a distributed control law for load balancing in content delivery networks

30
A Distributed Control Law for Load Balancing in Content Delivery Networks Presented by: C.Keerthi 11211F0012 Under the guidance of: Mr.D.B.Jagannadha Rao ( Associate professor)

Upload: sruthi-kamal

Post on 23-Jun-2015

537 views

Category:

Presentations & Public Speaking


0 download

TRANSCRIPT

Page 1: A Distributed Control Law for Load Balancing in Content Delivery Networks

A Distributed Control Law for Load Balancing in Content Delivery Networks

Presented by: C.Keerthi 11211F0012 MCA

Under the guidance of: Mr.D.B.Jagannadha Rao

( Associate professor)

Page 2: A Distributed Control Law for Load Balancing in Content Delivery Networks

AbstractIntroductionExisting & Proposed SystemsSoftware Requirement SpecificationModules & System ArchitectureDesign and System ModelsScreen ShotsConclusionFuture Enhancements

Page 3: A Distributed Control Law for Load Balancing in Content Delivery Networks

ABSTRACTIn this project, I may face the challenging issue of

defining and implementing an effective law for load balancing in Content Delivery Networks (CDNs).

I want to derive and prove a description about the network queues equilibrium.

This result is then leveraged in order to devise a novel distributed and time-continuous algorithm for load balancing, which is also reformulated in a time-discrete version

Page 4: A Distributed Control Law for Load Balancing in Content Delivery Networks

INTRODUCTION A Content Delivery Network (CDN) represents a

popular and useful solution to effectively support emerging Web applications by adopting a distributed overlay of servers.

By replicating content on several servers, a CDN is capable to partially solve congestion issues due to high client request rates , thus reducing latency while at the same time increasing content availability.

Page 5: A Distributed Control Law for Load Balancing in Content Delivery Networks

1. We present a novel load balancing algorithm to minimize both load imbalance factor and metric movement cost as much as possible.

2. The unique feature of our proposal is that each participating peer estimates and represents the “system state” as the probability distributions for the capacities of nodes and the loads of virtual servers.

3. The approximated probability distributions not only help estimate the expected load a peer should perceive but also provide hints for each peer in the system to schedule the transfers of virtual servers.

Page 6: A Distributed Control Law for Load Balancing in Content Delivery Networks

4 The participating peers in our proposal operate independently, and they need not rely on dedicated nodes to pair virtual servers and participating peers, eliminating the performance bottleneck and single point of failure.

5. Each peer in our proposal independently and solely manipulates partial information of the system and then reassigns its virtual servers to other peers based on the approximated system state.

Page 7: A Distributed Control Law for Load Balancing in Content Delivery Networks

System is not able to effectively face anomalous events like flash crowds.

Latency is increases drastically with load balancingSelection of two random servers is heavy load processA novel load balancing algorithm to minimize both load

imbalance factor and metric movement cost is very high.

Provide hints for each peer in the system to schedule the transfers of virtual servers.

Virtual servers establishment is mandatory, but not optional

Page 8: A Distributed Control Law for Load Balancing in Content Delivery Networks

1. I present a new mechanism for redirecting incoming client requests to the most appropriate server, thus balancing the overall system requests load. Another well-known static solution is the Round Robin algorithm (RR).

2. This algorithm selects a different server for each incoming request in a cyclic mode. Each server is loaded with the same number of requests without making any assumption on the state, neither of the network nor of the servers.

3. A modified version of such an algorithm is the Next-Neighbor Load Sharing. Instead of selecting two random servers, this algorithm just randomly selects one server and assigns the request to either that server or its neighbor based on their respective loads

Page 9: A Distributed Control Law for Load Balancing in Content Delivery Networks

Replicating content on several serversCDN is capable to solve congestion issues due to

high client request rates No need to establish virtual serverWithout disturbing any client redirector take care

about the input requestsLatency is decreases drastically with load balancing

Page 10: A Distributed Control Law for Load Balancing in Content Delivery Networks
Page 11: A Distributed Control Law for Load Balancing in Content Delivery Networks

Modules:

oBack end server

oSurrogate server

oRedirector

oClient

Page 12: A Distributed Control Law for Load Balancing in Content Delivery Networks

Back end server

I. An original server called back-end server containing new data to be diffused.

II. The basic network interface for all platforms at the syscalls read() and write(). Every modern OS provides its own syscall to help network servers transfer files as fast as possible.

Page 13: A Distributed Control Law for Load Balancing in Content Delivery Networks

Surrogate server

Content Delivery Network (CDN) replicates contents over several mirrored servers, named surrogate servers, which strategically place contents at various locations in order to deal with the flash crowds. CDN improves network performance by maximizing bandwidth usage, improving content accessibility and maintaining content updates through content replication thus offering fast and reliable applications and services by distributing contents to proxy servers often located closer to users.

Page 14: A Distributed Control Law for Load Balancing in Content Delivery Networks

3. Redirector The CDN architecture may consist of many surrogate

servers that could deliver copies of same content to one or more users. The CDN include the request routing infrastructure, the distribution infrastructure, and the accounting infrastructures. The request-routing infrastructure could consist of a mechanism to redirect content requests from a client to a suitable surrogate, or a group of surrogate, server(s).

Page 15: A Distributed Control Law for Load Balancing in Content Delivery Networks

4. Client when a surrogate is not able to fulfill the client’s request, the

request will be directed to the nearest surrogate or origin to fulfill that request. which dynamically redirects client requests based on selected policies. Response time experienced by clients after issuing a request. The decision process about these two aspects could be in contraposition. As an example, a “better response time” server is usually chosen based on geographical distance from the client, i.e., network proximity; on the other hand, the overall system throughput is typically optimized through load balancing across a set of servers.

Page 16: A Distributed Control Law for Load Balancing in Content Delivery Networks

HARDWARE REQUIREMENTS:

System    :   Any Processor above 500

MHz.

Hard Disk  :   40 GB.

Ram    :   512 MB.

Page 17: A Distributed Control Law for Load Balancing in Content Delivery Networks

SOFTWARE REQUIREMENTS

Operating System : Windows XP / 7 / 8

Technology : JSE

Software : JDK software

Version : JDK 1.5 /1.6 /1.7

Front end : Java Swings, AWT

Back end : No Database

IDE : Net beans 6.8 /7.1/7.3

Page 18: A Distributed Control Law for Load Balancing in Content Delivery Networks

DESIGN• The purpose of the design phase is to plan a solution

of the problem specified by the requirement document.

• In other words, starting with what is needed, design takes us toward how to satisfy the needs.

SYSTEM MODEL

• The unified Modeling Language (UML) is a standard language for writing software blueprints.

Page 19: A Distributed Control Law for Load Balancing in Content Delivery Networks

USECASE DIAGRAM

User Surrogate Server

Redirect Server

Client Connect to Server

Send Request

Check Load

Low Load

High Load

Send Response

Redirect to Other Server

Page 20: A Distributed Control Law for Load Balancing in Content Delivery Networks

CLASS DIAGRAM

BackEnd Server

+String srName+String cName+int cid+int sid

+void getSurrogateServer()+void getClinetInfo()+void getRequestInfo()+void getResponseInfo()

Surrogate Server

+String srName+int sid+int serviceRate+int arrivalRate

+void getClientRequest()+void checkLoad()+void responsetoClient()+void sendtoRedirectServer()

RedirectServer

+String srName+int sid+int cid

+void getClientRequest()+void redirectToOtherServer()

Client

+int cid+String cName

+void sendRequest()+void receiveResponse()

Page 21: A Distributed Control Law for Load Balancing in Content Delivery Networks

SCREEN SHOTS

Page 22: A Distributed Control Law for Load Balancing in Content Delivery Networks
Page 23: A Distributed Control Law for Load Balancing in Content Delivery Networks
Page 24: A Distributed Control Law for Load Balancing in Content Delivery Networks
Page 25: A Distributed Control Law for Load Balancing in Content Delivery Networks
Page 26: A Distributed Control Law for Load Balancing in Content Delivery Networks
Page 27: A Distributed Control Law for Load Balancing in Content Delivery Networks
Page 28: A Distributed Control Law for Load Balancing in Content Delivery Networks

CONCLUSION

Conclude by introduce a control theoretical analysis of the closed-loop congestion control problem in packet networks.

The control theoretical approach is used in a proportional rate controller, where packets are admitted into the network in accordance with network buffer occupancy.

Page 29: A Distributed Control Law for Load Balancing in Content Delivery Networks

FUTURE ENHANCEMENT. The second category (”the box is grey”) groups

approaches that use measurements to estimate available bandwidth, level of contention or even the temporary characteristics of congestion. Due to the possibility of wrong estimations and measurements, the network is considered a grey box.

The third category (”the box is green”) contains the bimodal congestion control, which calculates explicitly the fair-share, as well as the network-assisted control, where the network communicates its state to the transport layer; the box now is becoming green.

Page 30: A Distributed Control Law for Load Balancing in Content Delivery Networks