SR&ED Case Study: Distributed Database Scalability

Background

A SaaS platform’s database architecture, designed for a single-region deployment, began experiencing intermittent data consistency errors as the company expanded to multi-region deployment to meet customer latency requirements.

The Challenge

The team needed a replication approach that maintained strong consistency for critical financial records while keeping read latency low across regions — a combination standard off-the-shelf replication configurations could not achieve for this workload.

Technological Uncertainty

It was not known in advance whether any replication strategy could meet both the consistency and latency requirements simultaneously for this specific transaction pattern, or whether a fundamental trade-off would make one requirement unachievable.

Experimental Development

The team systematically tested multiple consistency models and conflict-resolution strategies under simulated multi-region load, measuring latency, consistency violations, and failure recovery behaviour for each configuration.

What Failed?

An initial eventual-consistency approach met latency targets but produced consistency violations on financial records under concurrent writes, and a subsequent strict-consistency configuration eliminated violations but pushed latency above acceptable limits for two regions.

Technological Advancement

The team developed a hybrid consistency model that applied strict consistency selectively to financial-record operations while allowing eventual consistency elsewhere, generating new architectural knowledge specific to the platform’s transaction patterns.

Potentially Relevant SR&ED Activities

  • Systematic testing of consistency models under simulated multi-region load
  • Development of a selective, hybrid consistency architecture
  • Failure-recovery and latency benchmarking across iterations

What Would Generally Not Qualify

Deploying the finalized architecture to additional regions using the same, already-validated configuration would not itself involve further eligible experimental development.

Documentation

Load-testing results across configurations, architecture decision records documenting the trade-offs considered, and incident postmortems from the failed approaches would support this claim.

About The Author

Dale Doering

Dale Doering is the owner of SRED Consultants Inc., helping businesses navigate the complexities of Scientific Research and Experimental Development (SR&ED) claims. With a strong understanding of the technical and interpretive requirements of the SR&ED program, Dale works with companies to identify eligible projects, document technological challenges, and clearly demonstrate the systematic experimentation or analysis undertaken to achieve advancement. His approach focuses on translating complex technical work into well-supported SR&ED claims, helping clients maximize eligible opportunities while maintaining a clear understanding of the program’s requirements.

Recent Posts

Frequently Asked Questions

What core problem was the engineering team trying to solve in this project?

The team needed to optimize their database architecture for multi-region expansion. Specifically, they were trying to achieve strong data consistency for critical financial records without sacrificing the low read latency required by customers across regions—a dual requirement that standard off-the-shelf database replication tools couldn’t achieve.

The technical uncertainty centered around whether any replication strategy could simultaneously meet both strict consistency and low-latency demands for the platform’s specific high-concurrency financial workload, or if standard architectural trade-offs would make one requirement impossible to achieve.

The team tested two primary standard configurations:

– Eventual Consistency: Met the low-latency targets but caused data consistency violations on financial records during concurrent writes.

– Strict Consistency: Successfully eliminated data errors but pushed read latency past acceptable performance thresholds in two geographic regions.

The team systematically developed a novel hybrid consistency model. This model dynamically applies strict consistency to high-risk financial transactions while allowing eventual consistency for non-critical platform data—generating new architectural knowledge specific to their transaction patterns.

– Qualifying Activities: Systematic load testing of consistency models, designing and benchmarking the hybrid consistency architecture, and analyzing failure-recovery/latency trade-offs during experimental iterations.

– Excluded Activities: Routine deployment of the validated hybrid architecture to additional regions using the same established setup (as it no longer contains technical uncertainty).

Related Post