Robust Spanning Tree Topology ( Data Collection And Data Dissemination For Distributed Environments)

Done By P. Adi Lakshmi (M.Tech(sss))

(07024D0512)JNT University,Kakinada.

Under the Guidance of L.Sumalatha,Associate Professor,JNT University,Kakinada.

Abstract Large-scale distributed applications are subject to frequent disruptions

due to resource contention and failure. Such disruptions are inherently unpredictable and, therefore, robustness is a desirable property for the distributed operating environment.

In this work, we describe and evaluate a robust topology for applications that operate on a spanning tree overlay network. The topology itself is able to simultaneously withstand disturbances and exhibit good performance.

The results show that our robust spanning trees achieve a desirable trade-off for two opposing metrics where traditional forms of spanning trees do not.

Existing System

Shortest paths: The distance in edge weights of the path from a node to the

other node is minimum. Such a tree is efficiently constructed by Dijkstra’s algorithm.

Fewest hops: The distance in number of hops along the path from each

node to the root node is minimum. This method is equivalent to SP when all edge weights are equal and therefore Dijkstra’s algorithm may be employed.

Disadvantages of Existing System

Fewest Hop: Fat and Shallow High Cost Power Consumption is high

Shortest Path: Deep and Skinny Data Loss is high

Proposed System

In our approach we first reduce the density of the topology

using the centralized algorithm. Then we deplete the data loss and

power consumption. This topology can be used both in centralized

system and distributed system. Robustness cannot be achieved in the

other existing system whereas it is one of the desirable property

which is achieved in our topology.

Software Requirements

Operating System : Windows 2000 or Later Version

Technologies : JDK 1.5

Data Bases : MS-Access

Front End : Java Swing

Hardware Requirements--Processor : Any Processor above 500 Mhz.

Ram : 128 MB.

Hard Disk : 10 GB.

Input device : Standard Keyboard and Mouse.

Output device : VGA and High Resolution Monitor.

Modules

Creating a normal topology structure.

Creation of Shortest Path Topology.

Creation of Fewest Hop Topology.

Creation of Robust Spanning Tree Topology.

Comparison of all the three topologies.

Creating a normal topology structure

In this module, we are constructing a normal topology structure. This

shows how the systems are connected between them.

The topology structure is constructed by identifying the possible

paths between the nodes that are connected.

The major problem in this topology structure is that it has redundancy

problems, which leads us to the improved topology structure to be

constructed.

Normally Connected Systems

Creation of Shortest Path Topology

In this module, we are converting the normal topology structure into

the Shortest Path Topology Structure.

This can be constructed by considering only the edge weights between

the nodes that are connected.

Creation of Fewest Hop Topology

In this module, we are constructing a topology based on Fewest

Hops Method.

This topology is constructed by considering the number of hops

between the source and the destination.

Creation of Robust Spanning Tree Topology

This module is our proposed system where we construct the topology

based on Robust Spanning Method.

This can be constructed by considering the advantages of both the

existing systems, Shortest Path and Fewest Hops.

a) Shortest Path

b) Fewest Hops

c) Robust Spanning Tree

Spanning trees for different topologies

The example spanning trees for existing topologies Shortest Path, Fewest Hops and proposed topology Robust Spanning tree.

Comparison of all the three topologies

In this module, we compare the

performance of all the three

constructed topologies.

Comparison is done based on the

latency time associated with each

topologies.

Here we show that the Robust

Spanning Tree Topology is

significantly better than the other

existing topologies.

Architecture

A

Comparing the three topologies

User

Architecture(Contd…)

UML Diagrams

Class Diagram:

Sequence Diagram:

Use case Diagram:

Screen Shots

Main / Startup Frame:

The above screen is the startup frame that allowing us to either add a new node to our topology or to move to Distributed environment for transmitting data from root node to destination node using different paths.

Frame to Add New Node:

This screen allows us to add a new node our topology. We need to provide Node Name, Port number and edge weight to connect the selected node in the list.

Distributed Environment:

This screen allows us to find all the possible paths and to send a message (either we can type or we may browse from existing text file) to the selected destination node. The screen also allows us to redirect Spanning page or Startup screen.

Distributed Environment:

We can observe here the list of possible paths for selected destination node ‘D’ and message in the provided box to send.

Spanning Page SP Topology Selected:

This screen allows us to transmit message to the selected destination node using three different topologies namely Shortest Path, Fewest hops and Robust topology.

Spanning Page Robust Topology Selected:

We can observe here the list of possible paths to send message to the selected node ‘D’ using Robust Topology.

Spanning Page Acknowledgement:

We can observe here when the selected node ‘D’ receive the message then it sends acknowledgement .

Comparison Chart for all three Topologies:

The above graph shows us the clear difference between our proposed robust topology and existing topologies in terms of latency time.

Conclusion:

Robustness is an important property for distributed computing

systems. These systems are subject to resource contention and, hence,

node failures and transmission delays are common enough to warrant

their consideration in system design. This is especially true when the

application designer has some control over the manner in which data is

routed and computations are performed, such as the choice of topology

for an overlay network. In this work, we presented a methodology for

constructing a spanning tree overlay network that exhibits robustness to

network disturbances. The construction technique employs a weighted

formula for hop count and path weight that changes the relative

importance as the distance from the root node changes. This results in

trees that perform well for a wide variety of metrics.

Future Enhancement:Our approach toward robustness is proactive rather than

reactive. It is natural to ask when a node realizes that its parent has failed,

why not simply choose another parent (assuming the node has multiple

neighbors)? This may or may not be desirable. If there are many nodes

that choose a new parent, then the properties of the tree will be unknown.

If the goal is to collect a reasonable amount of data over a long

period of time, then it would be better to use a topology about which we

have some statistics. It seems that the pertinent question is: At what point

is it worth rerunning the spanning tree construction algorithm to construct

a new tree? This is one subject of our future work.

Thank You!