Executed a Seamless Migration to Google Cloud Within 4 Months
Case Study
The Client
OnlineSales.ai is a leading monetization and hyperlocal co-op marketing platform headquartered in San Francisco. Founded in 2017, the company has around 200 employees. It marketplaces brands across industry verticals to unlock exponential growth trajectories. The platform hosts a comprehensive app suite that converges advertising, marketing, and demand planning for retailers and brands. OnlineSales.ai is a versatile platform catering to retailers and brands of all sizes, offering a white-labeled and self-serve solution for activating brands across various channels.
Business Context
The retail media market is booming, offering retailers significant opportunities to engage customers through targeted advertising. To capitalize on this trend, OnlienSales.ai offers AI-driven solutions for retailers lacking in-house expertise, enabling them to display targeted ads in real-time. To deliver their services efficiently, they needed a better cloud environment.
Challenge
Due to latency and scalability issues, OnlineSales.ai struggled with its existing cloud service provider. The company sought to migrate to a better cloud environment to enhance ad targeting, improve operational efficiency, and enable scalability in a global market. They required a seamless migration with minimal disruption to business.
Opportunity
Niveus suggested migration to Google Cloud Platform (GCP) to improve service availability and scalability, drawn by its numerous benefits. This was a comprehensive migration aimed at boosting service availability, enhancing operational efficiency, and securing a scalable infrastructure for future growth.
Business Solution
Niveus supported OnlineSales.ai in building their architecture on Google Cloud, and eventually with the migration. The teams worked together to understand the various services, and Niveus helped choose the best combination for the business. Niveus used various database migration strategies at different stages, including migrating from RDS to Cloud SQL, AWS Elasticsearch to Elastic Cloud, and MongoDB on AWS VM to MongoDB Atlas.
- The requests coming from the users were routed to the application hosted in GCE/GKE through Global Load Balancer, along with Cloud Armor being used for filtering malicious attacks and Cloud DNS being used for DNS resolution
- Google Cloud CDN was also configured with the load balancer to serve content closer to users, which accelerated the websites and applications
- The associated data from the application was then ingested into the persistent layer that consisted of Cloud SQL, Memorystore Redis, and Apache Solr with RDS migration to Cloud SQL
- The Push logs/commands from the application layer were transported to the ETL layer that comprised confluent Kafka and Cloud Pub/Sub, which further loaded the required data into the transformation layer that included 200 offline services
- The transformation layer further made API requests to the load layer and also uploaded required CSV data in a batch process
- Cloud Composer was configured for data orchestration and for batch upload of data to BigQuery
Key Value Delivered
- Executed a seamless migration to GCP within a 4-month timeframe
- Ensured seamless support during the cloud transition, achieving zero downtime in a complex migration process
- Streamlined the data pipeline using BigQuery and Dataproc, significantly enhancing Reporting API response times and overall efficiency
- Leveraged Dataproc to cut down engineers’ provisioning tasks by two days, reallocating their focus towards solution development
- Optimized the ad server, achieving notably improved response times
Business Impact
60%
Faster AI analytics for targeted advertising
45%
Reduction in ad placement time
30%
Reduction in ad server response time.
Customer Testimonial
Technology Stack
Cloud Storage
Compute Engine
Cloud Memorystore
Cloud SQL
BigQuery
Cloud Pub/Sub
Cloud Dataproc
Cloud Composer
Cloud CDN
Cloud Load Balancing
