Case Study: GumGum slashes AWS infrastructure costs and halves deployment time with Spot

GumGum Utilizes Spotinst To Help Drastically Reduce Costs Without Compromising Availability Or Scalability

GumGum, a company that applies computer vision across advertising, sports and social media, faced rapidly rising AWS infrastructure costs as demand grew — their Ad Server handled 4 million requests/minute at low traffic and spiked to over 18 million, requiring more than 130 c5.9xlarge instances and about $120K/month in EC2 spend in a single datacenter. After finding Reserved Instances unsuitable for highly variable peak loads, GumGum turned to Spot (Spotinst) to leverage Spot Instances while preserving uptime.

Spot provided a fully managed Spot Instances service with predictive termination handling, an SLA-backed availability guarantee, and a simple console, plus default Blue/Green deployments for CodeDeploy. Using Spot’s solution allowed GumGum to scale flexibly on deeply discounted capacity without compromising availability, and the Blue/Green workflows cut deployment time by over 50%, increasing developer capacity and delivering significant cost reductions.

GumGum

Vaibhav Puranik

SVP of Engineering