Grzegorz Chmaj

A P2P computing system for overlay networks

A distributed computing system is able to perform data computation and distribution of results at the same time. The input task is divided into blocks, which are then sent to system participants that offer their resources in order to perform calculations. Next, a partial result is sent back by the participants to the task manager (usually one central node). In the case when system participants want to get the final result, the central node may become overloaded, especially if many nodes request the result at the same time. In this paper we propose a novel distributed computation system, which does not use the central node as the source of the final result, but assumes that partial results are sent between system participants. This way we avoid overloading the central node, as well as network congestion. There are two major types of distributed computing systems: grids and Peer-to-Peer (P2P) computing systems. In this work we focus on the latter case. Consequently, we assume that the computing system works on the top of an overlay network. We present a complete description of the P2P computing system, considering both computation and result distribution. To verify the proposed architecture we develop our own simulator. The obtained results show the system performance expressed by the operation cost for various types of network flows: unicast, anycast and Peer-to-Peer. Moreover, the simulations prove that our computing system provides about 66% lower cost compared to a centralized computing system. Research highlights? We propose a P2P computing system providing computations and result distribution. ? The system includes decision strategies for computing peers. ? Simulations are used to evaluate the proposed approach. ? P2P flows provide lower OPEX costs compared to unicast and anycast flows. ? The distributed approach offers about 66% lower costs than a centralized system.

Progress in Systems Engineering

This collection of proceedings from the International Conference on Systems Engineering, Las Vegas, 2014 is orientated toward systems engineering, including topics like aero-space, power systems, industrial automation and robotics, systems theory, control theory, artificial intelligence, signal processing, decision support, pattern recognition and machine learning, information and communication technologies, image processing, and computer vision as well as its applications. The volumes main focus is on models, algorithms, and software tools that facilitate efficient and convenient utilization of modern achievements in systems engineering.

Modeling computational limitations in H-Phy and Overlay-NoC architectures

High performance computing demands constant growth in computational power and services that can be offered by modern supercomputers. It requires technological and designing advances in the multiprocessor internal structures as well as novel computing models considering the very high computing demands. One of the increasingly important requirements of computing platforms is a functionality that allows efficient managing computational resources, i.e., monitor them, restrict an access to some part of the resources, account for computational service, or ensure reliability and quality of service when some resources are broken or disabled. In this paper, we present a new model describing computational limitations for processing tasks on multiprocessor systems. The model is implemented in Hardware-Physical (H-Phy) and Overlay-Network-on-Chip (Overlay-NoC) architectures. Both architectures and the model are described and analyzed. Experimentation system is also presented, together with simulation assumptions, results of research and their study. The paper provides complete models of H-Phy and Overlay-NoC structures with an ability to restrict processing resources.

Heuristic algorithms for optimization of task allocation and result distribution in peer-to-peer computing systems

Abstract Recently, distributed computing system have been gaining much attention due to a growing demand for various kinds of effective computations in both industry and academia. In this paper, we focus on Peer-to-Peer (P2P) computing systems, also called public-resource computing systems or global computing systems. P2P computing systems, contrary to grids, use personal computers and other relatively simple electronic equipment (e.g., the PlayStation console) to process sophisticated computational projects. A significant example of the P2P computing idea is the BOINC (Berkeley Open Infrastructure for Network Computing) project. To improve the performance of the computing system, we propose to use the P2P approach to distribute results of computational projects, i.e., results are transmitted in the system like in P2P file sharing systems (e.g., BitTorrent). In this work, we concentrate on offline optimization of the P2P computing system including two elements: scheduling of computations and data distribution. The objective is to minimize the system OPEX cost related to data processing and data transmission. We formulate an Integer Linear Problem (ILP) to model the system and apply this formulation to obtain optimal results using the CPLEX solver. Next, we propose two heuristic algorithms that provide results very

Software Development Approach for Discrete Simulators

Simulation is the most common approach to perform the problem research. Among several types of simulation, the most common way is the discrete simulation, which assumes the division of the time scale into fixed length time slots. Depending on investigated problem, simulation packages may be used or it could be necessary to design and create own simulation system. In this paper, we propose the complete pre-study scheme and the most commonly appearing implementation problems with suggested solutions. We also describe how to implement the exemplary simulator in C++.

Overlay-NoC and H-Phy based computing using modern Chip Multiprocessors

Constant growth in demand for computational power requires advances in the internal mechanisms of multiprocessor computing structures. Such architectures may include many (sometimes, even millions of) processors performing processing tasks. Each technique that increases efficiency leads to significant benefits in operational energy and task execution time. Due the scale of multiprocessor computing structures, the importance of achieving faster and efficient systems is invaluable. In this paper, we present two different approaches for processing tasks on multiprocessor architectures: Hardware-Physical (H-Phy) and Overlay-Network-on-Chip (Overlay-NoC). Both methods are described and compared. We also present the research plan, models, simulation assumptions, and results of research. The paper is summarized with conclusions and future work plan.

Tracker-node model for energy consumption in reconfigurable processing systems

In this paper, we present an energy dissipation model for reconfigurable systems in which FPGAs have the property of online reprogramming. The proposed system contains regular nodes and one control node. Each regular node contains both CPU - capable of software processing, and FPGA unit which after being programmed with bitstream serves as the hardware processing parts. Nodes are connected in some structure and the connections form the transport layer. The system is capable of processing tasks in a distributed manner and communication, control and processing parts are taken into consideration in the energy equations. The model has also been used for algorithms that formed the complete system that is used for experimentation.

Heuristic Algorithm for Optimization of P2P-Based Public-Resource Computing Systems

In recent years network computing systems have been becoming important due to the increasing need for data processing and exchange. In this paper we focus on a public-resource computing system that uses Peer-to-Peer approach for data distribution. We assume that the considered system works on the top of an overlay network. We formulate an Integer Program optimization model of the system. Next an effective heuristic algorithm is developed to solve that model. Results of numerical experiments showing comparison of the heuristic against solutions provided by CPLEX solver are presented.

Energy-efficient distributed computing solutions for Internet of Things with ZigBee devices

In recent years, we are witnessing a trend with increasing number of networked and computing-capable devices being integrated into everyday environment. This trend is expected to continue. With computing devices available as logic structures, they might use each others processing capabilities to achieve a given goal. In this paper, we propose an architecture to process tasks using distributed structure of Internet of Things devices. We also include ZigBee devices that are not connected to the Internet, but participate with the processing swarm using local network. This significantly extends the flexibility and potential of the IoT structure, while being still not a well-researched area. Unlike many high-level realizations for IoT processing, we present a realization operating on the communications, computing and near protocol level that achieves energy consumption efficiency. Our work is suitable to be the base for higher-level realizations, especially for systems with devices operating on battery power. At the same time, the architecture presented in this paper uses minimal centralization, moving maximum responsibilities to regular devices. The proposed realizations are described using linear programming models and their high efficiency is evaluated.

A Knowledge Based System for Minimum Rectangle Nesting

Nesting algorithms deal with the optimal placement of shapes in specified regions subject to specified constraints. In this paper, a complex algorithm for solving two-dimensional nesting problem is proposed. Arbitrary geometric shapes are first quantized into a binary form. These binary representations are subsequently processed by operators which nest the shapes in a rectangle of minimum area. After nesting is completed, the binary representations are converted back to the original geometric form. Investigations have shown that the nesting effect is driven by quantization accuracy. Therefore, better accuracy is possible given more computing time. However, the proposed knowledge based system can significantly reduce the time of nesting, by intelligently pairing shapes, based on prior knowledge of their form.