Case Studies > Tapad Excels in Cross-Platform Advertising with Sub-Millisecond Query Responses

Tapad Excels in Cross-Platform Advertising with Sub-Millisecond Query Responses

Customer Company Size

Mid-size Company

Region

America

Country

United States

Product

Aerospike
Java Virtual Machine (JVM)

Tech Stack

NoSQL Database
Key-Value Store
Java
Solid-State Drives (SSDs)

Implementation Scale

Enterprise-wide Deployment

Impact Metrics

Productivity Improvements
Customer Satisfaction
Innovation Output

Technology Category

Analytics & Modeling - Predictive Analytics
Application Infrastructure & Middleware - Database Management & Storage
Infrastructure as a Service (IaaS) - Cloud Storage Services

Applicable Industries

Software

Applicable Functions

Business Operation
Sales & Marketing

Use Cases

Real-Time Location System (RTLS)
Digital Twin

Services

System Integration

About The Customer

Tapad is a leader in cross-platform advertising technology, helping advertisers achieve a unified view of consumers and present consistent ads across multiple touch-points. The company maximizes a campaign’s impact across multiple screens by leveraging extensive real-time bidding (RTB) enabled supply and data relationships to target the right users, in the right context, on the right device. Tapad's bidding algorithms draw both on predictive learning and historic performance to calculate, in real time, the likelihood of an impression to meet client’s targeting parameters and achieve ROI goals. The result is an unprecedented level of targeting, relevance, and impact on a brand’s bottom line. Tapad relies on the Aerospike real-time NoSQL database and key-value store to power its decision engine and manage more than 150 billion ad impressions per month spread across almost 2 billion web browsers, mobile phones, tablets, and TVs.

The Challenge

For Tapad, the ability to predict the best placements and correct pricing for advertisers’ campaigns, as well as offer RTB to marketers across multiple channels, was critical to the success of its technology. To support this functionality, the company required a database that could reliably process and manage the data linked to billions of mobile display impressions and also respond to inquiries within milliseconds. The key-value store approach of NoSQL databases was ideally suited for Tapad’s needs, and the company began to investigate several well-known systems. However, Tapad found that despite offering high throughput, many products lacked the predictable response times needed to support the platform’s users. Only Aerospike stood up to the test of Tapad’s proof of concept, demonstrating the performance, reliability, and predictability required for their demanding production environment requirements.

The Solution

Tapad has five Aerospike server clusters deployed in its data center. The Tapad platform, which is built on a Java Virtual Machine (JVM), uses the Java interface of the Aerospike database to respond to queries by combining real-time data with information from the company’s other big data storage solutions. Aerospike consistently processes queries within sub-milliseconds, even as it processes some 50,000 transactions per second (TPS) per server. The predictable ultra-fast performance is critical because the 50 millisecond time window for each opportunity includes the network hops between other companies’ data centers, which gives Tapad only about 25 milliseconds to respond with an offer. Tapad’s rapidly growing business is driving the demand to manage ever larger amounts of data. Recently, Tapad began running large new campaigns with a nearly all-new user base. At Tapad’s request, the Aerospike support team came in to evaluate whether the current system could handle the new load. Testing revealed that while Tapad had the necessary read/write capacity, there was not enough storage disk space. The Aerospike advisors then recommended adding solid-state drives (SSDs) to meet Tapad’s growing storage demands.

Operational Impact

Tapad's platform uses Aerospike to respond to queries by combining real-time data with information from the company’s other big data storage solutions.
Aerospike consistently processes queries within sub-milliseconds, even as it processes some 50,000 transactions per second (TPS) per server.
The predictable ultra-fast performance is critical because the 50 millisecond time window for each opportunity includes the network hops between other companies’ data centers, which gives Tapad only about 25 milliseconds to respond with an offer.
Tapad’s rapidly growing business is driving the demand to manage ever larger amounts of data.
Aerospike’s advisors played an important role in counseling the company on its storage upgrade, but the value is also in the day-to-day interactions.

Quantitative Benefit

Aerospike processes some 50,000 transactions per second (TPS) per server.
Tapad manages more than 150 billion ad impressions per month spread across almost 2 billion web browsers, mobile phones, tablets, and TVs.
The 50 millisecond time window for each opportunity includes the network hops between other companies’ data centers, which gives Tapad only about 25 milliseconds to respond with an offer.