Tibica: AWS Performance Optimization

About the Customer

Tibica is a serverless application marketplace offering cloud-native tools for teams that want to skip the infrastructure overhead. Their product line covers headless content management, code versioning, and access control, all built on AWS and designed to scale without managing servers.

Challenge

In July 2023, Tibica came to Perfsys with a latency problem. Their serverless, event-driven architecture was doing exactly what it was supposed to: scaling automatically, keeping costs down, eliminating the need for dedicated infrastructure. But there was a catch.

Response times were slower than they should have been. And in a marketplace where dozens of workflow steps run in sequence, those small delays stack up fast. Tibica needed to know where the time was going — and whether there was a practical fix.

Solution

Perfsys ran a technical assessment of Tibica's Lambda functions and found something straightforward: almost all of them were running on x86_64 architecture, which was the default choice — but not the most efficient one for their workload.

AWS Lambda supports two processor architectures. The x86_64 option is the standard. The arm64 option runs on AWS Graviton2, a processor AWS designed specifically for cloud workloads. For functions doing lightweight work, such as API Gateway integrations and routing logic, Graviton2 consistently delivers better throughput at lower compute cost.

According to AWS's own benchmarks , arm64 can provide up to 19% better performance and 20% lower cost compared to x86_64 for the same workload.

Tibica's Lambda functions fit that profile exactly. The recommendation was clear: migrate to arm64.

Source: AWS Partner Network (APN) Blog

Implementation

The migration took three weeks. Perfsys updated Tibica's Lambda function configurations to target the arm64 architecture, validated the change across all affected API Gateway integrations, and confirmed that function behavior was identical post-migration.

No application logic was rewritten. No infrastructure was rebuilt. The change was at the configuration level — which is exactly what makes architecture-layer optimizations worth evaluating before reaching for more complex solutions.

Results

The numbers came in quickly after deployment.

Average Lambda response time dropped from 2.7 seconds to 2.2 seconds — a 17% improvement. The overall user workflow response time, which compounds Lambda calls across multiple steps, decreased from 53.8 seconds to 41.9 seconds, a 22% reduction. And because arm64 uses compute resources more efficiently, Tibica's Lambda costs dropped by approximately 20%.

For a marketplace where users move through multiple application steps in a single session, a 22% faster workflow is something they feel. And a 20% reduction in Lambda costs is a recurring saving that compounds every month.

Conclusion

A three-week configuration change delivered measurable performance gains and a permanent reduction in operating costs. No rewrites, no downtime, no migration complexity.

If your serverless applications are running on x86_64 and you haven't evaluated Graviton2 yet, it's worth the conversation. Contact Perfsys to discuss what an AWS Cloud Assessment could uncover in your environment.

You can also see how Perfsys approached a different kind of latency challenge for Tibica in the Serverless Latency Optimization case study — where direct DynamoDB queries through API Gateway cut GET request times by 65%.

See also:

• Tibica: Serverless Cloud Marketplace on AWS — how Perfsys built Tibica's full AWS serverless infrastructure from the ground up
• Tibica: Serverless Latency Optimization — 65% faster GET requests by eliminating Lambda cold starts with direct DynamoDB queries
• AWS Consulting & Development — services overview
• MVP Development Services — serverless-first MVP builds on AWS

Anastasiia Tokareva

Software Engineer