Bachelor of Science (B.Sc.): Peoples’ Friendship University of Russia, Applied Mathematics and Informatics (diploma with honors)


Master of Science (M.Sc.): Peoples’ Friendship University of Russia, Applied Mathematics and Informatics (diploma with honors)


Candidate of Science: Peoples’ Friendship University of Russia, Theoretical Foundations of Computer Science


Title of docent: Theoretical Foundations of Computer Science, Ministry of Education and Science of the Russian Federation


  • Ongoing teaching responsibilities:
    • Models for Analyzing Performance Measures of Next Generation Networks
    • Multiservice Telecommunication Networks
    • Business Process Modeling
    • Electronic Business
    • Notations for Modeling Business Processes in Telecommunication Companies
    • TM Forum Frameworx: Business Process Framework (eTOM) and Information Framework (SID)
  • Supervision of students: 3 Ph.D. students, more than 15 M.Sc. theses completed, and more than 30 B.Sc. theses completed
  • Teaching award: grant of Vladimir Potanin Foundation for Young Teachers of Russian Universities (2013)

Research interests

mathematical teletraffic theory, models with real-time traffic, models with non with real-time traffic, queuing theory, queuing systems with service interruption, queuing system with unreliable servers, stochastic geometry, mathematical modeling, fifth generation wireless networks, multiservice networks, performance analysis of networks, machine type communications


mathematical teletraffic theory, queuing theory, 5G, IoT, M2M


  • H-index: 7
  • Research funding:
    • Russian Foundation for Basic Research, “Research and Development of Information and Telecommunication Technologies and Models of Wireless Machine Type Communications for Urban Infrastructure Management” (2016–2017) (No. 15-37-70016)
    • Russian Foundation for Basic Research, “Developing a Set of Markov Models for Analyzing Performance Measures of Admission Control Schemes in Priority-Based Multiservice Wireless Networks” (2016–2017) (No. 16-37-00421)
The concept of getting services and applications at «any time» and at «any place» requires the corresponding development of cellular networks, namely in LTE networks. At present, there is a lack of quality of service related recommendations describing various popular services, e.g., video conferencing. The problem is to find the optimal bit rate values for this service while not affecting the background lower priority services. In this paper we propose an instrument for solving this problem. First, we obtain mathematical model in the form of a queueing system with multicast high-priority traffic and unicast background traffic. The admission control assumes the adaptive bit rate change of multicast traffic and unicast traffic interruption. Second, we obtain the recursive algorithm for calculating mean bit rate and other performance measures. Third, we study the problem of optimizing mean bit rate.
Shared access to spectrum by several parties seems to become one of the most promising approaches to solve the problem of radio spectrum shortage. The framework proposed by ETSI, licensed shared access (LSA), gives the owner absolute priority in spectrum access, to the detriment of the secondary user, LSA licensee. The latter can access the spectrum only if the owner’s QoS is not violated. If the users of both parties need continuous service without interruptions, the rules of shared access should guarantee the possibility of simultaneous access. Balancing the radio resource occupation between parties could take quite a long time compared to the dynamics of the system due to the coordination process by the national regulation authority (NRA). We examine a scheme of the simultaneous access to spectrum by the owner and the LSA licensee that minimizes the coordination activities via NRA. According to this scheme, when the owner needs the spectrum, the power of the LSA licensee’s eNB/UEs is limited. From the LSA licensee’s perspective, the scheme is described in the form of a queuing system with reliable (single-tenant band) and unreliable (multi-tenant band) servers. We show that the infinitesimal generator of the system has a block tridiagonal form. The results are illustrated numerically by estimating the average bit rate of viral videos, which varies due to aeronautical telemetry corresponding to the owner’s traffic.
As next-generation mobile networks are rapidly taking shape driven by the target standardization requirements and initial trial implementations, a range of accompanying technologies prepare to support them with more reliable wireless access and improved service provisioning. Among these are more advanced spectrum sharing options enabled by the emerging Licensed Shared Access (LSA) regulatory framework, which aims to efficiently employ the capacity of underutilized frequency bands in a controlled manner. The concept of LSA promises to equip network operators with the much needed additional spectrum on the secondary basis and thus brings changes to the existing cellular network management. Hence, additional research is in prompt demand to determine the required levels of Quality of Service (QoS) and service provisioning reliability, especially in cases of dynamic geographical and temporal LSA sharing. Motivated by this recent urge and having at our disposal a fully-functional 3GPP LTE cellular deployment, we have committed to implement and trial the principles of dynamic LSA-compatible spectrum management. This paper is our first disclosure on the comprehensive experimental evaluation of this promising technology. We expect that these unprecedented practical results together with the key lessons learned will become a valuable reference point for the subsequent integration of flexible LSA-based services, suitable for inter-operator and multi-tenant spectrum sharing.
As the demand for higher bit rates in LTE networks increases, new technologies to enhance the effectiveness of radio resource utilization are introduced. One of them is Multimedia Broadcast Multicast Service (MBMS) that enables the usage of multicast technology for multiple resource demanding services, e. g. file repair procedure, intended for resending lost or damaged file segments to users. An important question is to find the periodicity of resending file segment, i. e. forthcoming interval for collecting user requests for repair the same file segment. To address this problem, we propose a Markov model of file repair procedure and formulas for calculating the performance measures—blocking probability, mean delay of file repair to start, and radio channel downtime probability. We also formulate and numerically solve a non-linear optimization problem for mean value of the mentioned polling interval.
While queuing theory has indeed been instrumental to various communication problems for over half a century, the unprecedented proliferation of wireless technology in the last decades brought along novel research challenges, where user location has become a crucial factor in determining the respective system performance. This recent shift turned important to characterize large cellular macrocells, as well as the emerging effects of network densification. However, the latter trend also called for increased attention to the actual user loading and uplink (UL) traffic dynamics, accentuating again the necessity of queuing analysis. Hence, by combining queuing theory and stochastic geometry in a feasible manner, we may quantify the dependence of system-level performance on the traffic loading. As performance of both session-and file-based UL transmissions has already been investigated recently in the context of heterogeneous networks, this paper explores a possibility of combining these two applications to provide a first-order evaluation of joint machine-to-machine (M2M) and human-to-human (H2H) transmissions in 3GPP LTE cellular systems. Employing a two-dimensional Markov chain for the aggregated process, we provide an approximation for the state transitions and, finally, arrive at a system-level approximation for the steady-state mode, which allows estimating a variety of system parameters averaged across space and time.
We perceive emerging developments in information and communication technologies (ICTs) in various areas of our dynamic live. The concept of getting services and applications ’any time’ and ’any place’ imposes the corresponding development in cellular networks. The rapid worldwide deployment of 4G LTE networks and increasing demand for new services are the key ICT trends that highlight the need for future improvements in technologies underling LTE. Although 3GPP concentrates all standardization activities in LTE, ITU-T recommendations reflect the main quality of service (QoS) requirements in ICTs. Nevertheless, not all QoS related ITU-T recommendations (i.e. E.800-E.899, G.1000-G.1999, Y.1500-Y.1599, Y.2100-Y.2199 series) correspond to today’s variety of products and services provided by cellular networks. This fact raises the need for updates in these recommendations. In this paper, we explore QoS measures specific to LTE services adaptively occupying radio resources. This so-called pre-emption process assumes the service degradation as well as service interruption. We propose a pre-emption based radio admission control (RAC) model for two resource demanding video services — video conferencing and video on demand. We apply teletraffic and queuing theories to get a recursive algorithm for calculating its performance measures, namely, blocking probability, pre-emption probability, and mean bit rate.
A problem of calculating the quality of the Triple Play service in an LTE mobile network under conditions of unicast and multicast transfer modes is considered. A mathematical model of the resource admission in the LTE network is constructed in the form of a system with explicit losses and three disciplines of servicing the unicast, multicast, and elastic traffics. Approximate methods of calculating the stationary probability distribution of the states of the models with unicast and elastic traffics are discussed. An approximate method based on the calculation of the marginal distribution of the number of users with unicast and multicast traffics is proposed for calculating the average transfer time of the elastic traffic. An exact algorithm for determining the stationary probability distribution of the system’s states that provides a considerable decrease in the dimension of the problem is presented.
Modern 4G wireless networks are multi-service networks, and an important issue is the development of an optimal radio admission control scheme (RAC) specific to various service types. The 3GPP recommendations for LTE and LTE-Advanced networks specify nine service classes that differ in terms of the priority level, bit-rate, and packet error loss. In this paper, we propose an RAC scheme for video telephony service that is provided in two modes: voice and video. Voice telephony users have a higher priority level than video telephony users; this is realised by the possibility of accepting voice calls owing to the degradation in video call quality. We analyse a model of a single cell with video telephony service, and we prove that the system state space is of an explicit form. This enables us to compute the performance measures, such as blocking probabilities, mean number of calls, and utilization factor.
Modern multi-service networks are inseparably linked with the commercial concept «triple play» that implies simultaneous provisioning of telephony, television (broadcast television and video on demand) and data transmission (mostly TCP-based best effort traffic) over a single broadband connection. These services generate traffics of three types — unicast streaming, multicast streaming and elastic traffics. In this paper, we propose and analyze a multi-service model of a triple play single-link network. Many research teams, including authors of the present paper, were not able to find any exact analytical solution or recurrent algorithm for models with a mixture of streaming and elastic traffics. We succeeded in developing and evaluating an approximation valid for the calculation of the elastic traffic mean transfer time for the proposed realistic traffic model
History of the development of theory of queues and mathematical teletraffic theory achieved at the MSU and PFUR are outlined. Basic theorems and algorithms for calculation of probability measures for multiservice loss networks with unicast connections, multicast connections, as well as with both types of connections are presented. A Triple Play network analysis is determined to be the key task for the future.