P2P Networking For Information Sharing: Client-Server Vs P2P Networking

Napster and Gnutella

Discuss about the P2P Networking for Information Sharing for Typical Client-Server.

In a typical client-server architecture that is most commonly used by organizations for information sharing within and out of their networks, a single server is used for storing information. In such a case, whenever, information is required by any member or node connected to network and the same to be shared to any other node, the member would take the file from this server, download it, and then perform sharing. Users have somewhat passive roles to play in this information sharing, as they do not contribute to information but only updates the repository (In Tae & Youn Sung, 2015).

On, the other hand, a new technology that has emerged over the past few years involves a Peer-to-Peer (P2P) network for information sharing. In such a network, any information can be made available, by any of the peer connected to the network, for distribution to the network. In such a network, any peer can connect to any other peer node and can download the data. This technology provides equal rights and capacities to all the connected nodes for information sharing (Parameswaran, Susarla, & B.Whinston, 2001).

Some examples of P2P information sharing networks include Napster that has a centralized repository, which is replicated using P2P network for sharing of music in MP3 versions and Gnutella, which is a file sharing network that allows users to share files (Oram, 2001).

Figure 1: Napster Architecture (Din, Nutanong, & Buyya, 2003)

Napster: Users can search and share MP3 files using Napster, which works of the centralized model of P2P, file sharing network architecture. In this model, a central server is used to maintain directories of MP3 files being shared over registered nodes and these directories are updated every time a user logs in. Client looking for files can automatically connect the a meta-server acting as an arbiter which is actually a less loaded server selected from the cluster and can handle up to 15000 users load. This server allows connected users to access and share files in the Napster community (Back, 2003).

Figure 2: Gnutella Topology (Din, Nutanong, & Buyya, 2003)

Gnutella: It uses a flat network of peer nodes called servents and these nodes maintain the directory of the content system, acting both as a server and as a client as and when required. A node can choose to accept or reject a connection based on available connection slots, protocol versions and so on. To avoid flooding in the network, Time-To-Live field is included in header of every message, which is decremented each time it passes through a servant such that when the TTL reaches a value of zero, message is dropped (Vollet, Ingham, & Ezhilchelvan, 2003).

Uses

P2P networks can majorly be used in two types of information sharing applications including multitasking and active networking.

P2P Multitasking: A multimedia streaming content provider can use P2P network for multitasking by sending data from one source and replicating the same on different nodes in far downstream allowing the new node to act as a secondary source. Such data can thus be easily distributed to the local user community making it easier for them to download content from any node, which would be in proximity (Parameswaran, Susarla, & B.Whinston, 2001.

P2P Active Networking: Active networks can be used to carry codes that can be used to identify traffic usage patterns such that decisions about content redistribution can be taken dynamically based on P2P and active networks. This can also help in establishing certain level of control on participants as well as enforce guidelines for preventing distribution of malicious codes.

P2P information sharing technology provides some major advantages over the regular client-server architecture used for information sharing such as enhanced load balancing, creation of dynamic repositories of information, improved tolerance for faults, content-based addressing, and improved searches.

Improved Load Balancing: proactive load-balancing schemes can be used in a P2P network as traffic and information item profiles can be used for deciding strategies for redistribution of content by locating closest locations for high demand items. These techniques require less sophisticated solution than required in a distributed caching system.  Further, strategies like automatic content relocation and replication can be used for dynamic monitoring of usage patterns such that content can be effectively replicated for optimizing the load balance.

Dynamic Repositories for Information:  In a P2P network, the information on one node can be easily shared and replicated across different nodes. Thus, the content availability grows with more spread over the network. With the combined networks containing comprehensive collection of all the content available for a community, even least searched items can be easily searched. For instance, most rare or an old song’s MP3 file can also be searched as easily and fast as any high search MP3 file in a Music repository working on P2P model (Buford, Yu, & Lua, 2009).

Redundancy and Tolerance for Faults: As P2P, networks allow replication of information across different nodes; the redundancy is high thereby enhancing the availability of information. Moreover, if information has a high level of demand, the same can be replicated to more nodes making it faster to reach and download for users. Moreover, with this decentralization of information across nodes, if one node fails in the network, it would not fail the entire network, making the system more faults tolerant (Lin, Lin, & Wang, 2008).

Benefits

Content-based addressing: On web pages, URL’s may not directly relate with the content they have. In a P2P network, the content stored at a node is made transparent such that the grouping of web pages addresses is done based on the content that specific nodes are carrying. This type of addressing makes searches easier for users who do not specific URLs but content identifiers. With more uniform resource identifiers in the form of content-based indexing, a more refined information repository that is easy and fast to search can be formed (Fukuda-Parr & Lopes, 2002).

Figure 3: Search model comparison (Parameswaran, Susarla, & B.Whinston, 2001

Better Searches: Internet search queries are executed by searching content only from public databases, which is not immediately updated especially in case the respective server goes down. On the other hand, in a P2P network information-sharing model, user node is indexed only when a user is online and the index is continuously synchronized with the status. At the time of search, a P2P network would not need to depend on search engine robots but can use the dynamically index content making it easier for P2P communities to access and share information within community (Parameswaran, Susarla, & B.Whinston, 2001.

The technology no doubt provides several technological advantages but it also comes with certain limitations such as:

Traffic Redistribution Challenge:  In P2P networks, traffic coming to one node can be redistributed to other nodes such that the information is replicated. If this traffic were needed to be reduced then it would require the source to be located close to the destination, which requires development of complex algorithms requiring high computing power. Thus, not all available computing resources can actually meet the computing power requirements of P2P networks and thus, it may not remain feasible for all information sharing solutions (Pagani, 2007).

Free-rider problem: As any one node can connect to any other node in the network, it becomes easy for one node to steal and replicate quality information available on other node. Such an information can then be bundled into offers where the information acts as one component to replicate the same over network treating the same as own.  In lack of capability to differentiate between the original and replicated content, it would become difficult for the receivers to identify if the information received is from originator or replicator. Moreover, receivers may not even care about the source as long as they are getting the required information (Krishnan, Smith, & Telang, 2002).

Limitations

Legal and regulatory Issues: Such a model presents opportunities for users to do copyrights infringement, malicious code propagation, and intellectual piracy and so on, as no control may be imposed on the type of information that is being shared by users in most P2P solutions. This problem only is resolved through a self-regulatory system that may be established for a community using the network. However, it can still face security risks as well as litigations from other interest groups demanding government intervention that can further lead to complete shutdown of activities if found objectionable by government (Zhang, Shi, Wang, & Zhang, 2007).

Noise: There is a possibility that the information listed in the P2P network may be cluttered with a significant level of noise making it difficult for users to get effective information (Parameswaran, Susarla, & B.Whinston, 2001.

Conclusions

Traditional file sharing systems used centralized server-client architecture in which every time a file is requested by a user, the request goes to server for its retrieval. In P2P network architecture, the centralized repository is actually replicated across nodes and a directory is maintained centrally such that every time, a user needs to fetch or share a file, the closest node can be reached to complete the operation. Such a technology has several benefits such as improved load balancing, dynamic repositories, improved fault tolerance; content based addressing and improved searches. However, the network also faces certain challenges or limitations such as requirement of high power for traffic redistribution, free-rider problem, noise in information shared, and legal or regulatory issues.

References

Back, T. (2003). P2P Information Sharing in Mobile Ad-hoc Networks. Helsinki University of Technology.

Buford, J., Yu, H., & Lua, E. K. (2009). P2P Networking and Applications. Mordan Kaufmann.

Din, C. H., Nutanong, S., & Buyya, R. (2003). Peer-to-Peer Networks for Content Sharing. Australia: The University of Melbourne.

Fukuda-Parr, S., & Lopes, C. (2002). Capacity for Development: New Solutions to Old Problems. New York: EarthScan Publications Limited.

In Tae, L., & Youn Sung, K. (2015). Technology Information Sharing and Technology Innovation Performance: An Empirical Study of the Mediating Role of Technology Development Capability. International Journal of u- and e- Service, 217-230.

Krishnan, R., Smith, M. D., & Telang, R. (2002). The Economics of Peer-To-Peer Networks . Pittsburgh,PA: Carnegie Mellon University.

Lin, F.-r., Lin, S.-c., & Wang, Y.-f. (2008). Can Peer-to-Peer Networks Facilitate Information Sharing in Collaborative Learning? IEEE, 2-10.

Oram, A. (2001). Peer to Peer: Harnessing the Benefits of Disruptive Technologies. O’Reilly & Associates.

Pagani, M. (2007). Encyclopedia of Multimedia Technology and Networking. Italy: IGI Global.

Parameswaran, M., Susarla, A., & B.Whinston, A. (2001). P2P Networking: An Information-Sharing Alternative. ICS.

Vollet, E., Ingham, D., & Ezhilchelvan, P. (2003). JMS on Mobile Ad-hoc Networks. Personal Wireless Communications (PWC) (pp. 8-15). Venice, Italy: IFIP WG 8.6.

Zhang, Y., Shi, H., Wang, X., & Zhang, J. (2007). Collaborative Legal Information Sharing on P2P Network. IEEE, 335-371.