While the first phase of QuaP2P examined the single quality aspects and some of their bilateral interdependencies, the second phase focuses on the methodology of researching multilateral interdependencies as well as on the characterization of these interdependencies. Therefore, a benchmark methodology for P2P systems will be developed, allowing for the investigation of the quality aspects of existing P2P systems.

For us, the term benchmark describes a test or a group of tests which were designed to compare the performance of a system to other (already existing) systems. Through the utilization of benchmarking and its methodology, basic P2P mechanisms from different contemplated layers of a P2P system (see Figure) shall be evaluated while also quantifying and characterizing the multilateral dependencies between the different quality aspects.

To achieve the mentioned goals, the second phase of QuaP2P shall deliver significant contributions for (1) a more objective assessment of the qualitative comparability of P2P mechanisms and for (2) a quantitative evaluation of them. Moreover, resulting from the tight composition and interaction of the diverse application scenarios, we expect to make system wide statements dealing with (3) the general selection and (4) the concrete parameterization of the single mechanisms. Other long-term objectives of our research project shall answer the questions (5) in which areas a P2P system outperforms a centralized solution, (6) how this influences the systematic development of such systems and (7) which applications can be set up on top of P2P architectures.

A P2P system consists of different mechanisms which provide the functionality of the system. To improve and control the quality of a P2P system, new or modified mechanisms are required. Therefore, the QuaP2P research project investigates new and existing mechanisms for P2P systems which may control and ideally improve the quality of the systems based on the different quality aspects defined during the first phase of the project.

For the second phase of QuaP2P, the research of quality of P2P mechanisms is not examined along the different layers, as it was practiced during the first phase. This type of investigation led to an isolated observation of single quality aspects which inhibit the detection of multilateral dependencies between several aspects. Therefore, the quality of P2P mechanisms is observed from a functional point of view. This new approach allows for a comparison of different mechanisms providing the same functionality. The comparison is realized through benchmarking of the mechanisms, enabling quantitative evaluation as well as to identify the multilateral interdependencies between the different quality aspects of a mechanism.

To perform our research on important P2P mechanisms, we first identified fundamental P2P mechanisms. These fundamental mechanisms are considered as important elements for existing and future applications based on the P2P paradigm. Using our layered architecture of a P2P system (see Figure), we identified them for each single layer which facilitates an extensive evaluation of the various mechanisms in our research group.

Our architecture consists of three functional, horizontal layers (search overlay layer, service layer, application layer) and one non-functional, vertical layer (real-time service layer).

At each layer, the mechanisms, providing the fundamental functionality, will be examined through benchmarking in terms of their quality.

Each of the four sub-projects of the DFG research group is assigned to a particular layer. At each layer, the mechanisms, provided by the respective layer, are evaluated through a specific benchmark that will be exclusively created for the mechanisms of that layer.

Altogether, the four different benchmarks will be combined in a benchmark-suite which allows for comparing the different mechanisms of a particular layer, identifies the multilateral interdependencies of the quality aspects and respects the functional dependencies between the four layers.