A High Performance Computing Web Search Engine Based on Big Data and Parallel Distributed Models

Abstract

This paper presents a high-performance web search system leveraging big data technology. Utilizing a heterogeneous architecture and a parallel distributed computing model based on the MapReduce framework, the system significantly enhances efficiency, scalability, and reliability. The design includes a storage management scheme that integrates cloud storage and grid computing technologies, facilitating efficient storage and rapid access to large-scale data. Key components such as an inverted index structure, vector space model, and semantic analysis models are employed to implement functionalities across the data, logic, and display layers. An experimental environment was set up on the Microsoft Azure cloud platform using the Common Crawl dataset for testing. Performance evaluation, based on metrics including response time, accuracy, and stability, demonstrates the system's superior performance compared to two existing systems, thereby validating its effectiveness.