Equivalent Servant Locator in Peer-To - Peer Networks

International Journal of Computer Trends and Technology- volume3Issue1- 2012
Equivalent Servant Locator in Peer- to- peer Networks

Prasanti Boyapati1, Dr. D. Srinivas Kumar2
1
Pursuing M.Tech(CSE), Nalanda Institute of Engineering & Technology, Siddharth Nagar, Sattenapalli, Guntur., Affiliated to JNTUK, Kakinada, A.P., India.
Principal & Professor, Department of Computer Science Engineering, Nalanda Institute of Engineering & Technology, Siddharth Nagar, Sattenapalli, Guntur., Affiliated to JNTUK, Kakinada, A.P., India.
Abstract - Peer-to-Peer systems offer an alternative to traditional client-server systems for some application domains. A P2P network is a distributed network composed of a large number of distributed, heterogeneous, autonomous, and highly dynamic peers in which participants share a part of their own resources such as processing power, storage capacity, softwares, and files contents. The participants in the P2P network can act as a server and a client at the same time. Such participants are called servants. This paper presents an equivalent servant locator for locating equivalent servants in peer- to- peer networks. Keywords- P2P network, Servant, client, server.
I. INTRODUCTION Computation in networks of processing nodes, each holding a part of the inputs and/or resources initially, can be classified into centralized or distributed computations( see figure 1). A centralized solution relies on one node being designated as the computer node that processes the entire application locally. In distributed computation, the processing steps of the application are divided among the participating nodes. The goal in such systems is to minimize communication and computation cost. Distributed systems can be further classified into a client-server model and a P2P model. In the client-server model, the server is the central registering unit, as well as the only provider of content and services. A client only requests content or the execution of services, without sharing
any of its own services. The client-server model can be flat where all clients only communicate with a single server or it can be hierarchical for improved scalability. Peer-to-Peer systems offer an alternative to traditional client-server systems for some application domains. A P2P network is a distributed network composed of a large number of distributed, heterogeneous, autonomous, and highly dynamic peers in which participants share a part of their own resources such as processing power, storage capacity, softwares, and files contents. The participants in the P2P network can act as a server and a client at the same time. They are accessible by other nodes directly, without passing intermediary entities. The P2P models can be pure or hybrid. In pure P2P any single, arbitrary chosen terminal entity can be removed from the network without having the network suffering any loss of network service. Hybrid P2P allows the existence of central entities in its network to provide parts of the offered network services.
ISSN: 2231-2803 http://www.internationaljournalssrg.org
Page 130
Fig. 1. computer systems There are several concepts underlying p2p systems: sharing resources, decentralization and self organization. Resource sharing implies that applications can not be set up by a single node. Shared resources can be physical resources such as disk space, CPU or network bandwidth, as well as, logical resources such as services or different forms of knowledge. Decentralization is an immediate consequence of sharing of resources. Decentralization is in particular interesting in order to avoid single point of failures and bottlenecks. When a p2p system becomes fully decentralized then there exists no longer a node that can centrally coordinate its activities or a database to store global information about the system centrally. Therefore nodes have to self-organize themselves,based on whatever local information is available and interacting with locally reachable nodes (neighbors). The global behavior then emerges as the result of all the local behaviors that occur. This paper focuses on services provided by equivalent servants and models and analyzes the performance of structured and unstructured overlays when used to provide such services. We present EQUATOR (EQUivalent servAnt locaTOR), a P2P-based architecture deployable in real networks for the provision of services based on equivalent servants. EQUATOR aims at guaranteeing high lookup performance, as well as high robustness to failures and churn phases, when a significant number of peers joins/leaves the network.
instances of such systems. Pure p2p systems are inherently scalable. Scalability in the system is usually restricted by the amount of centralized operation necessary and such system largely avoid central instances or servers. This kind of systems are inherently fault-tolerant, since there is no central point of failure and the loss of a peer or even a number of peers can easily be compensated. They also have a greater degree of autonomous control over their data and resources. On the other hand such systems present slow information discovery and there is no guarantee about quality of services. Also because of the lack of a global view at the system level, it is difficult to predict the system behavior.
2.2 Hybrid Architecture In hybrid P2P systems [4], there is a central server that maintains directories of information about registered users to the network, in the form of metadata. The end-to-end interaction (data exchange) is between two peer clients. There are two kinds of hybrid systems: centralized indexing and decentralized indexing. In centralized indexing [figure 2] a central server maintains an index of the data or files that are currently being shared by active peers, Each peer maintains a connection to the central server, through which the queries are sent. This architecture is used by Jabber[5]. Such systems with the central server are simple and they operate quickly and efficiently for discovery information. Searches are comprehensive and they can provide guarantee in searches. On the other hand they are vulnerable to censorship and malicious attack. Because of central servers they have a single point of failure. They are not inherently scalable, because of limitations on the size of the database and its capacity to respond to queries. As central directories are not always updated, they have to be refreshed periodically.
II. PEER- TO- PEER ARCHITECTURES Decentralization is one of the major concept of p2p systems. This includes distributed storage, processing, information sharing and also control information. Based on the degree of decentralization in a p2p system, we can classify them into two categories: 2.1 Purely decentralized A pure p2p system is a distributed system without any centralized control. In such systems all nodes are equivalent in functionality. In such networks the nodes are named as servant (SERver+cliENT), the term servent represents the capability of the nodes of a peerto-peer network of acting at the same time as server as well as a client. Gnutella[1], Freenet[2], Chord[3] are
Page 131
being part of the scale-free topology and some normal users accessing the offered service.
Fig. 2. centralized indexing In decentralized indexing [figure 3], a central server registers the users to the system and facilitates the peer discovery process. In these systems some of the nodes assume a more important role than the rest of nodes. They are called supernodes[4]. These nodes maintain the central indexes for the information shared by local peers connected to them and proxy search requests on behalf of these peers. Queries are therefore sent to SuperNodes, not to other peers. Fig 4. Architecture of EQUATOR 3.1. EQUATOR bootstrap service In real P2P networks, entering nodes cannot have any knowledge about the existing overlay and therefore a Bootstrap Service is required in order to give such nodes the opportunity to join the network. In particular, the Barabsi-Albert model requires a set of m0 peers to be available at the initial step of the overlay setup. A simple solution (adopted in many existing overlays such as KaZaA [6]) consists in setting up some static peers and pre-configuring their addresses on each client. In EQUATOR, we prefer a more flexible approach that relies on multiple bootstrap servers reachable through appropriate DNS records (e.g., SRV entries), thus guaranteeing redundancy and load balancing. Bootstrap servers globally store information about m0 participating peers; when a peer joins the overlay, it adds itself to the list in case the number of entries in the bootstrap servers is n<m0. Fig. 3. distributed indexing III. EQUIVALENT SERVANT LOCATOR This section presents EQUATOR, an unstructured overlay based on the Barabsi-Albert model (and hence on a scale-free topology), which adopts a set of construction and operating rules that are suitable for a real network. Furthermore, an epidemic dissemination algorithm is used to spread the network knowledge among the participating peers. A portion of a possible EQUATOR overlay is shown in Fig. 4, with some peers 3.2 Node popularity In a network based on epidemic dissemination , nodes send advertisement messages to other nodes in order to maintain the overlay. In a scale free topology the popularity is equivalent to the in- degree of the node. Since it is unfeasible for an equator node to be aware of its in- degree, EQUATOR adopts as popularity metric the number of advertisement messages a node receives, which is proportional to its in- degree. In particular a node can estimate its popularity by maintaining statistics about the average
Page 132
number of received messages per minute.The popularity of a node is used both in overlay construction and maintenance. 3.3 Cache refresh In EQUATOR, the knowledge of the network at any time t is limited to a few nodes , i.e., the ones that are in the two caches(servant cache and overlay cache). Refresh is the key technique that allows the deployment of small caches, which limits overhead due to both cache management and advertisement and lookup traffic. Furthermore, it reduces the possibility to have an old servant, which may be dead or currently unavailable (actually servicing a request) in the servant cache. In fact, a frequent cache refresh ensures the set of indexed servants changes frequently, resulting in a sort of round robin among them. Since the cache refresh rate at a node is proportional to the number of advertisement messages received and, consequently, to its popularity, this effect is maximized at the hubs, which have the opportunity to virtually offer a large number of servants, notwithstanding the limited size of the servant cache. Frequent entry refresh is also important for the overlay cache to allow the overlay to be dynamic and hence more robust. When a new peer joins, its overlay cache only contains the bootstrap nodes retrieved from the EQUATOR Bootstrap Service. Thanks to the refresh, nodes can insert new peers in their overlay cache and update the popularity information of the peers they already knows. This increasing knowledge of the network allows nodes to incrementally contribute to the construction of the scale-free topology as they can apply a more and more accurate preferential attachment. Hence, the overlay results in a scale-free topology, although variable over time. Furthermore, a frequent refresh ensures nodes are aware of live peers and hence well connected to the rest of the overlay. 3.4 Service lookup procedures from normal users While the overlay contains all the peers that are available to offer some of their resources (i.e., are potential servants), many hosts may join the system as normal users in order to simply exploit the overlay services and without taking an active part in the overlay. Users are most interested in service lookup functionalities and therefore have an advantage at connecting to peers that know many servants. In fact, in
our model service requests are distributed among the participating peers proportionally to their popularity, i.e., requests are preferably directed to hubs. Consequently, preferential attachment is beneficial also for users and therefore we need to implement an approximation of this algorithm also with respect to these nodes. The service lookup procedure we defined for normal users works as follows. Each user maintains a node cache, referred to as lookup cache. Whenever a user logins in EQUATOR, her EQUATOR instance connects to the Bootstrap Service and retrieves the initial m0 nodes. The user node selects one of them randomly and downloads its overlay cache. This procedure is repeated periodically in order to guarantee both the user node to have up-to-date knowledge of existing peers and service lookups to be well distributed among the peers. In fact, simply populating the lookup cache with nodes retrieved from the Bootstrap Service would possibly result in concentrating the lookup traffic among a few peers, with possible congestions.
4. CONCLUSION A P2P network is a distributed network composed of a large number of distributed, heterogeneous, autonomous, and highly dynamic peers in which participants share a part of their own resources such as processing power, storage capacity, softwares, and files contents. The participants in the P2P network can act as a server and a client at the same time. Such participants are called servants. This paper presents an equivalent servant locator for locating equivalent servants in peerto- peer networks.
REFERENCES
[1] M. Ripeanu, I. Foster, and A. Iamnitchi. Mapping the gnutella network: Properties of large-scale peer-to-peer systems and implications for system design. IEEE Internet Computing Journal, 6(1), 2002. [2] Ian Clarke, Oskar Sandberg, Brandon Wiley, and Teodore W.Hong. Freenet: A distributed anonymous information torage and retrieval system. Lecture Notes in Computer Science, 2009:46+, 2001. [3] Ion Stoica, Robert Morris, David Karger, Frans Kaashoek, and Hari Balakrishnan. Chord: A scalable Peer-To-Peer
Page 133
lookup service for internet applications. In Proceedings of the 2001 ACM SIGCOMM Conference, pages 149160, 2001. [4] Beverly Yang and Hector Garcia-Molina. Designing a super-peer network. In Proceeding of 19th International Conference on Data Engineering, page 49, 2003. [5] Jabber, http://www.jabber.org. [6] Kazaa media desktop, 2001. [Online]. Available: http://www.kazaa.com/
AUTHORS PROFILE
Prasanti Boyapati, Pursuing M.Tech(CSE) from Nalanda Institute of Engineering & Technology,Siddharth Nagar, Sattenapalli, Guntur Affiliated to JNTUK, Kakinada, A.P., India. My research Interests are Network management. Dr. D. Srinivas Kumar, Working as Principal & Professor, Nalanda Institute of Technology, Siddharth Nagar, Sattenapalli, Guntur, Affiliated to JNTUK, Kakinada, A.P., India. My research Interests are Image processing,Biometrics. He is a Life member of IEEE.
Page 134

Equivalent Servant Locator in Peer-To - Peer Networks

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Equivalent Servant Locator in Peer-To - Peer Networks

Caricato da

Copyright:

Formati disponibili

International Journal of Computer Trends and Technology- volume3Issue1- 2012