I.V. Kartashevskiy - Ph. D. (Eng.), Associate Professor, Volga State University of Telecommunications and Informatics (Samara). E-mail: ivk@psati.ru

Due to the high load radio (with limited frequency resources) network traffic at the access layer has a pronounced self-similar property. In the case when the correlation properties of the sequences of inter-arrival times and sequence of processing times can be disregarded we show that a model of access to the software-defined radio frequency resource can be represented as a queuing system G/G/1, which allowed any heavy-tailed distributions for time slots. Queuing system G/G/1 can be replaced with queuing system HK/HL/1 on the basis of the solution approximation problem. The aim of replacing this systems is to simplify the analysis of the system G/G/1. After replacement it can be carried out using a method based on the assumption of ergodicity for sequence of waiting time intervals and taking into account the rational form of the Laplace transform of the exponent. Parameters of hyper-exponential distributions are determined on the basis of solving the problem of the approximation of an arbitrary probability density function with decaying exponential polynomial. Here shown the method of queuing systems HK/HL/1 analysis which allows to estimate the average waiting time for requests in the queue. To substantiate the calculation reliability we make a comparison with the results obtained by the classical method for queuing system G/G/1 proposed by L. Kleinrock. Also a comparison of the results of the calculation of the average waiting time in queue with the results for the queuing system H2/H2/1, upon receipt hyper-exponential distribution parameters by equating the first three points of the corresponding distributions HK and G.

 

1. J. MitolaIII. CognitiveRadio. AnIntegratedAgentArchitectureforSoftwareDefinedRadio.  DoctorofTechnologyDissertation, RoyalInstituteofTechnology, Sweden. May 2000.
2. J. MitolaIII , G.Q. Maguire, Jr. Cognitiveradio: makingsoftwareradiosmorepersonal // IEEEWirelessCommunications. 1999. V. 6. № 4. P. 13−18.
3. Dibakar Das, Alhussein A.AbouzeidDelayanalysisofmultihopcognitiveradionetworksusingnetworkofvirtualpriorityqueues // Proceedingsof 2014 IEEEWirelessCommunicationsandNetworkingConference (WCNC). P. 1514−1519.
4. Arash Azarfar, Jean-François Frigon, Brunilde Sansò.Priorityqueueingmodelsforcognitiveradionetworkswithtrafficdifferentiation // EURASIPJournalonWirelessCommunicationsandNetworking. 2014.
5. S. Wang, J. Zhang, L. Tong. Delayanalysisforcognitiveradionetworkswithrandomaccess: afluidqueueview // INFOCOM. 2010.
6. Hung Tran, Trung Q Duong, Hans-Jürgen Zepernick. Delayperformanceofcognitiveradionetworksforpoint-to-pointandpoint-to-multipointcommunications// EURASIPJournalonWirelessCommunicationsandNetworking 2011. 2012:9.
7. Hamid Farmanbar, Hang Zhang.Trafficengineeringforsoftware-definedradioaccessnetworks // Proceedingsof 2014 IEEENetworkOperationsandManagementSymposium (NOMS). P. 1−7.
8. SHelukhin O.I. Multifraktaly. Infokommunikacionnye prilozhenija M.: Gorjachaja Linija - Telekom. 2011. 578 s.
9. Nazarov A.N., Sychev K.I. Modeli i metody rascheta pokazatelejj kachestva funkcionirovanija uzlovogo oborudovanija i strukturno-setevykh parametrov setejj svjazi sledujushhego pokolenija. Izd. 2-e, pererab i dop. Krasnojarsk: Izd-vo OOO «Polikom». 2011. 491 s.
10. Klejjnrok L. Teorija massovogo obsluzhivanija. M.: Mashinostroenie. 1979. 432 s.
11. Kartashevskiy I.V., Tarasov V.N.Approximationofinputdistributionsforqueuingsystemwithhyper-exponentialarrivaltime // ProceedingsofIEEEConferenceProblemsofInfocommunications. ScienceandTechnology (PICS&T). 2015. SecondInternationalScientific-PracticalConference. 13−15 Oct. 2015. P. 15−17.
12. Blatov I.A., Kartashevskijj V.G., Kireeva N.V., CHupakhina L.R. Approksimacija proizvolnojj plotnosti raspredelenija summami ehksponent // Vestnik Voronezhskogo gosudarstvennogo universiteta. Ser. Sistemnyjj analiz i informacionnye tekhnologii. 2013. № 2. S. 53−58.
13. Feldmann A, Whitt W.Fittingmixturesofexponentialstolong-taildistributionstoanalyzenetworkperformancemodels // PerformanceEvoluation 31. 1998. P. 245−279.
14. Keilson J., Machihara F.HyperexponentialwaitingtimestructureinhyperexponentialHK/HL/1system // JournaloftheOperationSocietyofJapan. September 1985. V. 28. № 3. P. 242−250.
15. Kartashevskijj I.V., Tarasov V.N. Sposoby approksimacija vkhodnykh raspredelenijj dlja sistemy G/G/1 i analiz poluchennykh rezultatov // Sistemy upravlenija i informacionnye tekhnologii. 2015. № 3.1(61). S. 182−185.