatives position Skai to keep delivering value, adapting to client needs, and leading in the advertising technology market.