This paper aims at implementing an efficient sensor middleware for maintaining and managing big data and analysing the efficiency of such a sensor middleware. The work focused on designing of sensor middleware and performance analysis of the sensor middleware. The Hadoop framework, which was used as a sensor middleware, was responsible for maintaining and managing the distributed sensor environment. A prominent monitoring tool such as Zabbix was used to monitor the Hadoop processes and provide alerts for custom trigger events. Performance analysis for the sensor middleware such as Hadoop was done using tools such as collectl and colplot. Performance analysis of Hadoop cluster was performed using an evaluation metric such as total runtime of the job. A brief discussion on the Hadoop tuning tool such as starfish, that makes the Hadoop cluster efficient through configuration changes, is also presented.