P2P Overlay Architecture

P2P overlay network has a wide spectrum of communication frameworks in which peers build a self-organized system that is fully distributed and cooperative. Figure 1 depicts the abstract P2P overlay architecture showing different components of the overlay.

The Network Communication layer describes the network characteristics of an end system/node connected via internet, small wireless or sensor-based devices.

The Overlay Nodes Management layer plays its part as management of peers which includes discovery of peers and optimized routing algorithm.

The Features Management layer deals with security, reliability, fault tolerance, and robustness of P2P system.

The Service Specific layer is to provide support to Application layer to utilize the underlying resources in an efficient manner. It schedules the parallel and computation-intensive tasks, and performs task like file management and provides content. Here meta-data describes the information about content stored across the peers and their location information.

The Application level layer describes the actual functionality implemented over the underlying P2P overlay networks.

Structured P2P:

In structured overlay network the network assigns keys to data items and organizes its peers into a graph that maps each data key to a peer. Such structured P2P systems use the Distributed Hash Tables as a substructure where data objects either values or location information is placed.

Unstructured P2P:

In unstructured overlay networks the overlay network organize peers in a random graph in flat or hierarchical manner and uses flooding or random traversing or expanding Time-To-Live search on the graph to query content stored on overlay networks.

Distributed Hash Tables:

­­­­ Current widely-used P2P networks rely on central directory servers or massive message flooding, clearly not scalable solutions. Distributed Hash Tables (DHT) are expected to eliminate flooding and central servers, but can require many long-haul message deliveries.

Although many theoretical schemes for minimizing routing information have been proposed and many designs for DHTs have recently become prominent discussion topics, we are unaware of any practical and efficient system combining both.

Latest Research in P2P:

In order to get the latest that is happening in this area I decided to get some latest research papers online and went through them. According to my limited research that I did on this topic I found out that the over lay under lay mapping is very important in this domain of network models. The nodes that are that are connected logically should be close to each other physically as well in order to avoid network congestions and other delays.

Following is a brief summary of the latest research being carried out.

EGOIST Overlay Routing using Selfish Neighbor Selection

This paper discusses the issue of connectivity management, folding new arrivals into an existing overlay. Work has been done on this matter before dealing with devising practical heuristics for specific applications designed to work well in real deployments, and providing abstractions for the underlying problem that are analytically tractable, especially via game-theoretic analysis. The authors of this paper have combined these two approaches and came up with a distributed overlay routing system called “Egoist”. Connectivity management is called upon when having to wire a newcomer into the existing mesh of nodes (bootstrapping), or when having to rewire the links be- tween overlay nodes to deal with churn and changing network conditions. Connectivity management is particularly challenging for overlay networks because over- lays often consist of nodes that are distributed across multiple administrative domains, in which auditing or enforcing global behavior can be difficult or impossible.

In a typical overlay network, a node must select a fixed number (k) of immediate overlay neighbors for routing traffic or queries for files. To solve this, the authors tried and tested the selfish neighbor selection technique which was different from traditional techniques. DHTs are able to provide the best possible indexing of objects in a network. On the other hand, routing of traffic on DHTs has been shown to be sub-optimal due to local forwarding [17, 24]. Egoist can be integrated as a different layer in DHTs; when an object is mapped onto a node, Egoist is responsible to optimally route the content. In Egoist, a newcomer overlay node vi connects to the system by querying a bootstrap node, from which it receives a list of potential overlay neighbors. The new- comer connects to at least one of these nodes, enabling it to participate in the link-state routing protocol running at the overlay layer. As a result, after some time, vi obtains the full residual graph G−i of the overlay. By running all-pairs shortest path algorithm on G−i, using Dijkstra’s algorithm, the newcomer is able to obtain the pair-wise distance (delay) function dG−i . In addition to this information, the newcomer estimates dij , the weight of a potential direct overlay link from it- self to node vj , for all vj ∈ V−i. Using the values of dij and dG−i , the newcomer connects to G−i using one of a number of wiring policies. Each node listens to all the control messages of the link state protocol and propagates them only to its immediate neighbors. In order to reduce system’s control traffic, each node propagates only unique messages by dropping messages that have been received more than once or have been superseded. In Egoist, a node selects its neighbor based on best response strategy. They also employed fast approximation versions based on local search instead of long arithmetic computation which will not only reduce computational cost but also enhance scalability. Egoist’s BR neighbor

selection strategy assumes that existing nodes never leave the overlay. Therefore, even in an extreme case in which some nodes are reachable through only a unique path, a node can count on this path always being in place. This can be a one set back of this technique. Egoist also deals with the cheating nodes (using the system to route one’s own traffic but denying routing to any incoming traffic from other nodes) very efficiently in which the nodes periodically select a random subset of remote nodes and “audit them” by querying the coordinate system for the delays of the outgoing links of the audited nodes and comparing them to the actual values that the audited nodes declare on the link-state routing protocol.

Delays are natural cost metrics for many applications, especially those involving interactive communication. To obtain the delay cost metric, a node needs to obtain estimates for its own delay to potential neighbors, and for the delay between pairs of overlay nodes already in the network. Using ping, one-way delay is estimated to be one half of the measured ping round-trip-times (RTT) averaged over enough samples. The performance of the system was measured on the basis of different cost metrics such as link and path delays, node load and available bandwidth. Results of the experiments are represented graphically which showed a positive outcome and the system out performed similar previous work done relating to this. They’ve also made it available for use for any third party application through their API. This way, both the application and its Egoist node run at the same node.

This was a very comprehensive paper about the work that authors of this paper carried out. All the basic working, idea behind it and the results were shown in detail explaining every aspect if the new system that they’ve designed. This work definitely have a lot of real world application as the one they’ve explained (Multiplayer P2P gaming) which makes it an extremely appealing and the best strong point of this paper.

Other Hot Areas:

·        

    Improving Unstructured Peer-to-Peer Systems by Adaptive Connection Establishment

    ·           Improving the Interaction between Overlay Routing and Traffic Engineering

    ·         ISP-Friendly Live P2P Streaming

    ·         Scalable Resilient Overlay Networks Using Destination-Guided Detouring