ITC571 – EMERGING

TECHNOLOGY AND
INNOVATIONS
CAPSTONE PROJECT SEMINAR

SUBMITTED BY: NILESH MISHRA

(11615981)
 TITLE: VIRTUAL TSA

AREA OF OPERATION :
CLOUD COMPUTING
TECHNOLOGY FOR
COMPANY: VMWARE
 INTRODUCTION
O The present report deals with improving the
experience of using cloud computing by
combining the time series analysis with
distributed resource scheduling (DRS) (a feature
of VMware cloud computing technology).
O Companies, with sufficient demand and
resources, are operating their own data centers
using VMware's vSphere to guarantee autonomy
and reliability which creates mapping problem
BACKGROUND OF PROJECT
O VMware uses a virtualization hierarchy shown to
the right.
O Tasks or applications run on virtual machines
(VMs), each of which run an OS abstracting the
ESX host's physical hardware.
O VMs can migrate between ESX Hosts, within
their cluster, to evenly distribute work.
O Applications use a variable amount of resources
(cpu, memory, network, disk), these resources
can often become over- or under-utilized which
means inefficient resource allocation
PROPOSED SOLUTION
O Having an efficient allocation of resources means lower
costs and energy usage.
O We provide a simple and intuitive interface to view stats
on the resource utilization of a datacenter, which allows
network administrators to make informed resource
scheduling decisions.
O We provide the tools necessary to compare various
statistics across multiple entities with the option of
correlation coefficients.
O We also provide functionality to do in-depth
investigations and forecasts of individual entities.
O Time-series analysis has not been applied to VMware's
software and our application offers an insightful
exploitation of the information available.
OVERVIEW AND SCOPE OF
PROJECT
O Virtual TSA can be divided into two categories the
first part is to perform the collection and visualization
of data that is collected from different ESXi hosts and
from vCenter Server (using SDK). Later this data is
stored in database which then will be used to extract
data.
O The second part is to help decision making of
administrators through a analyzer module which is
responsible for interaction with database. The
analyzer after receiving data performs different
analysis using techniques like moving average, FT
(Fourier Transform) etc..
FUNCTIONALITY OFFERED
BY VIRTUAL TSA
O The product is tasked to perform the
following functionality
O Use data analysis and represent it in visual
form
O Provide recommendations for vm migration
based on time series analysis
O Provide solutions to what if queries relating to
host management
 REQUIREMENTS FOR
VIRTUAL TSA
O The following are the requirements that are
necessary for implementing the Virtual Time
Series Analysis (Virtual TSA)
O HARDWARE: three ESXi servers, 1 NFS server, 1
vCenter server, 1 network switch (Gigabit)
O SOFTWARE:
O vSphere, vCenter, ESXi Hypervisor
O JRE (Java Run Time environment) for
programming
O Java, Python, C++
O VMware time series analysis libraries
 WORKING OF TSA
O The working of the current project "VIRTUAL
TSA" can be explained as user interacting
with the application and the application in
turn sends and receives data from MODEL,
then the model request data from database.
The data extracted will be put through Time
Series Analysis and the result which
depends upon the type of query the user
has asked will be displayed on the GUI.
FLOW OF INFORMATION IN
VIRTUAL TSA
The flow of information is shown as below
 DEVELOPED PHYSICAL
CLOUD INFRASTRUCTURE
O The cloud infrastructure used for testing is
 DESIRED OUTPUT
O The output of this project will be producing a software
interface that will provide the functional
requirements of
O A browser based Virtual time series application
which will display VM statistics at cluster level
O Produce a splash screen with option to select cluster
to analyze and investigate
O Analyze any previous dataset
O Option to produce graphs depending upon a
selected question from the list of what-if queries
O Live updates in the form of graphs
O Comparison of statistics based on active data set
using the comparison tab
ORIGINAL OUTPUT
THANK YOU