Implementasi Highly Available Website Dengan Distributed Replicated Block Device

Abstract

As an important IT infrastructure, website is a system which requires high reliability and availability levels. Website meets the criteria as a highly available system because website must provide services to clients in real time, handle a large amount of data, and not lose data during transaction. A highly available system must meet the condition of being able to run continuously as well as guaranteeing consistency on data requests. This study designed a website with high availability. The approach was building network cluster with failover and replicated block device functions. Failover was built to provide service availability, while replicated block device provides data consistency during failure of service.  With failover cluster and replicated block device approaches, a cluster which is able to handle service failures of web server and database server on the website. The result of this study was the services of the website could run well if there was any failure in node members of the cluster. The system was able to provide 99,999 (five nines) availability on database server services and 99,98  (three nines) on web server services.