Network Bottleneck Avoidance

using Edge Routers

Presented By:Ankur Singhal

Mayank Manchanda

OUTLINE

Introduction

Current Scenario

Problem Statement

Solution Approach

Feasibility Study

Scope

Existing System

Proposed System

Findings and Conclusion

Future Work

INTRODUCTION

The Internet’s excellent scalability and robustness result in part from the end-to-end nature of Internet congestion control. End-to-end congestion control algorithms alone, however, are unable to prevent the congestion collapse and unfairness created by applications that are unresponsive to network congestion.

The fundamental philosophy behind the Internet is expressed by the scalability argument- no protocol, mechanism, or service should be introduced into the Internet if it does not scale well. A key corollary to the scalability argument is the end-to-end argument: to maintain scalability, algorithmic complexity should be pushed to the edges of the network whenever possible.

CURRENT SCENARIO

There is a need for faster data transfer without the loss of packet while transmission either through congestion, communication channel failure or node failure.

Users don’t want an overhead of retransmission of lost or dropped packets.

Utilization of maximum bandwidth available to the systems.

PROBLEM STATEMENT

Congestion collapse from undelivered packets

Unfair bandwidth allocation to competing network flows

SOLUTION APPROACH

Compare, at the borders of a network, the rates at which packets from each application flow are entering and leaving the network.

If a flow’s packets are entering the network faster than they are leaving it, then the network is likely buffering or, worse yet, discarding the flow’s packets.

Ensuring that each flow’s packets do not enter the network at a rate greater than they are able to leave the network.

Using NBA with ECSFQ for allocation of fair bandwidth.

FEASIBILITY STUDY TECHNICAL REQUIREMENTS Advanced Java Socket Programming File Concepts Threads.

ECNOMICAL REQUIREMENTS Proposed system is cheaper Easily adaptable for both user and developer.

OPERATIONAL REQUIREMENTS Message sending Routing the packet’s Controlling the packet flow Avoids packet loss.

HARDWARE REQUIREMENTS

The minimum configuration required to run this project are:Main processor : Pentium III (or) IV RAM : 128MBHard Disk : 10MBClock Speed : 550 MHZSystem Bus Speed : 400 MHzCache RAM : 256 KB

SOFTWARE REQUIREMENTS

Language : JDK1.7 & Above. Front End Design : Swing CONCEPTS Operating System : Windows XP and higher

version.

SCOPE OF THE PROJECT

Network Bottleneck Avoider entails the exchange of feedback between routers at the borders of a network

Restrict unresponsive traffic flows Prevents congestion within the network.

EXISTING SYSTEM

Congestion collapse. Retransmission. Unfair bandwidth allocation core-stateless fair queuing WFQ (Waited Fair Queuing) not

sufficient for avoiding congestion

PROPOSED SYSTEM

Buffering of packets is carried out in the edge routers

No possibility of any undelivered packets present in the network

Fair allocation of bandwidth is ensured

MODULE ALLOCATION

Module 1 SOURCE MODULE.

Module 2INGRESS ROUTER MODULE.

Module 3ROUTER MODULE.

Module 4EGRESS ROUTER MODULE.

Module 5DESTINATION MODULE

DATA FLOW DIAGRAM

Forward Forward Feedback Feedback

Source

Source

Source

Destination

InRouterRouter

Router OutRouterRouter

Destination

Destination

BackwardFeedback

BackwardFeedback

SOURCE MODULE

Input Parameters:

Source Machine Name is retrieved from the OS. Destination Machine Name is typed by User. Message is typed by User.

Output Parameters: Data Packets.

SOURCE DATA FLOW DIAGRAM

USER INPUT

DESTINATION AND MESSAGE

PACKET SPLIT INTO 48 BYTE

SENDING DATA TO ROUTER

RECEIVING ACK FROM ROUTER

PACKET RECEIVED IN ROUER

INGRESS MODULE

Input Parameters: Data Packets from Source Machine. Backward feedback from the Router.

Output Parameters: Data Packets. Forward feedback.

ALGORITHM IMPLEMENTED LEAKY BUCKET ALGORITHM

TIME SLIDING ALGORITHM

Arrival of the Forward Feedback at the OutRouter Router

Start the timer

If Packets are arrived

Wait until the packet is forward

Current Packet is send

Wait until the Packet is arrived

Yes

No

Acknowledgement is backward to InRouter

If Packets are forwarded

Yes

No

If no Packet to ForwardedStop the timerYes

Forward the next packet

No

ROUTER MODULE

Input Parameters:

Data Packets from Ingress Machine. Forward feedback from the Router or Ingress

Router. Backward feedback from the Router or Egress

Router. Hop count.Output Parameters:

Data Packets. Forward feedback. Incremented Hop count. Backward feedback.

DATA FLOW DIAGRAM ROUTER

USER INPUT

DESTINATION AND MESSAGE

PACKET SPLIT INTO 48 BYTE

SENDING DATA TO ROUTER

RECEIVING ACK FROM ROUTER

PACKET RECEIVED IN ROUTER

EGRESS ROUTER MODULE Input Parameters: Data Packets from Router. Forward feedback from the Router.

Output Parameters: Data Packets. Backward feedback

RATE CONTROL ALGORITHM

IfCurrentRTT<e.base

RTT

On arrival of backward feedback packet p from OutRouter router e

Current RTT =Current Time -p.times tamp

Delta RTT=-Current RTT-e.base RTT

e.base RTT=Current RTT

RTTs Elapsed= (Current Time-e.last FeedbackTime)/CurrentRTT

e.last FeedbackTime=Current Time

For each flow f listed in p

Rate Quantum=min (MSS/currentRTT, f.egreesRate/QF)

AB

True

False

f.phase=CONGESTION_AVOIDANCE

If (deltaRTT*f.InRouterRate<MSS*e.hopcount)

f.InRouter Rate=f.InRouterRate*2^RTTsElapsed

If f.phase = = CONGESTION_ AVOIDANCE

If (deltaRTT*f.InRouterRate<MSS*e.hopcount)

f.InRouterRate=f.InRouter Rate + rate Quantum*RTTsElapsed

f. InRouter Rate=f.OutRouterRate-rateQuantum

NEXT

True

True

False

True

False

If f.phase = = SLOW_START

BA

DESTINATION MODULE

Message received from the egress router will be stored in the corresponding folder as a text file depends upon the Source Machine Name.

DESTINATION DATA FLOW DIAGRAM

DESTINATION

RETRIVE THE ALLOCATED BYTE

COMBINE THE PACKET IN(48 BYTES)

FINALLY SAVE THEM IN TEXT FILE

FINDINGS One flow is a TCP flow generated by an

application that always has data to send

The other flow is a constant bit rate UDP flow generated by an application that is unresponsive to congestion

FINDINGS Severe congestion collapse using FIFO only

FINDINGS Moderate congestion collapse using ECSFQ only

FINDINGS No congestion collapse using NBA with FIFO.

CONCLUSION NBA is able to prevent congestion collapse from

undelivered packets. NBA ensures at the border of the network that each flow’s packets do not enter the network faster than they are able to leave it.

Simulation results show that NBA successfully prevents congestion collapse from undelivered packets. They also show that, while NBA is unable to eliminate unfairness on its own, but it can be seen that it will be able to achieve approximate global max-min fairness for competing network flows when combined with ECSFQ, they approximate global max-min fairness in a completely core-stateless fashion.

FUTURE WORK

Combining Enhanced Core Stateless Fair Queuing mechanism With the Network Bottleneck Avoider to achieve fair allocation of bandwidth within the link.

Simulate and compare the performance of NBA system using multiple routers.

Examine backward feedback packets from more than one egress router.

Determining the most congested ingress to egress path (i.e., the one with the lowest flow egress rate)

THANK YOU