Dependability Properties of P2P Architectures

A system’s dependability can be thought of as being the trustworthiness of the system. The difficulty when attempting to measure dependability is that it is typically a context sensitive property. While one user might regard a system to be dependable for the particular activities they use it for, another user might regard it to be totally undependable for their activities. As well as being context sensitive, dependability is also regarded as being multi-dimensional, in that it can be influenced by a variety of other architectural properties. These properties include security, reliability, availability and performance. Because of this influence it is difficult to consider dependability without also considering these additional properties. Furthermore, these properties can in turn be affected by other sets of properties. Consequently, a system’s dependability can often be influenced by factors that would not typically be regarded as dependability properties in their own right. Because of this network of property influences within a P2P system, it is necessary to identify and consider all properties that can exist within P2P architectures, and to relate these to the dependability of the system. Identifying dependability properties and achieving dependability within P2P systems is further complicated by the fact that numerous P2P logical architectures exist and no single architecture is likely to be suitable for all application types. For example, applications such as Napster [1] ultimately benefit most from a semi-centralised architecture, where as FreeNet [2] is most suitable to be run over a decentralised architecture. The type of architecture used can influence the dependability properties of the system. Take for example, security. Fully decentralised P2P systems are likely to be better suited at handling denial of service attacks, semicentralised P2P systems would be better suited for handling peer certification. This paper aims to identify the main dependability properties (and related properties) that can play a part within P2P systems. This, in turn, can be used to help inform the creation of more dependable systems. Given the influence the choice of architecture can have, this paper first provides an overview of the main P2P architectures, before going on to identifying the different properties. Future work will provide a detailed analysis of the effect the architectures can have on these properties.