AN OVERLAY-BASED DATA MINING ARCHITECTURE TOLERANT TO PHYSICAL NETWORK DISRUPTIONS

-Big data generated from various aspects like online transactions, social websites, logs and search queries is increasing rapidly and thus the demand for data mining has risen as a noteworthy zone. An overlaybased parallel information mining executes completely dispersed information administration and handles processing by utilizing the overlay system, which can achieve high flexibility. The talk incorporates a survey of best in class systems and stages for preparing and overseeing huge information and also the endeavours expected on enormous information mining. Nonetheless, the overlay based parallel mining structural planning is not fit for achieving data mining administrations if there is an occurrence of the physical system disturbance that is created due to switch/correspondence line breakdowns on the grounds that various hubs are expelled from the overlay system. To get the estimated arrangement and better results, the proposed framework utilizes K-medoids algorithm for cluster formation and overlay based system. Proposed work gives enhancement in terms Energy Consumption in data gathering, reduced delay and Node Coverage