Volume 7 Issue 5 2017


Page Title Full Text
32-34 Analysing and Visualization of Customers Data in Online Shopping using Big Data tools
K.Rajesh and P.Sai Prasad

In this paper the authors presented that how to analyse the Customers data when it is huge and various raw format data. So, for this we have used the Hadoop Big Data tools map-reduce Pig and Hive which makes easier to analyse the various raw formats of data.

Keyword: Hadoop, Big Data, Map- reduce, Hive

Full Text PDF
35-40 Comparing VMware's DRS Migration with a Human Expert Migration
Kavyashree.V and Prasad A.Y

As the technology is growing rapidly, huge amount of data is being generated. The generated data can either be from simple devices like mobiles, weather, banks and many more. These data is loaded to the virtual environment. Virtualization provides a greater flexibility in terms of sharing up of resources. Next challenge that we face is balancing of load on such virtual machines (VM). VMware DRS is a tool that automatically balances the load in VMs. Here we consider large telecommunication application and measure the performance of host before and after the migration in terms of CPU utilization. Then we compare the results of VMware DRs migration with Human Expert. DRS could balance load in 40% of cases and in rest of the cases it could not balance the load as expected. In few cases it did unnecessary migration such as back and forth migration.

Keywords—Cloud Computing, Distributed resource Scheduler (DRS), Virtual Machine Migration, Virtualization, VMware

Full Text PDF
41-43 An Enhanced Lockstep Broadcasting in Big Data
Divyadharsini S and Krishneswari k

Transferring of bigdata in wsn is complex due to its size and energy consumption available at node. Also to  find the nearby transferring node in terms of neighbour node prediction. In our proposed system we find and estimate the neighbouring node by broadcasting signal to near node . it consumes low energy as compared to other techniques to compute neighbour. After neighbour computation we randomly select cluster heads and form the cluster based on the distance and energy metrics. After the clustering of environment we choose the data to transfer in the WSN . This network now split the bigdata into chunks of number of cluster avail. Then transfer the data..Due  to  limited resources  of  the  multi-sensor  system,  it  is  a  challenging  task  to reduce the energy consumption to survive a network for a longer period. Keeping in view the challenges above, this paper presents a novel technique of using a  hybrid algorithm for clustering and cluster  member  selection  in  the  wireless  multi-sensor  system. After  the  selection  of  cluster  head  and  member  nodes,  the  data fusion  technique  is  proposed  that  is  used  for  partitioning  and processing the data. The proposed scheme efficiently reduces the blind  broadcast  messages  but  also  decreases  the  signal  overhead due  to  cluster  formation.  Afterward,  the  routing  technique  is provided  based  one  the  layered  architecture.

Full Text PDF