S26117122

1,857 views
1,820 views

Published on

IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,857
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
1
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

S26117122

  1. 1. Venkata Santhosh Deepak, Mudiganti Vijaya Bhaskar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 www.ijera.com Vol. 2, Issue 6, November- December 2012, pp.117-122 Implementation Of Coquos On Unstructured P2P Network To Achieve Continuous Queries Venkata Santhosh Deepak*1 ,Mudiganti Vijaya Bhaskar*2 *1 M.Tech Student, Gokul Institute of Technology & Science, Bobilli , INDIA *2 Asst.Professor, CSE Dept, Gokul Institute of Technology & Science, Bobilli, INDIAAbstract Considering on-demand contentdiscovery mechanism, currently P2P contentdistribution systems are constructed. Byincluding two major capabilities like amechanism through which peers register longterm interest in network which notify new dataitems whenever they occur, and second capabilityis that advertising its content which enhances theutility of the systems. Presents proposals needintricate indexing and routing schemes to fulfilthe above two capabilities which not only makethem highly complex but also render the overlaynetwork less flexible toward transient peers. Forwhich we proposed a Scalable and effectivemiddleware which is called as CoQUOS tosupport unremitting queries in unstructured Fig.1: Decentralized P2P Architectureoverlay networks. We proposed two techniquesnamely, cluster-resilient random walk algorithm A second research area that is closelyand dynamic probability-based query related to work presented in this paper is that ofregistration scheme. First scheme used for pub-sub systems (event delivery systems. The earlypropagating the queries to various regions of the pub-sub systems adopted a centralized architecturenetwork and second scheme to ensure that the for subscription maintenance and matching andregistrations are well distributed in the overlay. notification of published data items, whereas theAnd we also proposed effective and efficient latter systems were either partially or completelyschemes for providing resilience to the churn of decentralized. More recently, researchers havethe P2P network and for ensuring a fair designed decentralized pub-sub systems on top ofdistribution of the notification load among the P2P networks several pub-sub systems have beenpeers. successfully implemented on structured P2P platforms. Unstructured P2P networks-based pub-Keywords- content delivery, continuous queries, sub systems are relatively less explored. The main difficulty in designing pub-sub systems onP2P networks, publish-subscribe systems, random unstructured P2P networks emerges from the factwalk. that the topologies of most unstructured P2P networks have no relationship to the data (or otherI. INTRODUCTION information) in individual peers. As a result, routing In the past few years, both structured and in these networks is, in general, independent ofunstructured P2P networks have experienced peers contents (a few unstructured P2P networkssignificant research searching through ad-hoc use content-sensitive heuristics for routing). Mostqueries has been the predominant information unstructured P2P networks-based pub-sub systemsdiscovery mechanism in P2P networks. Several try to overcome these difficulties by carefullyresearchers have studied efficient and scalable controlling the manner in which the topology of thealternatives to the flooding-based searching in overlay network evolves, and by adopting intricateunstructured P2P networks Random walk and its indexing strategies for routing subscriptions andvariants have been explored as alternatives to the notifications. Unfortunately, these mechanisms arebroadcast strategy. However, the ad-hoc query highly complex and they introduce significantparadigm is inadequate for advanced P2P content overlay management overheads thereby limiting thesharing applications. scalability of the corresponding systems. As mentioned in the introduction, the Sub- 2-Sub requires peers to be clustered on the basis of their 117 | P a g e
  2. 2. Venkata Santhosh Deepak, Mudiganti Vijaya Bhaskar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 www.ijera.com Vol. 2, Issue 6, November- December 2012, pp.117-122subscriptions. The publisher of a data item is research challenge is to develop a completelyrequired to reach the cluster that exactly corresponds distributed mechanism to register the continuousto the data item being published, and then start the queries at various regions of the P2P network so thatdissemination process. The RDF-based pub-sub the system has high notification effectiveness.scheme designed in the context of ELENA project Further, it is also necessary to develop low-costalso suffers from similar limitations. This system techniques for providing resilience to the churn ofalso requires the peers of the overlay to be the overlay network and for achieving fairmeticulously organized, in this case, to form an distribution of notification load among the peers inintricate bi-level Hyper cup architecture. Besides the network. Towards addressing these challenges,having limited scalability, the system is very this paper makes four novel contributions.vulnerable to super peer failures. The CoQUOS • First, we present a novel query propagationsystem differs fundamentally from P2Pbased pub- technique called Cluster Resilient Random Walksub schemes. All of the above-mentioned systems (CRW). This technique retains the overallare essentially pub-sub systems which are framework of the random walk paradigm. However,implemented on a P2P platform. In contrast, the at each step of propagation, CRW favors neighborsgoal of our work is to enhance the content discovery that are more likely to send messages deeper into thecapabilities of unstructured P2P content distribution network thereby enabling the continuous queries tonetworks. With the goal of achieving guaranteed reach different topological regions of the overlaynotification, most unstructured network. P2P-based pub-sub systems include • Second, a dynamic probability scheme is proposedcomplex overlay management strategies. On the for enabling the recipients of a continuous query toother hand, our CoQUOS system does not impose make independent decisions on whether to registerany restrictions on the topology of the overlay, the query. In this scheme, a query that has not beenemploys lightweight techniques, yet it provides high registered in the past several hops has a highersuccess rates. However, CoQUOS might not be chance of getting registered in its next hop, whichappropriate for applications where guaranteed ensures that registrations are well distributed alongnotification is necessary. Smart Seer is a structured the path of a query message.P2P overlay-based continuous query system fordocument repositories. Unlike Smart Seer, CoQUOSdoes not rely upon DHT for mapping continuousqueries to peer nodes hosting them or for routingnotifications. Instead, it incorporates lightweightmechanisms like CRW and dynamic probabilityscheme for registering continuous queries onvarious peer nodes of the overlay. Peer CQ is aP2Pbased continual query system for informationchange monitoring. It too uses the DHT to distributechange monitoring queries to the nodes of thenetwork. Researchers have studied the properties ofrandom walk as well as its utility for various P2Papplications such as computing aggregate queriesand uniform sampling of peers in unstructured P2Pnetworks. Several variants of random walk such aspopularity biased random walk have been proposed Fig.2: CoQUOS system overviewfor P2P searching. Show that the performance ofrandom walk degrades if the overlay networkexhibits considerable degree of node clustering. • Third, we discuss a passive replication-based scheme for preserving high notificationCRW overcomes this limitation by favoring out-of- effectiveness of the system even when thecluster peers at each hop of message propagation. underlying P2P network experiences significant churn.II. SYSTEM OVERVIEW • Finally, we propose a local load re-distribution The fundamental idea of the CoQUOS strategy to achieve fair distribution of notificationsystem is to register the continuous query on a set of loads among the participating peers.peers that are located in various topological regionsof the overlay network. These query replicas are a) Cluster-Resilient Random Walkused by the CoQUOS system to notify the Flooding-based broadcast is an option forrespective source peers of matching advertisements circulating continuous queries. However, this wouldissued by other peers. Considering the decentralized be analogous to a breadth first traversal of thenature of unstructured P2P networks, an important 118 | P a g e
  3. 3. Venkata Santhosh Deepak, Mudiganti Vijaya Bhaskar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 www.ijera.com Vol. 2, Issue 6, November- December 2012, pp.117-122network. As previous studies have reported, in this any of its beacon nodes, thus leading to low successscheme, messages remain in close vicinity of the rates.source node and do not go deep into the network. The objectives of our experimental study of theRandom walk is another message propagation CoQUOS system are fourfold:paradigm that has received considerable attention 1. Evaluating the effectiveness of the CRWfrom the P2P research community. In the context of technique in propagating continuous queries,P2P networks, random walk works as follows: 2. Studying the performance of the dynamicWhen a peer node Pi receives a message whose TTL probability approach for query registrations,has not expired, it selects one of its neighbors 3. Evaluating the churn resilience and the loadcompletely at random and forwards the message to balancing mechanisms, andthat peer. Since, at each step the message is 4. Evaluating the communication costs.forwarded to only one neighbor, the message loadimposed by random walk is very low. Random walk III. SYSTEM DESIGN &corresponds to a depth-first traversal of the network, IMPLEMENTATIONand a message propagated through random walks For implementation independenthas a higher probability of reaching remote regions specification and verification, and for subsequentof the network than its flooding-based counterpart. synthesis of specifications into efficientIn this paper, we use the terms random walk and implementations.pure random walk (PRW) interchangeably. The proposal is divided into below sub-tasks: i) adopt/further develop a model for formal, high- level system specification and verification. ii) Demonstrate the efficacy of the developed model by applying it to a suitable part of the consortium demonstrator, the network terminal for broadband access. iii) Develop a systematic method to refine the specification into synthesizable code and a prototype tool which supports the refinement process and links it to synthesis and compilation tools.Fig.3 : Illustration of cluster resilient random walkb) Dynamic-Probability-Based Query RegistrationThe CRW scheme provides a mechanism forpropagating a continuous query. But, how does anode receiving this message decide whether toregister the query? A straightforward solution wouldbe to register a query at every node it visits.However, this would result in large numbers ofunnecessary subscriptions, which affects theefficiency of the network. Alternatively, each peer Fig.4 : System Architecturereceiving a query message can decide register itwith a certain fixed probability, say Rp. We call this The Modelling represents a collection ofscheme the fixed probability-based query best engineering practices that have provenregistration scheme (FP scheme). Although this successful in the modelling of large and complexstrategy seems intuitive, it cannot guarantee high systems. The modelling is a very important part ofnotification success rates for every query. The developing objects oriented software and thereason is that for some continuous queries a long software development process. The modelling usesseries of peers in the path of the query message may mostly graphical notations to express the design ofall decide not to register the query, whereas another software projects. Using the modelling helpssequence of consecutive nodes may all decide to project teams communicate, explore potentialhost the query. The announcements originated near designs, and validate the architectural design of thethe dry patches of a query‟s this might fail to reach software. 119 | P a g e
  4. 4. Venkata Santhosh Deepak, Mudiganti Vijaya Bhaskar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 www.ijera.com Vol. 2, Issue 6, November- December 2012, pp.117-122 System Login Change Password Server Data Info Add Server ServerInfo Self Data 3 : Change Password() 1 : Server Data Info() PeerInfo Register PeerAdmin Data Upload Data User Info 2 : Self Data() Logout Peer Change Password 4 : Logout() Fig.7 : Interoperational Collavoration Diagram for Logout Peers IV. RESULTS Fig.5 : Interoperational Usecase diagram for adminstrator Peer Login Server Data Info Self Data Change Password Logout 1 : login() 2 : Server Data Info() 3 : Self Data() 4 : Change Password() 5 : Logout() Fig.8 : Viewing all Customer Details Fig.6 : Interoperational Sequence Diagram for Peers 120 | P a g e
  5. 5. Venkata Santhosh Deepak, Mudiganti Vijaya Bhaskar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 www.ijera.com Vol. 2, Issue 6, November- December 2012, pp.117-122Fig.9 : Information for the User about the data in theserver Fig.11 : Details of the User CONCLUSION Allowing peers of a formless P2P content sharing networks for registering longstanding queries and receiving notification when a new matching item appears can considerably increase effectiveness. We overcome the complexity of pub- sub system on unstructured overlays by our continues query paradigm study in this paper. Here we presented CoQUOS system which support continues queries in Unstructured P2P network by including different features such as lazy replication technique for countering network churn, cluster resilient random walk for query propagation and dynamic probability scheme for query registration. References [1] R. Baldoni, C. Marchetti, A. Virgillito, and R. Vitenberg. Contentbased Publish- Subscribe over Structured Overlay Networks. In Proceedings of ICDCS , 2005. [2] G. Banavar, T. Chandra, B. Mukherjee, J. Nagarajarao, R. E. Strom, and D. C. Sturman. An Ef_cient Multicast Protocol for Content- Based Publish-SubscribeFig.10 : Informatio of the Peers Systems. In Proceedings of ICDCS 1999, 1999. [3] S. Androutsellis-Theotokis and D. Spinellis. A Survey of Peerto- Peer Content Distribution Technologies. ACM Comput. Surv., 2004. 121 | P a g e
  6. 6. Venkata Santhosh Deepak, Mudiganti Vijaya Bhaskar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 www.ijera.com Vol. 2, Issue 6, November- December 2012, pp.117-122 [4] J. Chen, L. Ramaswamy, and A. Meka. Message Diffusion in Unstructured Overlay Networks. In Proceedings of NCA, 2007. [5] T. Bu and D. F. Towsley. On Distinguishing between Internet Power Law Topology Generators. In INFOCOM, 2002. [6] A. Carzaniga, D. S. Rosenblum, and A. L. Wolf. Design and evaluation of a wide-area event noti_cation service. ACM Transactions on Computer Systems, 19(3):332.383, 2001. [7] P. Reynolds and A. Vahdat, “Efficient Peer-to-Peer Keyword Searching,” Proc. ACM/IFIP/USENIX Int‟l Middleware Conf. ‟03, 2003. [8] B. Segall, D. Arnold, J. Boot, M. Henderson, and T. Phelps, “Content Based Routing with Elvin4,” Proc. Australian Unix System User Group Conf., 2000. [9] K. Sripanidkulchai, “The Popularity of Gnutella Queries and Its Implications on Scalability,” O‟Reilly‟s www.openp2p.com website. Feb. 2001. [10] I. Stoica, R.Morris, D.Karger, M.F. Kaashoek, and H. Balakrishnan, “Chord: A Scalable Peer-to-Peer Lookup Service for Internet Applications,” Proc. ACM SIGCOMM ‟01, Aug. 2001. [11] D.J. Watts and S.H. Strogatz, “Collective Dynamics of „Small- World‟ Networks,” Nature, vol. 393, no. 6684, pp. 440-442, 1998.Author List: Venkata Santhosh Deepak receivedB.Tech in Computer science from JNTUH, PursuingM.Tech in Computer Networks from Gokulinstitute of Engineering & Technology Affiliated toJNTUK.. His research area of interest is WebTechnology.Mudiganti Vijaya Bhaskar received B.Tech inComputer science from AU, M.Tech in ComputerScience from AU. Pursuing Ph.d in soft computingin AU. His research area of interest is SoftComputing. 122 | P a g e

×