Health Services and health data are critical and very important. Companies are busy analyzing health data and providing support to online transaction processing in the same field. Rarely a Big data solution is found that is meant to ensure proper, efficient and timely processing of health data to improve health-related services. Research is pursued to find out the various requirements of the varied applications of health services of different scales. 

Health data is huge, contains different types of data, and is generated and transmitted fast with a requirement to be received and processed fast. There is no assurance that the data may not have any errors. A huge amount of data is to be processed correctly to respond to different queries. Query response time is very critical. The performance of a system to answer these queries is very important.  The data model also plays a role in improving or decreasing the performance of the queries. Depending on the performance, the data model must be modified. It is also to be found which Big data solution is a better choice to handle such queries efficiently. 

Hadoop Map-Reduce and Hive technologies both have been in the forefront of Big Data development and adoption. Hive provides the original SQL framework on Hadoop. It is a runtime Hadoop support structure that provides Hive Query Language (HQL). Using HQL various queries can be executed efficiently on the large volume of data set. Here we are discussing two Bigdata platforms, namely Hadoop Map-Reducand Hive to find which one is reliable and flexible enough to handle large volumes of medical data. The YCSB guidelines should be followed for that.  

Nowadays, scientists are willing to work on metrology, genomics, connectomics , and complex physics simulations requiring Bigdata analytics. The Big Data analytics is required as the sizes of the data sets are growing rapidly. These datasets are collected from information sources like sensor nodes, remote sensing aerial devices, software logs, cameras, microphones, radars and wireless sensor networks etc. Processing Exa and terabytes of data is very costly with respect to processing units. The main concept of Big Data processing is to divide the data into small units that can be processed and after processing, accumulate the results from all these units. The business intelligence uses descriptive statistics with large amounts of informative data to measure things or detect trends by using inductive statistics and concepts from nonlinear system identification. 

Several research papers have been published in the area of performance analysis of various Big data tools. In authors have presented Hive as an open-source data ware-housing solution built on top of Hadoop. Hive as a tool that provides a system catalog which keeps metadata about files within the system.This allows Hive to function as a traditional warehouse which can interface with standard reporting tools like Micro Strategy. High-performance Integrated Virtual Environment (HIVE)  is proposed to have a robust architecture to perform next generation sequence data analysis. It is stated to be very robust and flexible due to the introduction of the abstraction layer in  between computational requests and operating system processes. It integrates metadata into object oriented data model and thus it handles all types of data. The access control and permission system of the honeycomb data model is secured and hierarchical so that data access privileges can be set  easily in a fine granular manner without having multiple rules. Data can be retrieved in HIVE from multiple sources. Varied visualizations of the result are available with the opportunity to export the results generated to be analyzed further by other external systems. Data can be shared and collaborated across the system because of this data model.  Apache Hive is a data warehouse solution built on top of Hadoop for providing data analysis, summarization, and query on large set of data. It provides the original SQL framework on Hadoop and gives an abstraction layer on top of Map Reduce to make it easier for analysts and data scientists to query data on the Hadoop File System.HIVE can take data from Hadoop, HDFS (Hadoop File System), local file system and it can write data to all of these. HBase allows random read and writes. HIVE offers efficient data aggregation, analysis and ad hoc querying on huge amount of data set. HIVE converts HQL queries to Directed Acyclic Graph of Map-Reduce jobs. These Map Reduce jobs are executed in Hadoop. 

Custom Map Reduce scripts can be plugged into queries. data can be queried in custom formats by using existing IO libraries. It has a Metastore containing schemas and statistics that is used to explore data effectively.  It supports primitive data types within tables and supports array, map and nested structures. It works with text, binary and column-oriented file formats.