PFM Case Study
PFM is developing a novel visitor count software suite which uses AWS for backend processing
PFM Case Study
Company Description PFM is developing a novel visitor count software suite which uses AWS for backend processing. Anonymized visitor data is particularly valuable to retailers, museums and municipal councils along with other organizations. Councils use this data to determine whether their policies and strategies are having the desired effect. Retailers can determine whether their locations are still in the right place and whether their locations are performing as expected.
Problem: PFM had built an architecture based on an Application Load Balancer, Redis memory cache and an array of EC2 instances feeding data into a large RDS instance. Whilst this worked it, the recurring monthly costs were significant and it was difficult to scale.
Solution: 9apps proposed a solution based on using the serverless technologies Lambda and DynamoDB.
Serverless technologies differ from conventional cloud instances in that charges are based on cpu and memory resources used instead of the time the cloud instance is running. This lead to a direct correlation in monthly AWS charges and the number of footfalls detected by PFM’s WiFI detectors. In the present situation their outgoings are consequently much less and will stay at a low level until normality returns. This design reduced monthly costs by 70%.
These technologies also offer a scalable, future proof architecture for the collection, aggregation and reporting of data. In addition to these goals the architecture will expedite product and feature development on top of an optimized uniform product platform.
The following AWS Services were used in the design:
- API Gateway
- Route 53
- Cloud Formation