WEBSOCKETS VS. HTTP POLLING: IMPLICATIONS FOR HIGH-FREQUENCYDATA STREAMING APPLICATIONS

Abstract

This study investigates the comparative performance of WebSockets and HTTP Polling for highfrequency data streaming applications within Nigerian network conditions. Focusing on key metrics such as latency, cost implications, and network resilience, the research employs an experimental design to simulate real-time scenarios relevant to industries like fintech and online customer support. Using Node.js for backend implementation and WANem for network simulation, the study analyzes performance data through regression techniques. Findings reveal that WebSockets significantly outperform HTTP Polling, demonstrating lower latency (p = 0.000) and greater network resilience (p = 0.000). Furthermore, the cost implications analysis indicates that HTTP Polling incurs higher expenses (p = 0.012), primarily due to repeated connection overheads. With 73.1% of latency variation and 68.9% of network resilience variance explained by protocol type, the study underscores the advantages of adopting WebSockets for applications requiring real-time updates. Recommendations include promoting WebSockets as the default protocol for high-frequency applications, optimizing network configurations, and encouraging infrastructure investments to enhance overall user experiences across Nigeria. This research contributes to the understanding of effective communication protocols in regions with variable internet conditions, providing valuable insights for developers and policymakers alike