INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS www.ijrcar.com Vol.4 Issue 7, Pg.: 23-27 July 2016 S. Paul Steven, X. Arogya Presskila & Dr.K.Ramesh Page 23 INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS ISSN 2320-7345 BIG DATA SOURCE ANALYSIS DATA SERVICES IN GEO-DISTRIBUTED DATA CENTERS 1 S. Paul Steven, 2 X. Arogya Presskila, 3 Dr.K.Ramesh 1 S. Paul Steven, PG Scholar, E-mail: paulsteven92@gmail.com 2 X. Arogya Presskila, Assistant Professor, E-mail: presskila@gmail.com 3 Dr.K.Ramesh, Assistant Professor, E-mail: rameshk7n@yahoo.co.in Abstract: - With the launch of the database systems, several companies have been using it over the last few decades that were mainly driven by improvement in hardware, availability of an oversized quantity of data, collection of data at an infinite rate, rise in applications and so on. The sketch of data management systems is gently fragmented based on application domains (i.e., applications depending on relational data, graph-based data, and stream data). State-of-the-art commercial and educational systems use disk-related and in-memory operations. In this paper, an in-memory system is aimed for storing huge amount of data and serving faster accesses on database systems. In business operations, quickness is not an option, however a required one. Hence, each pathway in database systems is exploited to further improve performance, including decreasing dependency on the hard disk, increasing lots of memory to make a lot of data reside within the memory, and even setting up an in-memory system where large quantity of data’s will be hold on within the memory. Most commercial database vendors have recently launched in-memory database processing systems. Efficient in- memory data management may be a necessity for various applications. Nevertheless, in-memory data management remains at its early stage, and is probably going to evolve over the next few years. Keywords: Big Data, In-Memory System, In-memory Database, Efficiency, Memory 1. Introduction Big data give rise to several researches to develop systems for helping ultra-low latency service and real-time data analytics. Current disk-based systems can no longer offer timely response due to the high access latency to hard disks. The unsatisfactory performance was initially detected by internet companies such as Amazon, Google, Facebook and Twitter, but now it is becoming an hurdle for other companies/organizations that desire to provide a meaningful service (e.g., advertising, social gaming). For example, trading companies demand to detect a quick change in the trading prices and react immediately (in several milliseconds), which is no way to achieve using traditional disk-based processing/storage systems [1]. To fit the strict real-time requirements for evaluating mass amount of data and fulfilling requests within milliseconds, an in-memory system/ database that carries the data in the random access memory (RAM) all the time is necessary [3] In the last decade, multi-core processors and the vacancy of massive amounts of main memory at plummeting cost are creating new breakthroughs, making it feasible to build in-memory systems which is a significant part, if not the entirety, of the database fits in memory. For instance, memory storage size and