Dependability within peer-to-peer systems

Abstract

There is an increasing interest in using architectures as a basis for software systems. However, by their very nature, achieving dependability within a P2P system can be difficult. This paper provides an initial analysis of the main issues that need to be considered when developing a dependable P2P system. It examines the key properties that can influence the dependability of a P2P system, and discusses the relationship between system dependability and the choice of logical network architecture. Dependability and Peer-to-PeerA system's dependability can be thought of as being its trustworthiness [9]. 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 undependable for their activities. Traditionally dependability has also been regarded as multidimensional, in that it can be influenced by a variety of other attributes. Key attributes include availability, reliability, responsiveness, safety and security [9].Peer-to-Peer (P2P) computing has become very popular in recent years. Essentially it can be thought of as a class of application that takes advantage of the resources and services that are available at the edge of the Internet [8]. There is an increasing interest in using P2P as a basis for software systems within industry, where it can be used to support activities such as communication between workers. However for P2P technology to be adopted within such an environment it also needs to be dependable and achieving dependability within a P2P system can be difficult.In particular P2P systems possess a number of specific properties that can have an influence on the system's dependability attributes (security, reliability, etc). For example, the type of peer discovery mechanism used can influence the responsiveness of a P2P system. A broadcast discovery mechanism can result in slower performance than with using a centralised peer lookup server. Consequently, when considering dependability within P2P system design, it is also becomes necessary to consider these specific P2P properties.Dependability within P2P systems is further complicated by the numerous P2P logical network architectures (abstractions of the underlying physical network) that exist and no single architecture is likely to be suitable for all application types. For example, Napster [7] benefits most from a semi-centralised architecture (provides more efficient resource searching), whereas a decentralised architecture is more suitable for FreeNet (better anonymity support) [2]. The different types of logical network architecture can also influence the P2P properties and general system dependability. Decentralised P2P systems are likely to be better suited at handling denial of service attacks, the central authority provided by semi-centralised systems would be better suited for handling peer certification. Designers, when deciding on a suitable logica...

A survey for measuring and managing employee dependability with EBusiness

“…For example, Boehm et al [21] suggest accuracy and performance are attributes of dependability; Walkerdine et al [22] also indicate responsiveness as a key attribute of dependability.…”