DOI: 10.7256/2306-4196.2014.4.12800

Abstract: In today’s world the problem of processing and storing huge amounts of data becomes increasingly pressing. Messages in social networks, photos, streaming video – altogether creates a high load on the server-side software. For this reason common approaches used in desktop-software design may be ineffective since they don’t take into account the huge load on the application created by the vast number of users. Currently, there is no clear definition for highly-loaded systems. In most cases this term is used in situations, when software fails to operate under some momentary load. There’s no specific values set at which a system can be considered highly-loaded, since each software is different and same amount of requests can lead to completely different loads on the resources. The given study of the database management systems consisted of several experiments, measuring the speed of common operation on databases, such as adding, selecting and deleting. Based on the result of these experiments the author makes conclusions and gives recommendations on choosing the database management system. The article reviews approaches in developing highly loaded systems, highlights their features and disadvantages and shows examples of the use of these approaches in popular web-services such as ÂÊîíòàêòå, Facebook, Google and ßíäåêñ. The articles brings a comparative analysis of MySQL and MongoDB database management systesms. In conclusion the author gives recommendations on selecting a database management system depending on the approach to designing architecture of a highly-loaded project.

DOI: 10.7256/2306-4196.2014.4.12648

Abstract: The article reviews some aspects of practical use of the pseudo-random number sensors in computational mathematics and cryptography. In particular the author studied inappropriate behavior of the Monte Carlo method in solving the task of risk assessment of uniform credit portfolio using regular pseudo-random number sensor of the Microsoft Office programming system. The article identifies limits of the regular pseudo-random number sensor. The author reconstructs it in terms of single-module residue arithmetic and on that basis proves unreasonableness of its use in cryptographic applications and explained certain aspects of inadequate behavior of the Monte Carlo in the given example. The article proposes a solution as alternative pseudo-random number sensor based on the Mersenne twister for solving the complex tasks of computational mathematics. The article shows results of corresponding numerical experiments. The research is based on the numerical experiments based on the Monte Carlo method. The reconstruction and study of the pseudo-random number sensor of the Microsoft Office programming system involved number and theoretic methods. Probability theory formalism is used in formulating of the model problem and interpreting the results of its solution. The paper for the first time shows the comparison of the "Value at Risk" results for a task of risk assessment of uniform credit portfolio received using Monte Carlo method with exact values, calculated using the methods of the algebraic theory of risk. This comparison allowed to determine the limitations for Monte Carlo based methods and other methods using regular pseudo-random number sensor of the Microsoft Office programming system. The author proposes alternative solution for the problem and shows its adequacy in the corresponding numerical experiment.

DOI: 10.7256/2306-4196.2014.4.12677

Abstract: The article is devoted to the problem of computer simulation of biochemical analysis of biological fluids using the Stat Fax® 1904 + R. The author describes methodology and stages of creating an educational simulator (formulation of the problem, project planning, defining the system, gathering of input data, translation of the model, verification and validation, implementation of the system into the educational process). The article presents main methods for calculation of substances’ concentration, imitated by the computer simulator (based on the standard, factor calculation, multi-point calibration and kinetic characteristics measuring). The author pays significant attention to the interface design of the application’s main window that visualize virtual desktop for carrying out biochemical analysis. The essential elements of the simulated desktop are the images of the device with working keyboard, display, printing area for displaying the results of analysis, tubes (blank, standard, sample) with corresponding buttons, as well as the power button. The task of simulation of the quantitative and qualitative characteristics of the test mixture is solved using random number generator. The results of the presented study can be used in building educational computer simulators not only in medicine but also in other areas of professional activity. 

DOI: 10.7256/2306-4196.2014.4.12683

Abstract: The course of physics in colleges and universities includes various wave processes: reflection and transmission of the pulse through the interface between two media, interference, wave propagation in a dispersive medium, the formation and interaction of solutions. Here it is important to combine theoretical and practical approaches to studying of these phenomena with computer models, allowing creating visual image of the phenomenon and analyzing its behavior in different conditions. The subjects of the study are the simple computer models and computational experiments helping to show the wave processes in the one-dimensional linear and nonlinear media. The experiments require mathematical and computer modeling, building a mathematical model, creating software simulating the studied phenomenon based on the numerical solution of the corresponding system of equations. The novelty of the work is in the fact that the author presents three simple computer programs written in Pascal, simulating pulse propagation in one-dimensional medium, its reflection from the boundary between two media, and its passage of the second medium, the wave propagation in a dispersive medium, the formation of different solutions and their interactions. The analysis of the results of computer modeling allows to state that the use of such computational experiments based on the simulation of a one-dimensional medium by a system of coupled springing or simple pendulums or solving the sine-Gordon equation really allows to study the wave processes at a higher level and to form interest in physics and information technologies. 

DOI: 10.7256/2306-4196.2014.4.12758

Abstract: The authors study arbitration system for data threads between ports of modern serial PCI Express bus. The article is devoted to the development of a model for that system. For the model the authors make an assumption, that switching matrix of the switch is non-blocking. Authors describe principles of the software model operation that allows to study the arbitration algorithm and the dependence of flow characteristics upon various factors. Using the mentioned above model the authors examines the influence of the virtual channels quantity and of the unevenness of the input ports load of the commutator on the amount of buffering memory of the port and virtual channels’ arbiters. The article uses computational experiment based on the software model of PCI Express bus arbitration as a method of the research. The produced model is based on the concepts and algorithms regulated by the official protocol specification for PCI Express. The authors present a software model of multistage arbitrator switch for PCI Express, in which varies the set of parameters of the arbitration specific for the real bus load on the cluster systems. The modular approach allows to modify the software model and include different priority schemes. The model designed and described in the article may be used in the building of the switch structure, as well as in configuration of an arbitrator, which can be useful when creating a cluster system with external PCI Express switches, which found practical use relatively recently. The model can be also used during the study of the PCI Express protocol.

DOI: 10.7256/2306-4196.2014.4.12606

Abstract: The report reviews tasks that shall be solved during interprocessor interface modeling in parallel computing system with combined (mixed) topology based on hypercubes: binary, generalized, generalized ring, generalized ring hypercube with multiple buses. The first three of those structures can be modeled based on the last one because the algorithms of specifying their topologies are practically the same. Modeling of the hypercube with multiple buses has its own specifics because the presence of buses considerably affects diameter, capacity and complexity of the inteprocessor interface structure, because the data can be transferred between the nodes of the structure using both direct connections and buses. As the result of this the it is important to find the optimal path for transferring data taking into account the state of closest nodes, congestion of bus (buses) linked to the current node. For that purpose a status word is formed, representing busyness of a bus and direct links of connected to the node, the state of the node itself, data acceptance, transfer or transit. To simplify selection of the direction and rout the model uses the cost of a bus. Simulation of routing and transactions are rather simple because the structure has a high degree of dimension.  The main element of the modeling system is switching processor that carries out the main functions of the commutation environment. The process of developing the modeling system has the following requirements for the designing of the software based on the iterative-incremental model using object-oriented approach and DBMS due to the need to work with large amount of data at the high speed of processing. It also requires high-level programming languages, which allows building real-time applications with features of multithreading and scalability. The project uses analytical methods of the theory of computing systems, simulation method, methodology of object-oriented design. This article discusses problems related to the study of existing and emerging types of parallel computing system topologies based on hypercubes. The novelty of the study lies in the fact that for the first time the topology of the parallel computing system based on the generalized hypercube with multiple buses is presented. The main parameters are defined, the characteristics are studied on the base of the simulation, the comparison of the parameters with the well-known hypercube topologies is given.