Reliability is one of the bigest challenges faced by service-oriented systems. Therefore, to solve this problem, we have proposed ReServE - Reliable Service Environment. ReServE increases fault-tolerance of SOA systems and ensures consistent processing despite failures. However, the proposed environment imposes also the performance overhead. Thus, in this paper, we extended ReServE and added a monitoring feature provided by the M3 service. As a consequence, the extended environment can adjust appropriately the load of its modules to the changing interaction and behaviour patterns of service oriented systems. We have experimentally shown that the proposed solution, while providing the required level of reliability, decreases significantly the performance overhead.
Radosław Januszewski, Rafał Różycki and Grzegorz Waligóra
Many fields of modern science rely more and more on the immense computing power of supercomputers. Modern, multi-thousand node systems can consume megawatts of electrical energy in highly uneven manner, challenging the data center infrastructure, both power and cooling coils. The traditional way of managing the infrastructure makes each subsystem of a data center (e.g. cooling) independent from all other in the way it relies only on local sensors to manage the infrastructure. The erratic nature of computing in a large data center makes this approach suboptimal. In the paper we show that by challenging the traditional split between the infrastructure and the computing equipment, one can gain significant boost in energy efficiency of the entire ecosystem. A solution that predicts cooling power demand basing on the information from a supercomputer resource manager, and then sets up the parameters of the cooling loop, is presented along with potential benefits in terms of reduction of the power draw.
Mohammad Aldabbas, Francesca Venteicher, Lenna Gerber and Marino Widmer
This paper addresses the issue of selecting a suitable location for a fire station in canton of Fribourg, as a result of a fire brigades’ merger, by applying Multiple Criteria Decision Analysis (MCDA) methods. Solving the problem of determining fire station locations through various methods has been analyzed in-depth by researchers. However, a different approach, based on application of ELECTRE I and ELECTRE II methods is advanced in this paper. The selection of the most suitable fire station site is obtained by applying the designated methods to five distinctive alternatives (called scenarios), taking into consideration the relatively limited information and specifics, and the extensive number of relevant criteria that summed up to sixty-one. Taking the merger of the three local fire departments as an example, the proposed methods for selecting a suitable location for the fire station demonstrate and justify the reason behind this choice. Research shows that the applied methods have been proven to be useful and powerful tools that exhibited acceptable levels of consistency when selecting the best project. The main finding is that one scenario in particular proved to be preferred over the others and most suitable in determining the fire station location.
Darja Solodovnikova, Laila Niedrite and Aivars Niedritis
Today, many efforts have been made to implement information systems for supporting research evaluation activities. To produce a good framework for research evaluation, the selection of appropriate measures is important. Quality aspects of the systems’ implementation should also not be overlooked. Incomplete or faulty data should not be used and metric computation formulas should be discussed and valid. Correctly integrated data from different information sources provide a complete picture of the scientific activity of an institution. Knowledge from the data integration field can be adapted in research information management. In this paper, we propose a research information system for bibliometric indicator analysis that is incorporated into the adaptive integration architecture based on ideas from the data warehousing framework for change support. A data model of the integrated dataset is also presented. This paper also provides a change management solution as a part of the data integration framework to keep the data integration process up to date. This framework is applied for the implementation of a publication data integration system for excellence-based research analysis at the University of Latvia.
Vadim V. Romanuke
The present paper considers an open problem of setting hyperparameters for convolutional neural networks aimed at image classification. Since selecting filter spatial extents for convolutional layers is a topical problem, it is approximately solved by accumulating statistics of the neural network performance. The network architecture is taken on the basis of the MNIST database experience. The eight-layered architecture having four convolutional layers is nearly best suitable for classifying small and medium size images. Image databases are formed of grayscale images whose size range is 28 × 28 to 64 × 64 by step 2. Except for the filter spatial extents, the rest of those eight layer hyperparameters are unalterable, and they are chosen scrupulously based on rules of thumb. A sequence of possible filter spatial extents is generated for each size. Then sets of four filter spatial extents producing the best performance are extracted. The rule of this extraction that allows selecting the best filter spatial extents is formalized with two conditions. Mainly, difference between maximal and minimal extents must be as minimal as possible. No unit filter spatial extent is recommended. The secondary condition is that the filter spatial extents should constitute a non-increasing set. Validation on MNIST and CIFAR- 10 databases justifies such a solution, which can be extended for building convolutional neural network classifiers of colour and larger images.
Rinalds Vīksna and Gints Jēkabsons
Social networking sites such as Facebook, Twitter and VKontakte, online stores such as eBay, Amazon and Alibaba as well as many other websites allow users to share their thoughts with their peers. Often those thoughts contain not only factual information, but also users’ opinion and feelings. This subjective information may be extracted using sentiment analysis methods, which are currently a topic of active research. Most studies are carried out on the basis of texts written in English, while other languages are being less researched. The present survey focuses on research conducted on the sentiment analysis for the Latvian and Russian languages.
Kristiāns Kronis and Marina Uhanova
The paper describes the implementation of organic benchmarks for Java EE and ASP.NET Core, which are used to compare the performance characteristics of the language runtimes. The benchmarks are created as REST services, which process data in the JSON format. The ASP.NET Core implementation utilises the Kestrel web server, while the Java EE implementation uses Apache TomEE, which is based on Apache Tomcat. A separate service is created for invoking the benchmarks and collecting their results. It uses Express with ES6 (for its async features), Redis and MySQL. A web-based interface for utilising this service and displaying the results is also created, using Angular 5.
Ainārs Auziņš, Jānis Eiduks, Alīna Vasiļevska and Reinis Dzenis
During object-relational database physical structure design, problems are caused by three factors: ambiguity of transformations of conceptual model, multiplicity of quality assessment criteria, and a lack of constructive model. In the present study a constructive hierarchical model of physical database structure has been developed. Implementations are used in XML, SQL and Java languages. Multi-criterial structure optimisation method has also been developed. Structure variation space is generated using transformation rule database. Prototype has been implemented within the framework of the research.
Vladimir Roganov, Michail Miheev, Elvira Roganova, Olga Grintsova and Jurijs Lavendels
The development of new software to improve the operation of modernised and developed technological facilities in different sectors of the national economy requires a systematic approach. For example, the use of video recording systems obtained during operations with the use of endoscopic equipment allows monitoring the work of doctors. Minor change of the used software allows using additionally processed video fragments for creation of training complexes. The authors of the present article took part in the development of many educational software and hardware systems. The first such system was the “Contact” system, developed in the eighties of the last century at Riga Polytechnic Institute. Later on, car simulators, air plan simulators, walking excavator simulators and the optical software-hardware training system “Three-Dimensional Medical Atlas” were developed. Analysis of various simulators and training systems showed that the computers used in them could not by themselves be a learning system. When creating a learning system, many factors must be considered so that the student does not receive false skills. The goal of the study is to analyse the training systems created for the professional training of medical personnel working with endoscopic equipment, in particular, with equipment equipped with 3D indicators.