Inter national J our nal of Electrical and Computer Engineering (IJECE) V ol. 9, No. 4, August 2019, pp. 3221 3227 ISSN: 2088-8708, DOI: 10.11591/ijece.v9i4.pp3221-3227 r 3221 P arallelising r eception and transmission in queues of secondary users Said Lakhal, Zouhair Guennoun Mohammadia School of Engineering, Mohammed 5 Uni v ersity in Rabat, ERSC formerly kno wn as LEC, Research Center E3S, Morocco Article Inf o Article history: Recei v ed Jun 29, 2018 Re vised Mar 13, 2019 Accepted Mar 21, 2019 K eyw ords: Cogniti v e radio netw ork Queue management T ransmission delay Data shifting Input flo w Output flo w ABSTRA CT In a cogniti v e radio netw ork, the secondary users place the pack ets to be transmitted on a queue, for controlling the order of arri v al, and adapting to the netw ork state. The pre vious conceptions assigned to each secondary user a single queue, which contains both: recei v ed and forw arded pack ets. Our present article di vides the main queue into tw o sub queues: one to recei v e the arri v ed pack ets, and the other to transmit the a v ailable pack ets. This approach reduces the transmission delay dues to, the shift- ing of data, placed on the single queue, and to the sequential process ing of reception and transmission. All, without increasing the memory capacity of the queue. Copyright c 2019 Insitute of Advanced Engineeering and Science . All rights r eserved. Corresponding A uthor: Said Lakhal, Mohammadia School of Engineering, Mohammed 5 Uni v ersity in Rabat, ERSC formerly kno wn as LEC, Research Center E3S, Ibn Sina A v enue, B.P 765, Agdal, Rabat, Morocco. 212697620603 Email: said.lakhal.rech@gmail.com 1. INTR ODUCTION The demand of spectra is increased in last tw o decades, grace of the intense transmission of v oices and videos via the netw ork. Therefore, the classical conceptions became inable to support these ne w challenges. As a result, the Federal Communications Commission [1, 2] decided to modify its spectrum allocation strate gy , with the aim of adopting a more fle xible polic y . These ef forts led to the birth of the f amous cogniti v e radio netw ork (CRN), by Mitola [3,4]. In this netw ork, the primary users (PUs) and the secondary users (SUs) alternate for e xploiting the spectra. The PUs ha v e priority to access spectra. By cons, the SUs w ait for the release of a fe w spectra in order to transmit data [5, 6]. During this w aiting period, a particular SU can accumulate multiple pack ets to send. These pack ets are or g anized on a queue to: k eep them, mark the order of arri v al of each one, and adapt to the netw ork fluctuations. The time elapsed between the arri v al of a pack et and its transmission is called the w aiting time or delay , which is in v ersely proportional to the throughput. In the classic approach, each SU is assigned a unique queue, reserv ed for both recei ving and forw ard- ing pack ets. A L yapuno v optimization technique is used in [7-9], to stabilize the queue and design an online flo w control. Both [10] and [11] ha v e tar geted the maximization of the SU’s throughput. In [10], a h ybrid queue management policies is proposed, and in [11], the authors introduced mean throughput maximization scheduling protocol which schedules the a v ailable SUs. The authors of [12] proposed a repeat queuing. Each pack et goes through three phas es: reception, shift and transmission [13, 14]. This design sequentially treats the reception and trans mission process, so the phase of shifting pack ets on the queue tak es a certain amount of J ournal homepage: http://iaescor e .com/journals/inde x.php/IJECE Evaluation Warning : The document was created with Spire.PDF for Python.
3222 r ISSN: 2088-8708 time, which will increase the delay of pack ets on the queue. C on s equently , it decreases the throughput through the CRN. In this article, instead of considering a single queue, we design tw o subqueues: one for the reception of the pack ets and the other for the transmission. When this last is emptied, the sub queues e xchange their roles. The si ze of each one equal to half the size of the queue, considered in the classical approach. Thus, the memory capacity to st ore the pack ets will not increase, the reception and transmission processes are parallely treated, as well as, the phase of shifting pack ets on the queue, will not tak e place. W ith this approach, we concei v e a ne w management procedure of the tw o sub queues. Therefore, we are reducing the delay , relati v e to the con v entional approaches. 2. RALA TED W ORKS Se v eral researchers ha v e addressed the SUs’ management queues. A L yapuno v optimization technique is used in [7], for controlling the partition of users into groups, which are modelled by the graph collaring, in order to share channels and stabilize queue according. The simulations demonstrated the lo w-comple xity of the proposed model. Based on the same optimization technique [8], according to the collision queues, the authors designed an online flo w control and resource allocation algorithm; in the aim to maximize the SUs’ throughput, subject to maximum collision constraints with the PUs. As a result, the desired objecti v e is reached. Al w ays for meeting the same objecti v e, a pack et of w orks are de v eloped in [10, 11, 15]. The authors of [10] proposed a h ybrid queue management policies (QMP) interwea v e/o v erlay , and an adapti v e QMP . The e v aluations sho wed that the h ybrid approach, leaded to the best SUs’ throughput, compared to the con v entional schemes. A priority queue scheduling algorithm is formulated in [11]; to a v oid collision between heterogeneous nodes, during data transmission, and impro v e the entire netw ork t hroughpu t . Another scheduling technique is presented in [15], for increasing the basic QoS parameters. The principle of such technique, consists of di viding the netw ork into tw o re gions, each one is controlled by a particular base station, and the spectrum is allocated on a priority basis according, to real-time and non-real-time data. The e xperimentations sho wed a decrease in terms of delay , collision probability , end-t o - end delay and o v erhead ratio; as well as, an increase in terms of the netw ork ef ficienc y and throughput. The pre-emption and non-preemption priorities attracted man y attentions in the pre vious w orks. In this topic, we will in v estig ate the authors’ contrib utions in [16-18]. A h ybrid approach is e xposed in [16], at which lo w priority SUs are no longer pre-empted by high priority SUs, when their number of interruptions reaches a certain threshold v alue. Therefore, the authors sho wed that the threshold adjustment according to the performance metric pro vided a promising performance. In [17], the authors presented a queuing model, pro viding the accurate a v erage system time, for general pack ets service time, and service interruption peri- ods, with an opportunistic spectrum access (OSA) netw orks. The y sho wed that, for the same a v erage CR transmission link a v ailability , the pack et system time significantly increases in a semi-static netw ork, with long operating and interruption periods, compared to an OSA netw ork with f ast alternating operating, and interruption periods. The pack ets are grouped with dif ferent prioriti es in a queue [18], represented by tw o dimensional state transition graph. The simulations demonstrated t w o results: First, the decrease of the a v erage w aiting time of high priority pack ets, with the gro wth interference po wer threshold. Second, the proportionality between the lo w priority pack et a v erage w aiting time, and the arri v al rate of the high priority pack et. F or modelling the characterize spectrum handof f beha viours with general service, the authors of [12] proposed a repeat queuing. After that , the y deri v ed the close-e xpression of the e xtended data deli v ery , and the system sojourn time in both: staying and changing scenarios. The analysis of spectrum handof f beha viours resulting from multiple inter - ruptions, clarified the traf fic-adapti v e polic y and the admissible re gion. As it is kno wn, the SUs dynamically allocate the free channels. F or this purpose, the authors of [19] proposed a dynamic channel-selection solu- tion, and a priority virt ual queue interf ace that determines the requir ed information e xchanges, and e v aluates the e xpected delays e xperienced, by v arious priority traf fics and competing users’ beha viours. Based on a dy- namic strate gy learning algorithm deplo yed at each user , the y significantly reduced the pack et loss rate, and outperformed the con v entional single-channel dynamic resource allocation. T w o types of retrial customers and a paired service are serv ed by a single system in [20]. After solving a Riemann boundary v alue problem, the authors determined the joint stationary orbit queue length distrib ution at service completion epochs. After the emer gence of the cogniti v e radio technology , se v eral applications ha v e Int J Elec & Comp Eng, V ol. 9, No. 4, August 2019 : 3221 3227 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Elec & Comp Eng ISSN: 2088-8708 r 3223 emer ged; among them we cite, b ut not limited to: smart grid [21], safety [22], military [23], wireless body area netw orks [24] and surv eillance [25]. 3. OPERA TIONS OF THE TW O MODELS In the con v entional approach, each SU has a single queue, on which are placed the pack ets to send and those to transmit. Based on the principle: first come, first serv ed; the netw ork manager shifts the data to the output of the queue aft er each transmission, to f acilitate transmission and reception at the same time. Unlik e our approach, which considers the same size of the queue, b ut se gmenting it into tw o sub-queues: one for the reception and the other for the transmission. By using this technique, the data shift operation on the queue will not tak e place, and as soon as the transmission sub-queue is emptied, the controller of the transmission will be mo v ed to the last reception sub queue, and so the sub queues e xchange their roles. The Figure 1 illustrates the operating of our approach (O A) and that of the classical approach (CA), i.e., by Q t and Q r , we designate the transmission and reception queues, respecti v ely . Figure 1. Operating of our model and that of the classical 4. MODELIZA TION OF THE TW O APPR O A CHES When the output flo w is greater than t he input flo w , each arri v ed pack et is transmitted before the arr i v al of the ne xt; therefore, the saturation problem does not arise. Otherwise, since a gi v en moment, the queue becomes unable to recei v e the arri v ed pack ets, and so, the saturation state appears. This study is interested in the second case. In the follo wing, we will model the saturation problem in our approach and that of the con v entional, i.e. T able 1 contains all used symbols in this modelization. P ar allelising r eception and tr ansmission in queues of secondary user s (Said Lakhal) Evaluation Warning : The document was created with Spire.PDF for Python.
3224 r ISSN: 2088-8708 T able 1. Symbols and their meanings Symbols Meanings os; S Q; I ; O Necessary time for shifting one byte in the CA, size of queue, input flo w , output flo w , resp. v 0 i ; w 0 i T ime spent, current size of queue after the i th transmission and before the 1 st saturation, resp, in the CA. z 0 n T ime spent after the n th transmission and before the 1 st saturation, in O A. d T ime to mo v e to the ne xt pack et on the queue, in O A. v i m T ime spent, after the i th saturation and the m th transmission, in the CA. w i m T ransmitted quantity after the i th saturation and the m th transmission, in the tw o approaches. z i m T ime spent after the i th saturation and the m th transmission, in O A. 4.1. Classical appr oach (a) Before the 1 st satruration v 0 1 = 2 , one unit for w aiting the reception and the other for transmitting. v 0 2 = v 0 1 + os (2 I O ) + 1 . Iterati v elty: v 0 n = v 0 n 1 + os ( nI ( n 1) O ) + 1 = v 0 n 1 + n os ( I O ) + os O + 1 W e put a = os ( I O ) ; b = os O + 1 . v 0 n = ( n 1)( n + 2) 2 a + ( n 1) b + 2 (1) w 0 n = ( n + 1) I n O = n ( I O ) + I (2) The saturation of queue arri v es when w 0 n S Q , i.e, n S Q I I O . The saturation threshold is indicated by: n 0 = int ( S Q I I O ) (b) After the 1 st saturation v 1 1 = v 0 n 0 + 1 v 1 m = v 1 m 1 + os ( S Q O ( m 1)) + 1 (3) W e put: c = os S Q + 1 (4) v 1 m = v 1 m 1 ( b 1)( m 1) + c v 1 m = m ( m 1) 2 ( b 1) + c ( m 1) + v 0 n 0 + 1 (5) Since the input flo w is greater than that of the output, the queue will accept the pack ets only after checking the follo wing condition: w 1 m I , i.e. m O I , i.e. m I O . Then, the threshold acceptation is: m 0 = int ( I O ) . The queue accepts the arri v ed pack et at the m th 0 transmission, after it rejects all arri v ed pack ets at: m 0 + 1 ; ::::; 2 m 0 1 transmissions. (c) After the i th saturation v i m = m ( m 1) 2 ( b 1) + c ( m 1) + v ( i 1) m + 1 v i m = i ( m 1) h m 2 ( b 1) + c i + v 0 n 0 + i (6) The queue accepts the arri v ed pack et at the im th 0 transmission, after it rejects all arri v ed pack ets at: im 0 + 1 ; ::::; ( i + 1) m 0 1 transmissions. Int J Elec & Comp Eng, V ol. 9, No. 4, August 2019 : 3221 3227 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Elec & Comp Eng ISSN: 2088-8708 r 3225 4.2. Our appr oach T able 2. sho ws the approachment in stages T able 2. Our approach Before the 1 st satruration After the 1 st saturation After the i th saturation z 0 1 = 2 z 1 1 = z 0 n 0 + 1 z i m = ( m 1) + z i 1 m + 1 z 0 n = z 0 n 1 + d + 1 z 1 m = z 1 m 1 + 1 z 0 n = ( n 1)( d + 1) + 2 (5) z 1 m = ( m 1) + z 0 n 0 + 1 (6) z i m = i ( m 1) + z 0 n 0 + i (7) The queue accepts the arri v ed pack et at the im th 0 transmission, after it rejects all arri v ed pack ets at: im 0 + 1 ; ::::; ( i + 1) m 0 1 transmissions. 4.3. Comparison between the tw o appr oaches Based on relation (1), v 0 is quadratic according the transmission i teration n . Then, its curv e is a parable. Besides a > 0 , therefore, this parable is con v e x. In the other hand, z 0 is e xpressed in relation (5) as a linear fonction, al w ays by referring to the transmission iteration and assuming that ( d + 1) > 0 , thus, the curv e of z 0 is an increasing line. Relation (4) e xpresses the time of the m th transmission after the i th saturation, in the classical ap- proach. This time is linear according to the saturation iteration. In the other side, relation (7) presents the v ariation of time depending on the saturation iteration, in our approach. 5. SIMULA TION By choosing positi v e v alues of a, b and d, we obtain the curv es of v 0 and z 0 , illustrated in the left part of Figure 2. Kno wing that: m 2 ( b 1) + c >> 1 and v 0 n 0 > z 0 n 0 , as a result we obtain the right part of Figure 2, illustrating the spent time according to the saturation iteration, for the tw o approaches. (a) (b) Figure 2. Comparison between the tw o models in terms of delay before and after the first saturation. In Figure 2, we remark that for the same transmission iterations, the delay in the classical appaoch is greater than that in our approach. This result, can be e xplained by tw o f actors: 1) The delay increase dues to the shift of the queue data in the cl assical model. 2) The arrangement of tw o queues, one for the reception and the other for the transmission, mak es it possible to carry out a parallel processing between reception and transmission, and so, we g ain more time in our model. P ar allelising r eception and tr ansmission in queues of secondary user s (Said Lakhal) Evaluation Warning : The document was created with Spire.PDF for Python.
3226 r ISSN: 2088-8708 6. CONCLUSION In this w ork, we ha v e de v eloped a queues management mechanism, based on the di vision of the main queue into tw o sub queues: one for the reception of the arri v ed pack ets and the other for the transmission of the a v ailable pack ets. W ith this design, we ha v e reduced the transmission delay dues to the shift of the data on the single queue in the classic design. Also, the consideration of tw o queues parallely ensures the transmission and reception. As a result, the time of the sequential treatment is g ained. REFERENCES [1] P .V arade et, al., ”Throughput Maximization of Cogniti v e Radio Multi Relay Netw ork with Interference Management, International J ournal of Electrical and Computer Engineering ( IJECE) , v ol. 8, no. 4, pp. 2230-2238, 2018. [2] I.Mustapha, et al., ”An Ener gy Ef ficient Reinforcement Learning Based Cooperati v e Channel Sensing for Cogniti v e Radio, Sensor Netw orks, Perv asi v e and Mobile Computing, 2016. [3] D. Damodaram and T .V enkatesw arlu, ”Ef ficient Hardw are Architecture for Cyclostationary Detector , Bulletin of Electrical Engineering and Informatics , v ol. 5, no. 3, pp. 340-346, 2016. [4] J. Mitola, ”Cogniti v e Radio An I nte grated Agent Architecture for Softw are Defined Radio, Ph.D. disser - tation, KTH Ro yal Institute of T echnology , Stockholm, Sweden, 2000. [5] M. K. Kaushik, et al., ”Sensing and Sharing Schemes for Spectral Ef ficienc y of Cogniti v e Radios, Inter - national J ournal of Electrical and Computer Engineering (IJECE) , v ol. 8, no. 5, pp. 2934-2941, 2018. [6] L.T ang and J.W u, ”Research and Analysis on Cogniti v e Radio Netw ork Security , W ireless Sensor Net- w ork, V ol.4, pp. 120-126,2012. [7] Y ituW ang et, al., ”Heterogeneous Spectrum Aggre g ation: Coe xistence from a Queue Stability Perspec- ti v e, IEEE T r ansactions on W ir eless Communications , 2018. [8] R.Ur g aonkar and M.J.Neely , ”Opportunistic Scheduling with Reliability Guarant ees in Cogniti v e Radio Netw orks, IEEE T r ansactions on Mobile Computing , v ol. 8, no. 6, 2019. [9] C.Qiu et, al., ”L yapuno v Optimized Cooperati v e Communications W ith Stochastic Ener gy Harv esting Relay , IEEE Internet of Things J ournal , v ol.5, No.2, pp. 1323-1333, 2018. [10] K.A.Mehr , ”Queue Management for T w o-User Cogniti v e Radio with Delay-Constrained Primary User , Computer Netw orks, 2018. [11] S.Sodag ari, ”Real-T ime Scheduling for Cogniti v e Radio Netw orks, IEEE Systems J ournal , v ol. 12, no. 3, pp. 2332-2343, 2018. [12] X.Y et, al., ”Spectrum Handof fs Based on Preempti v e Repeat Priority Queue in Cogniti v e Radio Net- w orks, Sensor s , v ol. 16, no. 7, pp. 1-19, 2016. [13] A.Azarf ar et, al., ”Analysis of Cogniti v e Radio Netw orks Based on a Queueing Model with Serv er Inter - ruptions, IEEE ICC - Co gnitive Radio and Networks Symposium , pp. 1703-1708, 2012. [14] T .M.N.Ng atched et, al., ”Analysis of Cogniti v e Radio Netw orks with Channel Assembling, Buf fering, and Imperfect Sensing, IEEE W ir eless Communications and Networking Conf er ence , Shanghai,China, Apr . 7-10 , pp. 952-957, 2013. [15] P .D.Rathika and S.Sophia , ”A Distrib uted Scheduling Approach for QoS impro v ement in Cogniti v e Radio Netw orks, Computer s and Electrical Engineering , v ol. 57, pp. 186-198, 2017. [16] T .E.F ahim, ”A No v el Hybrid Priority Discipline for Multi-Class Secondary Users in Cogniti v e Radio Netw orks, Simulation Modelling Pr actice and Theory , v ol. 84, pp. 69-82, 2018. [17] A.Azarf ar et, al., ”Priority Queueing Models for Cogniti v e Radio Netw orks with T raf fic Dif ferentiation, EURASIP J ournal on W ir eless Communications and Networking , 2014(1):206, December 2014. [18] C.Long et, al., ”Queueing Analysis for Preempti v e T ransmission in Underlay Cogniti v e Radio Netw orks, International J ournal and Communication Systems , v ol. 29, no. 6, 2016. [19] H.P .Shiang and M.V .D.Schaar , ”Queuing-Based Dynamic Channel Selection for Heterogeneous Multi- media Applications Ov er Cogniti v e Radio Netw orks, IEEE T r ansactions on Multimedia , v ol. 10, no. 5, pp. 896-909, 2008. [20] I.Dimitriou, ”A Queueing Model with tw o Classes of Retrial Customers and P aired Services, Spring er US , v ol. 238, no. 1, pp. 123-143, 2016. [21] VC.Gungor and D.Ahin, ”Cogniti v e Radio Netw orks for Smart Grid Applications: A Promising T echnol- ogy to Ov ercome Spectrum Inef ficienc y , IEEE V ehicular T ec hnolo gy Ma gazine , pp. 41-46, 2012. Int J Elec & Comp Eng, V ol. 9, No. 4, August 2019 : 3221 3227 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Elec & Comp Eng ISSN: 2088-8708 r 3227 [22] A.A.Alkheir and H.T . Mouftah, ”Cogniti v e Radio for Public Safety Communications, In book: W ireless Public Safety Netw orks V ol. II, Edition: first, Chapter: 10, Publisher: Else vier , Editors: Daniel C ˆ amara, Na vid Nikaein, pp.1-22, 2016. [23] T .J. W illink, ”SDR and Cogniti v e Radio for Military Applications, Emer ging W ireless T echnologies, pp. 1-20, 2007. [24] R.C.Santiago and I.Balasingham, ”Cogniti v e Radio for Medical Body Area Netw orks Using Ultra W ide- band, IEEE W ir eless Communications , pp. 74-81, August 2012. [25] U.S.Premarathne et, al., ”Secure and Reliable Surv eillance o v er Cogniti v e Radio Sensor Netw orks in Smart Grid, Perv asi v e and Mobile Computing, 2015. BIOGRAPHIES OF A UTHORS S. Lakhal obtained the diploma of application engineer in computer sciences in 1998, from the Uni v ersity Sidi Mohamed Ben Abdelah, Fes, Morocco, M.Sc. de gree in modelization in 2006, from Mohammadia School of Engineering. He is currently a researc her at the Laboratory of Electronics and T elecommunications, Mohammadia School of Engineers (EMI), Rabat, Morocco. His current research interests are Computing, Radio cogniti v e, Algorithmic and comple xity , Modelization. Z. Guennoun recei v ed his engineering de gree in Electronics and T elecommunications from the Electronics a nd Electrical Montefiore Institute, ULG Lie ge, Belgium in 1987; his M.Sc. de gree in Communication Systems from the EMI School of Engineering, Rabat, Morocco in 1993; and his PhD de gree from the same school in 1996. He visited the Centre for Communication Research (CCR) in Bristol Uni v ersity , UK, during the period of 1990-1994 to prepare a split PhD. During 1988-1996 he w ork ed as an Assistant Lecturer in the EMI School of engineering, and from 1996 he is w orking in the same school as a Professor Lecturer . His fields of interest are digital signal processing, error control coding, speech and image processing. Currently in char ge of the laboratory of Electronics and T ele communications (LEC) at EMI. IEEE member since 1990; and member of the Moroccan IEEE section e x ecuti v e committee. P ar allelising r eception and tr ansmission in queues of secondary user s (Said Lakhal) Evaluation Warning : The document was created with Spire.PDF for Python.