Análisis e implementación de alta disponibilidad mediante clustering en sistemas de call center basados en VoIP.

Abstract

The High-Availability Analysis and Implementation through Clustering in Call Center Systems (CC) based on Voice over IP protocol (VoIP) was carried out to minimize the no-access services provided by them. A prototype was implemented in the virtual environment using the emulator of computers VMware Workstation; the scenery was made up of three servers in which the operative system CentOS, Elastix for the VoIP environment, Heartbeat for the communication between nodes, Pacemaker as a Resources Administrator of the Cluster and Rsync for the file synchronization, were installed. The cluster was subjected to events that cause migration of services to collect data of system recovery times, showing that in cluster scenery the no-access time is 45 seconds, which implies a high availability percentage of 99,997%; on the other hand, in no-cluster sceneries, the times depend on different situations; taking as an example, in the case in which it takes 3 days to survey the complete system 83,333% availability is reached. It was determined that the use of High-Availability clusters permits to obtain systems tolerant to faults, avoiding the fact that the bad functioning of the hardware or software effects the provided service and, above all it permits to minimize the perception of such faults by the users.