Event Management Server for Real-Time Ad Bidding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online advertising campaigns face challenges in real-time data access from various sources, which hinders efficient bidding in auctions for advertisement slots, particularly in personalized retargeting, leading to delayed user retargeting and reduced campaign effectiveness.
Innovation Solution
A system architecture that includes an event management server processing and normalizing data from user devices, partitioning it by time thresholds, and storing it in a fast key-value events datastore, allowing a bidder server to generate bids based on recent user events within a predetermined time range, ensuring real-time data availability for auction bidding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If real-time access to data from different sources is implemented, then bid efficiency is improved, but system complexity increases
Solution Approach 1:
The system segments data from different sources into separate event streams, each processed independently through normalization and partitioning. This allows the complex task of handling multiple data sources to be broken down into manageable components, improving bid efficiency while controlling system complexity through modular architecture.
Solution Approach 2:
An event management server acts as an intermediary between data sources and the bidding system. This mediator normalizes, partitions, and stores event data in a key-value datastore, enabling real-time access for bidding while abstracting the complexity of data collection and processing from the bid generation process.
2Speed
If data is processed and stored in real-time, then user retargeting speed is improved, but data processing complexity increases
Solution Approach 1:
The system performs preliminary processing of event data by normalizing formats and partitioning by time thresholds before storage. This pre-processing prepares data for rapid retrieval during bidding, improving retargeting speed while managing processing complexity through automated batch operations rather than real-time complex computations.
Solution Approach 2:
The system transforms event data into a standardized key-value format with time-based partitioning parameters. This parameter transformation simplifies data structure and enables fast retrieval, improving processing speed while reducing the complexity of handling diverse data formats from multiple sources.
3Productivity
If event data is partitioned by time thresholds, then data retrieval efficiency is improved, but storage complexity increases
Solution Approach 1:
The system segments event data into time-based partitions using key-value storage. This segmentation allows efficient retrieval of recent events by time threshold while maintaining storage organization through simple key-based indexing, improving retrieval efficiency without significant storage complexity.
Solution Approach 2:
The system transforms event data into a simplified key-value format where time thresholds become storage keys. This parameter transformation enables efficient data retrieval based on time ranges while keeping the storage structure simple and manageable, avoiding the complexity of traditional relational databases.
Data Source
AI summary
This disclosure describes systems, methods, and computer-readable media related to online advertisement campaign recommendations. A server may obtain data from an events datastore indicative of missing time slices from a partition of events data. The server may obtain first events data for the missing time slices. The server may identify redundant events data in the first events data that is redundant to second events data stored in the events datastore. The server may delete the redundant events data from the first events data to generate a modified first events data. The server may transmit the modified first events data to the events datastore.