Document Type

Article

Publication Date

Summer 7-1-2026

Abstract

This paper presents the design and evaluation of a Distributed Ticket Booking System (DTBS) built to handle extreme concurrency in high-traffic digital environments. Motivated by real-world failures such as the 2022 Taylor Swift Eras Tour sale—where 14 million simultaneous requests overwhelmed Ticketmaster—the system addresses the fundamental conflict between strong consistency and high availability under massive demand. The proposed solution adopts the Command Query Responsibility Segregation (CQRS) pattern to decouple read and write operations, and applies a hybrid consistency model grounded in the CAP theorem: a CP-oriented booking subsystem with Redis-based distributed locking, and an AP-oriented search subsystem using Elasticsearch for low-latency event discovery. Apache Kafka manages asynchronous messaging and a virtual queuing mechanism for fair traffic serialization during demand surges. Comprehensive functional and load testing validates the design. Elasticsearch reduces p99 search latency by 76% compared to direct PostgreSQL queries. The booking service sustains 700 requests per second under 500 virtual concurrent users with a median response time of 168 ms. All 11 functional test cases pass, confirming zero double-booking under simulated contention. The results demonstrate that context-specific application of CAP trade-offs enables a system that is both highly performant for reads and strictly consistent for writes.

Program or Discipline Name

Computer Information Sciences

Download

Share

COinS