About the Customer:
SuperPlay is a leading Israeli gaming company that develops and publishes mobile games for a global audience. The company specializes in creating engaging social casino and casual gaming experiences, reaching millions of players worldwide. As a technology-driven gaming company, SuperPlay relies on robust cloud infrastructure to deliver seamless gaming experiences to their users, leveraging AWS services including Amazon Elastic Kubernetes Service (EKS), Amazon EC2, and Amazon ElastiCache for Redis as their core infrastructure components
Customer Challenge:
SuperPlay, a gaming company specializing in real-time mobile experiences, relies heavily on high-performance data operations to support fast-paced interactions and a seamless user experience. To meet the low-latency demands of their applications, SuperPlay chose Amazon ElastiCache for Redis over traditional relational databases like RDS after benchmarking revealed significantly faster query response times with Redis for their data volumes.
The application layer operates on Amazon EKS running on EC2 instances and connects to ElastiCache using Redis client libraries optimized for low-latency communication (such as `redis-py` or `ioredis`, depending on the language stack). This setup allowed SuperPlay to deliver the real-time responsiveness required for their gaming events.
However, as the user base grew rapidly, so did the volume of data and the number of queries. To maintain performance, SuperPlay had to scale out their ElastiCache cluster by adding more nodes, which led to a sharp increase in infrastructure costs. Despite leveraging Reserved Instances, the growing cost impact made the current approach unsustainable. SuperPlay faced a critical need to optimize their Redis strategy without sacrificing performance, ensuring continued scalability and cost-efficiency as their game traffic expanded.
Partner Solution:
GlobalDots implemented a three-step strategy to curb rising costs:
Introduction of ValKey as a Redis alternative:
The first step involved migrating from Redis to ValKey, a more cost-effective solution. Since ValKey supports all Redis commands and allows live replication between ElastiCache Redis and ElastiCache ValKey, the migration process was relatively straightforward.
Scaling smarter, not wider:
Upon investigation, we found that the primary driver for scaling out was connection limits. Rather than adding more nodes, we opted to scale up – upgrading existing nodes to more powerful instances. This approach addressed the bottleneck more efficiently and significantly reduced overall costs.
Optimizing Reserved Instances:
In the final step, we upgraded the Reserved Instances from ElastiCache Redis to ElastiCache ValKey, aligning the reservations with current usage. Long-term commitments were made only after confirming stable performance and predictable usage patterns, ensuring cost efficiency over 1- or 3-year terms.
Throughout the migration, GlobalDots took full ownership of monitoring costs and usage, enabling precise commitments and optimal selection of instance types and sizes. This ensured the best balance of performance and price for the system’s needs.
Results and Benefits:
Usage hours between the two periods increased by 92.5%. However, the daily average cost increased by only 4% during this period, thanks to the Redis-ValKey migration and a more focused, aggressive, and effective reservation orchestration, as well as choosing wisely the node’s instance size.
Figure 2+3: The subtle rise in costs, even when the CPU consumption almost doubled
About the Partner:
Leveraging the Power of AWS Partnership, this case study demonstrates that cost savings can be achieved without compromising performance. In fact, the expert team at GlobalDots prides itself on delivering solutions that help our partners simultaneously reduce spend and elevate their product and service offerings.
