Deploying automatic speech recognition system for scalability, reliability, and security with Kubernetes

• Main Author: Tjandy Putra
• Other Authors: Chng Eng Siong
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/171933
• Institution: Nanyang Technological University

This project undertakes the task of refining the existing Automatic Speech Recognition (ASR) system’s deployment, orchestrated by Kubernetes on the cloud. While focusing on reliability, scalability, and security, it also aspires to maintain a balanced approach to cost-effectiveness. In pursuit of enhanced data and processing reliability, Apache Kafka has been considered. For bolstering security measures, the incorporation of technologies such as Kyverno and Falco has been explored. Kyverno serves to enforce adherence to essential cluster rules, aiming to mitigate human-induced discrepancies, whereas Falco is introduced with a vision to provide cluster system administrators with potential insights into any unforeseen malicious activities. Beyond these solutions, the report also examines other elements that play pivotal roles in enhancing the overall architecture. The document seeks to elaborate on these modifications, offering a detailed perspective on how each element collaboratively contributes to the system’s advancement.