Caulis replaces Redis with Aerospike to deliver sub-100ms real-time fraud detection at scale
About Caulis
Caulis is a cybersecurity company based in Japan, providing fraud detection as a service to banks, utilities, and enterprise customers. Its flagship product, FraudAlert, offers real-time identity and behavior-based fraud detection across multiple touchpoints, including login, money transfers, phishing attempts, and more. Caulis has also worked with Japanese government initiatives such as the FATF and Regulatory Sandbox System to advance cross-industry data sharing and secure digital infrastructure. In March 2024, Caulis successfully went public.
Real-time fraud detection with unreliable infrastructure
Caulis needs to deliver real-time fraud alerts across critical sectors like finance and utilities. But its Redis-based system couldn’t support the sub-500ms SLAs without compromising reliability or uptime. Ongoing incidents and instability made it clear that a foundational change was needed.
Redis could not scale without downtime
Redis Cluster 3.x and 4.x suffered from instability and lacked the ability to scale in or out without service disruption—impacting uptime and performance.
Backup reliability and restore issues
Frequent backup corruption and long recovery times undermined trust in system resilience.
Complex operations and maintenance
Redis clustering was difficult to manage. Despite expert consultation, internal teams struggled to stabilize and scale the system effectively.
Real-time demands of fraud detection
Caulis needed consistent, sub-500ms latency to provide real-time alerts for banks and other institutions—delays meant increased risk of financial loss.
A scalable, high-performance NoSQL platform
Caulis evaluated Aerospike alongside Aurora and DynamoDB. While all three offered managed services and support, Aerospike stood out for its cost performance, operational simplicity, and real-time speed at scale.
Matched Redis speeds with greater stability
Aerospike delivered Redis-level performance in all internal tests, while avoiding Redis’s failure points.
Scalable without downtime
Enabled both horizontal and vertical scaling with zero service interruption.
Easy to operate, restore, and back up
Aerospike simplified maintenance, enabled reliable backups, supported fast deployment using mirrored AWS traffic, and provided responsive enterprise support.
Cost-effective fixed pricing
Predictable pricing based on data size made Aerospike financially sustainable for Caulis, which was still evolving its product-market fit at the time.
Faster, safer, and simpler fraud detection
With Aerospike, Caulis stabilized its infrastructure, improved detection speed, and streamlined operations to focus on growth and innovation.
Consistent sub-100ms response time
Average latency now ~100ms with 99.9% of requests under 500ms, enabling real-time alerts to financial institutions.
System stability and confidence restored
Aerospike’s performance and reliability eliminated Redis-related incidents and backup failures. Production migration was completed seamlessly with zero downtime by validating Aerospike’s behavior using mirrored AWS traffic before cutover.
Simplified operations and maintenance
Teams are now focused on product innovation, not firefighting database issues.
Lower TCO with better scalability
Aerospike’s fixed pricing and efficient infrastructure enabled Caulis to grow while staying cost-effective.
Testimonials
Additional resources
For a deeper understanding and more insights, explore these additional resources.
- Blog
Optimizing Fraud Detection: Caulis’s Redis to Aerospike Migration
Discover how Caulis enhanced fraud detection by migrating from Redis to Aerospike. Learn about the challenges, solutions, and benefits of this strategic move. - Video
From Redis to Aerospike: Supercharging fraud detection with scalable NoSQL
Yasuyuki Hayashi of Caulis shares the move from Redis to Aerospike for FraudAlert, boosting scalability, reliability, and cost-effectiveness in real-time fraud detection.
Get started with Aerospike
For high-performance, scalable data management and ultra-low latency, ideal for handling massive datasets and real-time applications.
Caulis replaces Redis with Aerospike to deliver sub-100ms real-time fraud detection at scale
About Caulis
Caulis is a cybersecurity company based in Japan, providing fraud detection as a service to banks, utilities, and enterprise customers. Its flagship product, FraudAlert, offers real-time identity and behavior-based fraud detection across multiple touchpoints, including login, money transfers, phishing attempts, and more. Caulis has also worked with Japanese government initiatives such as the FATF and Regulatory Sandbox System to advance cross-industry data sharing and secure digital infrastructure. In March 2024, Caulis successfully went public.
Real-time fraud detection with unreliable infrastructure
Caulis needs to deliver real-time fraud alerts across critical sectors like finance and utilities. But its Redis-based system couldn’t support the sub-500ms SLAs without compromising reliability or uptime. Ongoing incidents and instability made it clear that a foundational change was needed.
Redis could not scale without downtime
Redis Cluster 3.x and 4.x suffered from instability and lacked the ability to scale in or out without service disruption—impacting uptime and performance.
Backup reliability and restore issues
Frequent backup corruption and long recovery times undermined trust in system resilience.
Complex operations and maintenance
Redis clustering was difficult to manage. Despite expert consultation, internal teams struggled to stabilize and scale the system effectively.
Real-time demands of fraud detection
Caulis needed consistent, sub-500ms latency to provide real-time alerts for banks and other institutions—delays meant increased risk of financial loss.
A scalable, high-performance NoSQL platform
Caulis evaluated Aerospike alongside Aurora and DynamoDB. While all three offered managed services and support, Aerospike stood out for its cost performance, operational simplicity, and real-time speed at scale.
Matched Redis speeds with greater stability
Aerospike delivered Redis-level performance in all internal tests, while avoiding Redis’s failure points.
Scalable without downtime
Enabled both horizontal and vertical scaling with zero service interruption.
Easy to operate, restore, and back up
Aerospike simplified maintenance, enabled reliable backups, supported fast deployment using mirrored AWS traffic, and provided responsive enterprise support.
Cost-effective fixed pricing
Predictable pricing based on data size made Aerospike financially sustainable for Caulis, which was still evolving its product-market fit at the time.
Faster, safer, and simpler fraud detection
With Aerospike, Caulis stabilized its infrastructure, improved detection speed, and streamlined operations to focus on growth and innovation.
Consistent sub-100ms response time
Average latency now ~100ms with 99.9% of requests under 500ms, enabling real-time alerts to financial institutions.
System stability and confidence restored
Aerospike’s performance and reliability eliminated Redis-related incidents and backup failures. Production migration was completed seamlessly with zero downtime by validating Aerospike’s behavior using mirrored AWS traffic before cutover.
Simplified operations and maintenance
Teams are now focused on product innovation, not firefighting database issues.
Lower TCO with better scalability
Aerospike’s fixed pricing and efficient infrastructure enabled Caulis to grow while staying cost-effective.
Testimonials
Additional resources
For a deeper understanding and more insights, explore these additional resources.
- Blog
Optimizing Fraud Detection: Caulis’s Redis to Aerospike Migration
Discover how Caulis enhanced fraud detection by migrating from Redis to Aerospike. Learn about the challenges, solutions, and benefits of this strategic move. - Video
From Redis to Aerospike: Supercharging fraud detection with scalable NoSQL
Yasuyuki Hayashi of Caulis shares the move from Redis to Aerospike for FraudAlert, boosting scalability, reliability, and cost-effectiveness in real-time fraud detection.
Get started with Aerospike
For high-performance, scalable data management and ultra-low latency, ideal for handling massive datasets and real-time applications.