Owing to the explosion of data in today’s world, datasets are enormous, geographically distributed and heterogeneous. Data mining aims extracting useful information from voluminous repositories where data is stored. Predictive analysis of hidden patterns in massive datasets poses to be a challenge. The problems faced while using the data warehousing model for such datasets were privacy, centralization of the data present at multiple independent sites, bandwidth limitation, complexity of integration, and analysis of the data at a global level. Distributed algorithms have been designed to address the same. Distributed data mining (DDM) techniques regard the distributed datasets as one virtual table and assume the existence of a global model which could be designed if the data were combined centrally. This paper presents distributed data mining systems and frameworks for analyzing data and mining the required knowledge from it. Emphasis has been laid on the architectures of such models. Factors like computation resources, communication, hardware, and usage of distributed resources of data have been considered while analyzing or designing distributed algorithms. Such algorithms primarily aim at memory expense and average distribution of working load. Distributed data finds its application in e-commerce, e-business, intrusion detection systems, and sensor networks.