Research Publications

During my post-graduate studies at Univesity of Hradec Kralove I coauthored many research publications published by IEEE Xplore and Springer Link.

Orchestration and Automation of NVF

Advances on P2P, Parallel, Grid, Cloud and Internet Computing | Published by Springer in 2018

This paper explains how to handle automation of Network Function Virtualization. There is an increase of popularity of moving out from legacy hardware networking devices to virtual network appliances. Vendor’s goal is to virtualize their physical products so these products can be used in cloud environments. There are also community projects like OPNFV. In our work we propose to way how to automatically orchestrate these Network Function Virtualization devices.

Performance Analysis of Cloud Computing Infrastructure

Mobile Web and Intelligent Information Systems | Published by Springer in 2017

This paper explains how to work with performance measurement tools to be able to make qualified decision during the election of the right cloud platform. The popularity of the cloud computing is increasing rapidly and there is a variety of cloud platforms from Open-source world as well as from the vendors. Choosing the right solution, either public cloud from large number of providers or private cloud, can significantly differ in cost and performance. This brings a lot of challenges to find suitable solution one for end user workload.

Metric Based Cloud Infrastructure Monitoring

Advances on P2P, Parallel, Grid, Cloud and Internet Computing | Published by Springer in 2018

Goal of the paper is to design dynamic, model-driven solution for agile service and network monitoring based on metrics suitable for managing large infrastructures and clustered applications in baremetal or cloud environments. The solution uses decetralised model where each node computes its states and passes the results to upper level cluster monitoring services. This is opposite approach than traditional monitoring where all logic is stored on central monitoring node or cluster of nodes.

Our proposed Monitoring solution offers automation of repetitive administration tasks and makes creation of monitoring infrastructure part of routine configuration tasks and make use of meta-data models already in place. Various monitoring scenarious are covered as Cluster and inter datacenter monitoring, advanced time-series analytics, corellations and trending with prediction algorythms and querying log event databases for unusual patterns. Emphasis is given to limiting the amount of output when failures occur presented to human operators, reducing the time to solve the problem and automating solution of most common issues.

Hardware Layer of Ambient Intelligence Environment Implementation

Computational Collective Intelligence | Published by Springer in 2017

Ambient Intelligence is growing phenomena caused by advances of speed and size of computational hardware. It is possible to collect data from various sensors, devices or services and react to evaluated event and start the predefined process. These systems are closely connected with Cloud based services as they provide publicly available interface for management and visualization, computational and storage capabilities for forecasts and other advanced data processing to create ambient intelligence environment. The true ambient environments react to the presence of people and the sensors and actuators have become part of the environment.

This article presents RoboPhery project aimed to provide abstraction layer for interfacing any low cost hardware sensors and actuators with MQTT and Time-series Database bindings, that can serve as sensory network for Ambient Intelligence environments. The service architecture is designed to be so simple at hardware level to support single-board micro-controllers like ESP2866, ESP32 modules as well as single-board computers based on ARM or x86 architectures. The communication among individual devices is handled by the standard MQTT messages. The same devices can be configured to support multiple use-cases on configuration level to lower the operational costs of the solution.

VNF Orchestration and Modeling with ETSI MANO Compliant Frameworks

Internet of Things, Smart Spaces, and Next Generation Networks and Systems | Published by Springer in 2017

The goal of this paper is to compare existing NFV (Network Function Virtualization) management solutions and propose SaltStack based solution with the same goal. Modern approach to govern NFV automation in large scale IT infrastructures is done by virtualization of various network services. Hardware networking devices are gradually replaced by virtual network appliances for the lower acquisition and maintenance costs. Vendors virtualize their physical products so they can be used along or within cloud environments. The ETSI (European Telecommunications Standards Institute) standard MANO (NFV Management and Organization) aims to unify the management of physical and virtual network services and devices to single point of control for configuration and management.

The paper covers the major NFV frameworks that use MANO as their reference architecture with focus on underlying interface technologies. The main focus of the paper is NFV-MANO orchestration engine using SaltStack capabilities that can model and manage heterogeneous NFV and NFVI resources according to the ETSI NFV-MANO specification along software resources already present. This gives us unique ability to cover all data center resources, both hardware or software based in a single model.

OpenSource Automation in Cloud Computing

Proceedings of the 4th International Conference on Computer Engineering and Networks | Published by Springer in 2015

This chapter covers basic OpenSource automation and configuration management tools that can be used to alleviate common operations tasks and processes in cloud systems. It shows a quick survey of major cloud computing solutions and introduces simple abstraction layer for the management of physical and virtual resources. The last chapter covers some of the common automation scenarios from both cloud computing provider and consumer perspectives and their possible open-source implementations.

Network Visualization Survey

Computational Collective Intelligence | Published by Springer in 2015

This paper compares modern approaches to draw complex graph data that create create compelling visualizations. Graphs are used to represent more and more complex systems that are used across various scientific domains. Some graph visualizations are used to model network topologies or service architectures in information service fields, model genomes in biomedicine or complex molecules in chemistry. Visualizations of graph data play important role in interpreting the meaning of graphed data. The adage “A picture is worth a thousand words” generally refers to the notion that a complex idea can be conveyed with just a single still image and that is actually the main reason for creating visualisations of any kind. Graph drawings from all domains follow the same rules covered by the graph theory. This work outlines different layout strategies used for drawing graph data. Tested graph drawing layouts are compared by standard quality measures to determine their suitability for various areas of usage.

High Level Models for IaaS Cloud Architectures

New Trends in Intelligent Information and Database Systems | Published by Springer in 2015

This paper explains how ontology can be used to model various IaaS architectures. OpenStack is the largest open source cloud computing IaaS platform. It has been gaining wide spread popularity among users as well as software and hardware vendors over past few years. It’s a very flexible system that can support a wide range of virtualization scenarios at scale. In our work we propose a formalization of OpenStack architectural model that can be automatically validated and provide suitable meta-data to configuration management tools. The OWL-DL based ontology defines service components and their relations and provides foundation for further reasoning. Model defined architectures can support simple all-in-one architecture as as well as large architectures with clustered service components to achieve High Availability.

Enterprise ICT Transformation to Agile Environment

Computational Collective Intelligence | Published by Springer in 2015

This paper explains how ontology can be used to model various IaaS architectures. OpenStack is the largest open source cloud computing IaaS platform. It has been gaining wide spread popularity among users as well as software and hardware vendors over past few years. It’s a very flexible system that can support a wide range of virtualization scenarios at scale. In our work we propose a formalization of OpenStack architectural model that can be automatically validated and provide suitable meta-data to configuration management tools. The OWL-DL based ontology defines service components and their relations and provides foundation for further reasoning. Model defined architectures can support simple all-in-one architecture as as well as large architectures with clustered service components to achieve High Availability.

Measurement of Cloud Computing Services Availability

Nature of Computation and Communication | Published by Springer in 2015

This paper explains how to work with performance measurement tools to be able to make qualified decision during the election of the right cloud platform. The popularity of the cloud computing is increasing rapidly and there is a variety of cloud platforms from Open-source world as well as from the vendors. Choosing the right solution, either public cloud from large number of providers or private cloud, can significantly differ in cost and performance. This brings a lot of challenges to find suitable solution one for end user workload.

Gateway Redundancy Protocols

IEEE 15th International Symposium on Computational Intelligence and Informatics | Published by IEEE in 2014

This paper covers how to increase the availability of data networks. Issue of high available services is very actual today. Redundancy is usually solved by the Spanning Tree Protocol on the 2nd layer of ISO/OSI model. This paper is dedicated to redundancy on 3rd layer of ISO/OSI model. The goal of paper is to explore possible solutions for the default gateway redundancy to increase availability of external networks. The survey is focused on FHRP (First Hop Redundancy Protocol) protocols, description of their functionality, network communication, and basic configuration. Next aim of the paper is comparison of these protocols and real time measurements of protocol performance. Outcome of the paper is functional description of HSRP, VRRP and GLBP protocols. Meaning of parameters in communication messages of HSRP and VRRP protocols is explained further. Paper contains basic configuration for every protocol with brief explanation of configuration commands. Measurements delay of choosing new active gateway has been performed on physical topology with physical devices. Latest version of FHRP was used during these measurements. The measured results indicate that performance of tested FHRP protocols is almost equivalent.

Private IaaS Cloud Platform Design

Univerzity of Hradec Kralove | Diploma Thesis | Published in 2013

The diploma thesis focuses on the issue of virtualization and cloud services. This is a dynamic area that has recently been undergoing stormy development, both technically and commercially. The theoretical part of the work describes the history of virtualization and the main virtualization and cloud solutions in terms of their impact on the market with these technologies.

In the practical part, selectable solutions for the implementation of a private IaaS platform. The selected platform is then deployed in two reference installations. The first installation will be in a virtualized VirtualBox environment. The second reference installation will be in the CEPSOS network laboratory at UHK FIM.

Digital Resource Scraping in Knowledge Management System

Univerzity of Hradec Kralove | Bachelor Thesis | Published in 2010

The goal of this bachelor thesis is to create a web application for managing and searching for digital resources. The application will have a clearly defined semantic layer that is sufficiently flexible and cooperates with other metadata standards.

A modern user interface is used to create and manage digital resources. Resources can be browsed using taxonomic and facet classifications or searched for by entered terms.