- Open Access
- Total Downloads : 17
- Authors : K V S K Prakash, G Lakshmikanth
- Paper ID : IJERTCONV3IS18032
- Volume & Issue : NCACI – 2015 (Volume 3 – Issue 18)
- Published (First Online): 24-04-2018
- ISSN (Online) : 2278-0181
- Publisher Name : IJERT
- License: This work is licensed under a Creative Commons Attribution 4.0 International License
A Lightweight Solution for Eliminating Redundant Resources on Cloud
Cloud Resources Optimization
K V S K Prakash (M.Tech)
Dept. of CSE
Sree Rama Engineering College Ramireddy Nagar, Karakambadi Road, Tirupati
G Lakshmikanth (M.Tech)
HOD, Dept. of CSE
Sree Rama Engineering College Ramireddy Nagar, Karakambadi Road, Tirupati
Abstract In this paper, we suggest an idea on eliminating the duplicate resources on cloud by checking small portion of the resources by comparing with the already downloaded stuff, such that a lot of burden on the Client would get relieved like Price, Time, and Effort etc The main idea is to offload the server in checking for duplicates such that the client itself can check for duplicate resources using predictive acknowledgement. This paper also enhances the idea by providing cloud Client and Server Migration, Redundant Resource Elimination. Its a absolutely new Idea that earlier would rely on the solution like Name matching, comparing old packets with the new packages arrived just. The previous solution is based on TCP and extension has been made to work on all Platforms and Protocols.
KeywordsCloud Elasticity, Redundant Resources, Cloud Resources Optimization, Bandwidth Optimization
As Market grows we all knew that Cloud Computing is a very vast and emerging means of providing services to the client in its various forms such as platform, infrastructure, and software. In addition to this now they added some extensions like Security as a service, Database as a Service, etc As each and every firm or organization offers its services through cloud in the present and very extensively in the nearby future all the Companys Services were available as a part of Cloud. So we got Cloud services as PAY PER USE service. Clients or customer pay only to their usage regarding the Bandwidth, CPU power, Servers using , Processing Power, VMs Consumption, Memory Consumption etc Data transfer costs also include heavy impact on the Cloud Services if the users download the resources which were already downloaded that might be with a different name or same name. If the redundant services are not removed from the Server very clearly they would increase the network traffic which may cause severe issues. Earlier the complete load is on the server which has to check whether any duplicate resource is existing or not. Actually this is a new idea when the server is not in use but the Server cant be expected to be free all the time. In fact Server is usually busy in serving the clients so checking or redundant resources on the server side would not worked for a longer time. Next Idea is to provide these checking both at server and client side. This also became vague over a period. Finally the server is off loaded from this checking where that job is taken care by
the Client itself such that whenever it found such redundant packets or data it will send an acknowledgement to the server (predictive acknowledgement).
The above Architecture explains about the Server and Client management of the Resources on the Cloud
The Server registers itself with CSP for uploading, downloading and deleting Resources on the cloud. The Client after Registration can only download the Resources from the Cloud after proper validation. If the Resource is already downloaded immediately it will send an PRED to the Server and the Server replies with an PRED ACK
Abbreviations and Acronyms
PRED-ACK Predictive Acknowledgement RRE Redundant Resource Elimination Protocol TCP Transmission Control Protocol
Many Techniques have been developed over the years for redundant packet elimination but initially that is done on purely server side only. Later it was extended to Server and Client, finally that was implemented on client side. This paper avoids all the third party middle-box techniques for finding the duplicate packets. Hash code will be generated for each and every packet, stored on the table and for each new packet this hash code will be evaluated and compared with the hash codes which are already stored. If the code is found it indicates the packet has been duplicated and a prediction is sent to the server for this duplicate packet. At last the Server will scrutinize the packet and sends an acknowledgement to the Client.
Generally in cloud based services checking for the redundant copies of resource is done on the server side which is a expensive task in terms of cost and effort. This can be shared both on server and client side, this is a little bit simplified version when compared to checking the whole thing on the server side. This complete mechanism should be shifted to on the client side. The previous solution doesnt address the problems like Cloud Elasticity, Resource Optimization. The Problem is the checking is done based on the name but on the content that whether it has been duplicated.
A Heavyweight model for eliminating the duplicates on the cloud using Server Side checking, Server and client side checking, Rabin Fingerprinting methods seems not applicable to a greater extent to the real time problems. Traffic Redundancy will be increased a lot and cant be controlled with these ideas. These will increase the Cost, Traffic , Bandwidth Consumptions, end users effort and Time. The identified problems are very fatal which needs immediate concern for handling and setting up right.
A Lightweight solution for all the above problems can be suggested by shifting the complete idea towards the client side with comparing the already received packets signature with the newly arrived packets. This idea finds duplicate resource before their download which will resolve all the above problems like Traffic Optimization, End users Effort and Time saving , cost will be reduced , Cloud Resources optimization can be happened, Bandwidth Consumption is reduced .
The basic idea is to download only a portion of the resource which has to be downloaded , compute its signature and find the hash code and compare it will the hash codes already stored on the store. If there is a match Immediately respond the CSP with the message that a resource has been replicated. Then server can acknowledge the client about the message
and can delete the duplicated file from the store such that all the resources on the cloud could be genuine and reliable
Server Side Algorithm:
Server Registers with the Cloud Services
Server places all its resources on the Cloud
Server has the permission of inserting , deleting and downloading any resource
Client Side Algorithm:
Client Registers with the Cloud Services
After Proper validation the client to allowed to see and download all the resources.
If a resource is downloaded, Hash code is generated using SHA Algorithm and stored on the Store.
If a new resource is downloaded, as usually HASH CODE will be generated and compared with the already stored codes, if there is a match send a prediction to the server.
After receiving the prediction from the server an acknowledgement will be sent to the Client.
The above flowchart depicts a complete flow of control between the server and the client for navigating through all the transactions that are happened on the cloud here our discussion will come under Resoures as a service with optimization of resources, cloud mobility that includes both server and client side migration, discusses how the bandwidth can be optimized, how redundant resources on the cloud can be eliminated through our PRED and PRED-ACK concept. Further our discussions can be applied even to UDP and other protocols and domains.
Cloud Computing offers better services by extending the idea to all the protocols such and TCP , UDP etc. and to all the domains such as WEB APPS, MOBILE APPS, Enterprise Apps.
A Novel optimistic idea for eliminating replicated resources over the Cloud. Sometimes same videos have the same content when they have different qualities of VIDEO and AUDIO especially with videos , or even with documents . Sometimes the starting portions of several videos might be the same which has unique content somewhere after playing some time. In such cases rather than reading first part of chunk of the data its better to read in the middle of the Document or video for calculating the HASH CODE.
E. Zohar, I. Cidon, and O. Mokryn, The power of prediction: Cloud bandwidth and cost reduction, in Proc. SIGCOMM, 2011, pp. 8697.
Prakash E J, A Nageshwara Rao, Periodically predicting clients bandwidth and cost acknowledgements sends to server to optimize resource usage, IJIRCCE, Vol.2, special issue 4,September 2014.
N. T. Spring and D. Wetherall, A protocol-independent technique for eliminating redundant network traffic, in Proc. SIGCOMM, 2000, vol.30, pp. 8795.
A.Muthitacharoen, B. Chen, and D. MaziÃ¨res, A low- bandwidth network file system, inProc. SOSP, 2001, pp. 174 187.
Katrina LaCurts, Jeffrey C. Mogul, Hari Balakrishnan, and Yoshio Turner, Cicada: Predictive Gaurantees for cloud network bandwidth, MIT-CSAIL-TR-2014-004,March 24,2014
S.Mccanne and M.Demmer, Content-based segmentation scheme for data compression in storage and transmission including hierarchical segment representation, US Patent 6828925, Dec. 2004.
A.Gupta, A. Akella, S. Seshan, S. Shenkar, and J. Wang,
Undestanding and exploiting network traffic redundancy, UW- Madison,USA,Tech.Rep, 1592,Apr, 2007.
E.Lev-Ran, I.Cidon andI.Z.Ben-Shaul, Methodandapparatus for reducing network traffic over low bandwidth links, US Patent 7636767, Nov. 2009.
M. Armbrust, A. Fox, R. Griffith, A.D.Joseph, R.Katz, A.Konwinski, G. Lee, D. Patterson, A. Rabkin, I. Stoica, and M. Zaharia, A view of cloud computing, Commun. ACM, vol. 53, no. 4, pp. 5058, 2010.
Suresh Chougala and Sharavana K, Design of traffic redundancy and elimination approach for reducing cloud bandwidth and cost, IJIRCCE, Vol. 2,issue 2, February 2014.
Xaunran Zong, Predicting application performance in the cloud, Masters thesis , Duke university,2011.
Rushil Dave, Mobile virtual network operator systems on cloud: An architectural and cost-benefit study, Masters thesis, august 8,2011.