Earticle

Cloud computing technology relieves the maintenance burden of corporate IT infrastructure, which is the highly expected area for cost saving to be achieved through corporate IT technology innovation. But logically more monitored items occur especially in cloud service using virtualization technology. In reality, monitoring tool is increasingly gaining more attention because of its capability of monitoring service quality, as virtualization-based equipments not seen physically. The system developed in this paper, allows for real-time total monitoring of all elements that configure services by implementing Agentless total computing monitoring system for comprehensive management of IT infrastructure service, as well as for service quality monitoring with Replay function through monitoring log analysis and Big Data analysis on resources traffic and performance data. Furthermore, Android application was developed in order to provide monitoring and alarming services.

데이터센터를 안정적이고 효율적으로 운영하기 위해서는 서버관리와 전력관리가 중요하다. 본 논문에서는 데이터센터의 서버관리 인터페이스와 전력관리 모듈을 이용하여 에너지 사용량을 모니터링 및 시각화하고, 서버의 온도와 각 장치의 작동상태, 전원상태와 같은 서버 모니터링 정보를 제공하여 데이터센터의 에너지소비분석에 도움을 주는 웹 기반 애플리케이션에 대한 소개와 구현방법을 제안한다. 이 애플리케이션은 서버관리표준 인터페이스인 IPMI 뿐만 아니라 제조사의 독자적인 서버관리기술을 사용하여 이기종 서버에 대한 통합 모니터링이 가능하며, 모니터링 대상서버 시스템에 별도의 에이전트 설치가 필요하지 않으므로 모니터링에 따라 시스템에 발생할 수 있는 부하가 거의 없고, 실시간으로 에너지 소비량과 모니터링 정보를 제공하기 때문에 데이터센터의 에너지소비패턴분석과 서버관리에 효과적으로 사용할 수 있다.

Server management and power management are important to operate a data center stably and efficiently. By this paper, we introduce an web based application that is able to monitor and visualize energy consumption, help to energy consumption analysis of a data center due to provide server monitoring information such as temperatures for server, status of each device, power status using server management interfaces and power management modules in data center, also suggest the methods to implement them. This application takes advantage of IPMI which is server management standard interfaces and server management technology in manufacturer's individual way so it can do integrated monitoring for heterogeneous severs, and there is little monitoring load inside of server system because it doesn't need to install agent program for monitoring target system, and it can be used successfully to energy consumption analysis, server management in a data center due to realtime provided energy consumption and monitoring information.