Abstract
We consider a two-link system with restricted accessibility that accommodates Poisson arriving calls from different service-classes and propose a multirate teletraffic loss model for its analysis. In a restricted accessibility system, a new call may be blocked even if available bandwidth units do exist at the time of its arrival. In the two-link system, each particular link has two thresholds which refer to the number of in-service calls in the link. The lowest threshold, named support threshold, defines up to which point the link can support calls offloaded from the other link. The highest threshold, named offloading threshold, defines the point where the link starts offloading calls to the other link. The two-link system with restricted accessibility is modelled as a loss model whose steady state probabilities do not have a product form solution. However, we propose approximate formulas for the calculation of call blocking probabilities. The accuracy of the formulas is verified through simulation and found to be quite satisfactory.
Similar content being viewed by others
1 Introduction
Quality of Service (QoS) mechanisms are necessary in contemporary communication networks in order to provide the required bandwidth needed by calls. In the case of call-level traffic in a single link, modelled as a loss system, such a QoS mechanism is a bandwidth sharing policy [1]. The simplest bandwidth sharing policy is the Complete Sharing (CS) policy. In the CS policy, a new call has full access to all non-occupied bandwidth units (b.u.) of the link and is accepted in the link if the required bandwidth is available. Otherwise, call blocking occurs. Often the terms ‘full accessibility’ or ‘full availability’ are used in the literature instead of the term ‘CS policy’ [1, 2]. However, the term ‘availability’ may also refer to the proportion of time the link is available [3]. Herein, we prefer the term ‘CS policy’.
The simplest teletraffic loss model that adopts the CS policy is the classical Erlang model [1]. In this model, the call arrival process is Poisson while each call requires one b.u. to be accepted in the system. If this b.u. is available then a call remains in the link for a generally distributed service time. Otherwise call blocking occurs. The fact that Call Blocking Probabilities (CBP) are calculated via the classical Erlang B formula has led to numerous extensions of Erlang’s model for the call-level analysis of wired (e.g., [4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]), wireless (e.g., [19,20,21,22,23,24,25,26,27,28,29,30]), satellite (e.g., [31,32,33]) and optical networks (e.g., [34,35,36,37,38,39]).
In the recent paper of [27], a two-link loss system is studied that accommodates Poisson arriving calls from a single service-class. Each link may support and provide service to calls offloaded from the other link. An offloaded call is a new call that arrives in a link but will be served (subject to bandwidth availability) by the other link. This offloading mechanism is achieved with the aid of two thresholds per link, a low and a high threshold. Both thresholds refer to the number of in-service calls in each link. The first (low) threshold is the support threshold while the second (high) one is the offloading threshold. The support threshold refers to the point up to which the link can support offloaded calls from the other link. On the other hand, the offloading threshold refers to the point where call offloading can start from one link to the other.
The model of [27] does not have a Product Form Solution (PFS) for the steady state probabilities. This is due to the fact that the offloading mechanism destroys Local Balance (LB) between adjacent states of the system. In order to calculate CBP, either a linear system of Global Balance (GB) equations should be solved or an approximate method that relies on the independence between the links and the classical Erlang B formula can be adopted.
A potential application of the offloading scheme of [27] is in the area of mobile/WiFi networks. To manage the increasing traffic in mobile networks, traffic can be offloaded to WiFi networks [40, 41]. To further increase the available bandwidth of WiFi access links, recent research focuses on the aggregation of backhaul access link capacities and on the bandwidth sharing policies that should be adopted. The impact of such an aggregation to CBP in the case of a single service-class can be well studied by the offloading scheme of [27].
In this paper, we extend the model of [27] to include the case of multirate Poisson traffic, i.e., we consider that the system accommodates calls of different service-classes with different bandwidth requirements. In addition, we incorporate the notion of restricted accessibility. In a restricted accessibility system, a new call may be blocked even if available b.u. do exist at the time of its arrival. Under the general term ‘restricted accessibility’ one may include:
-
bandwidth sharing policies such as the threshold policy [15, 17], the probabilistic threshold policy [28, 42] or even the bandwidth reservation policy [4, 9, 13]. In the threshold policy, a new call (of a particular service-class) is not allowed to enter a system if the number of in-service calls (of that service-class) together with the new call exceeds a predefined threshold. In the probabilistic threshold policy, a new call may enter the system with (a state dependent) probability when the number of in-service calls plus the new call exceeds the threshold. In the bandwidth reservation policy, a new call will be blocked when the only available b.u. at the time of its arrival in one link are already reserved for other calls (e.g., of the same or the other link).
-
the case where each particular state of the system (except from the initial state where the system is empty) is associated with a blocking probability. This means that a new call may be blocked and lost with a certain blocking probability when the system is in a particular state at the time of its arrival. Such an approach is useful when modeling interference between neighbouring cells (e.g., in CDMA systems) [3, 43]. In this paper, we focus on this type of restricted accessibility and study the case where the blocking probability is different for each state. To this end, we provide an approximate method for the CBP determination which is verified via simulation. The CBP calculation in the proposed two-link model is based on the classical Erlang Multirate Loss Model (EMLM) [44, 45], which refers to a single link.
The remainder of this paper is organized as follows: In Sect. 2, we review the system of [27]. In Sect. 3, we propose the extension of [27] which includes the case of multirate traffic and the notion of restricted accessibility. In Sect. 4, we provide analytical and simulation CBP results for the proposed model. We conclude in Sect. 5.
2 Review of the Two-Link System for Single-Rate Traffic
We consider a loss system of two links with capacities \(C_1\) and \(C_{2}\) b.u., respectively. Each link accommodates Poisson arriving calls of a single service-class which require one b.u. in order to be connected in a link. Let \(\lambda _{l}\) be the arrival rate in link l (l = 1, 2). Also let \({j}_{l}\) be the occupied b.u. in link l. Then, \(0\le j_{1} \le C_{1}\) and \(0\le j_{2} \le C_{2}\). Since a call requires one b.u. the value of \({j}_{l}\) also expresses the number of in-service calls in link l.
Each link l (l = 1, 2) has two different thresholds: the support (low) threshold \({\textit{th}}_{1l}\) and the offloading (high) threshold \({\textit{th}}_{2l}\), with \({\textit{th}}_{1l}\mathrm {<}{\textit{th}}_{2l}\) and \(0\le th_{1l}, th_{2l} \le 1\). Assuming that \(\left\lfloor x\right\rfloor \) is the largest integer not exceeding x, then the role of these thresholds, in link l, is described as follows (see Fig. 1):
-
1.
If \(0\le j_{l} <\left\lfloor th_{1l} C_{l} \right\rfloor \) then link l is in a support mode of operation, i.e., it accepts and serves not only new calls that initially arrive in link l but also new calls offloaded from link m \(\left( m=1,2,\, \, m\ne l\right) \).
-
2.
If \(\left\lfloor th_{1l} C_{l} \right\rfloor \le j_{l} <\left\lfloor th_{2l} C_{l} \right\rfloor \) then link l is in a normal mode of operation, i.e., it does not accept calls offloaded from link m. It only accepts calls that initially arrive in link l.
-
3.
If \(\left\lfloor th_{2l} C_{l} \right\rfloor \le j_{l} \) then link l operates in offloading mode of operation, i.e., a new call that initially arrives in link l will be offloaded to link m. If link m operates in support mode (i.e., \(0\le j_{m} <\left\lfloor th_{1m} C_{m} \right\rfloor \)) then the call will be accepted in link m. If link m does not operate in support mode and at the same time \(j_{l} \le C_{l} -1\), the call will be accepted in link l. Otherwise the call will be blocked and lost without further affecting the system.
Based on the above, the call admission of a new call that initially arrives in link l (l = 1, 2) is summarized in the following steps:
(1) If \(\left( 0\le j_{l} <\left\lfloor th_{2l} C_{l} \right\rfloor \right) \) then the call is accepted by the link l and remains for a generally distributed service-time with mean \(\mu ^{-1}\).
(2) If \(\left\lfloor th_{2l} C_{l} \right\rfloor \le j_{l}\) then: (2a) if \(0\le j_{m} <\left\lfloor th_{1m} C_{m} \right\rfloor \) the call is offloaded to link m and remains for a generally distributed service-time with mean \(\mu ^{-1}\), (2b) if \(\left\lfloor th_{1m} C_{m} \right\rfloor \le j_{m} \), then link m is in a normal mode of operation and does not support offloaded calls from link l. In that case, the call will be handled by link l. In addition, if \(j_{l} \le C_{l} -1\), then the call is accepted in link l and remains for a generally distributed service-time with mean \(\mu ^{-1}\). Otherwise, call blocking occurs.
Due to the support and offloading modes of operation of the two links, the corresponding 2-D Markov chain of the system is not reversible and therefore LB between adjacent states is destroyed. Thus, the steady state distribution, \(P({\varvec{j}})=P({ j}_{1} ,j_{2} )\), of this system cannot be described by a PFS. To determine the values of \(P({ j}_{1} ,j_{2} )\) (and consequently CBP) there exist two different methods.
The first method provides accurate results compared to simulation. The drawback of the method is that it requires the knowledge of the system’s state space and the solution of the set of linear GB equations for each state \({\varvec{j}}=({ j}_{1} ,j_{2} )\) expressed as rate into state j = rate out of state j:
where:
Before we proceed with the CBP determination, via the first method, we emphasize that this method can be quite complex even for a system with links of small capacity.
Having obtained \(P(j_{1} ,j_{2})\), we can calculate the CBP in each link, \(P_{b_{1} }^{'}\) and \(P_{b_{2} }^{'}\) via (3) and (4), respectively [27]:
The rationale behind (3) is that a call can be blocked in the first link if there is no available bandwidth in that link (i.e., if \(\textit{j}_{1}\) = \(\textit{C}_{1}\)) and the second link is not in support mode (i.e, \(\left\lfloor th_{12} C_{2} \right\rfloor \le j_{2} \)). A similar description applies to (4).
In addition, we can calculate the total CBP in the two-link system via the following weighted summation [27]:
The second method leads to approximate CBP results (compared to the corresponding simulation results) by assuming that both links operate independently from one another. This means that each link l operates as an independent Erlang loss system of capacity \(\textit{C}_{l}\) (l = 1, 2). Such an assumption simplifies (at the cost of accuracy) the necessary calculations for the CBP determination.
The CBP in the first and the second link can be approximated by (6) and (7), respectively:
where \(P_{l} \left( C_{l} \right) \) refers to the CBP in link l (l = 1, 2) which can be determined either via the closed form of the classical Erlang B formula (8a), or via its recurrent form, which is appealing for large values of offered traffic-load \(a_{l}\) and capacity \(C_l\) (8b):
As far as the values of \(P_{l} \left( j_{l} \ge \left\lfloor th_{1l} C_{l} \right\rfloor \right) \) are concerned, they are given by:
where \(P_{l} (j_{l} )\) is determined via the truncated Poisson distribution:
Finally, the total blocking probability can be determined via (5), where \(P_{b_{1} }^{'}\) and \(P_{b_{2} }^{'} \) are replaced by \(P_{b_{1} } \) and \(P_{b_{2} }\) (calculated via (6) and (7)), respectively.
An alternative recursive way for the determination of \(P_{l} (j_{l} )\), \({j}_{l}\) =1,...,\(\textit{C}_{l}\), is based on the link independence assumption. In the Erlang loss model, used to describe each link l, there exist LB between the adjacent states \(\textit{j}_{l\ }\)– 1 and \(\textit{j}_{l}\) which has the form [46]:
Based on (11), we can determine the unnormalized values of \(P_{l}^{'} (j_{l} )\)’s considering an initial value of \(\textit{P}_{l}^{'} (0)\) = 1. The normalized values of \(\textit{P}_{l}^{'} (j_{l} )\) ’s are given by:
Based on (12), we can calculate \(P_{b_{1}}\), \(P_{b_{2}}\) and the total CBP, via (6), (7) and (5), respectively.
3 The Proposed Model
In the proposed model, we consider again the loss system of the two links. Each link accommodates Poisson arriving calls of K different service-classes. Calls of service-class k (k = 1,..., K) require \(\textit{b}_{k}\) b.u. in order to be connected in a link. Let and be the arrival rates in the first and second link of service-class k calls, respectively. We also denote by j\(_{1}\) and j\(_{2}\) the occupied b.u. in the first and second link, respectively. Similar to Sect. 2, each link l (l = 1, 2) has a support threshold \({\textit{th}}_{1l}\) and an offloading threshold \({\textit{th}}_{2l}\), with \({\textit{th}}_{1l}\) \(\mathrm {<}\) \({\textit{th}}_{2l}\) and \(0\le th_{1l} ,th_{2l} \le 1\). To incorporate the notion of restricted accessibility in the proposed model, we assume that each state \(\textit{j}_{l}\) of link l, except from the initial state where link l is empty (i.e., when \(\textit{j}_{l\ }\) = 0) is associated with a blocking probability, \(pb_{l,k} \left( j_{l} \right) \) which can be different for each service-class k. Clearly, when there is no available b.u. for calls of service-class k in link l (i.e., when \(j_{l} \ge C_{l} -b_{k} +1\)), then \(pb_{l,k} \left( j_{l} \right) \,=\,1\). On the same hand, when the system is empty, then \(pb_{l,k} \left( 0\right) =0\).
The call admission of a new service-class k call that initially arrives in link l (l = 1, 2) is summarized in the following steps:
(1) If \(\left( 0\le j_{l} <\left\lfloor th_{2l} C_{l} \right\rfloor \right) \) then the call will be handled by link l. In addition, if \(j_{l} +b_{k} \le C_{l} \), then the call is accepted in link l with probability \(1-pb_{l,k} \left( j_{l} \right) \) and remains for a generally distributed service-time with mean \(\mu _{k}^{-1}\).
(2) If \(\left\lfloor th_{2l} C_{l} \right\rfloor \le j_{l}\) then:
(2a) if \(\left( 0\le j_{m} <\left\lfloor th_{1m} C_{m} \right\rfloor \right) \) the call is offloaded to link m and assuming that \(j_{m} +b_{k} \le C_{m} \), the call is accepted link m with probability \(1-pb_{{ m},k} \left( j_{m} \right) \) and remains in link m for a generally distributed service-time with mean \(\mu _{k}^{-1} \),
(2b) if \(\left\lfloor th_{1m} C_{m} \right\rfloor \le j_{m} \), then link m is in a normal mode of operation and does not support offloaded calls from link l. In that case, the call will be handled by link l. If \(j_{l} +b_{k} \le C_{l} \), then the call is accepted in link l with probability \(1-pb_{l,k} \left( j_{l} \right) \) and remains for a generally distributed service-time with mean \(\mu _{k}^{-1}\). Otherwise, the call is blocked and lost.
To determine in an approximate but efficient way the CBP of service-class k calls we assume that the two links operate independently from one another. In that case, each independent link behaves as an EMLM system under restricted accessibility [43], and therefore the CBP of service-class k calls in the first and the second link can be approximated by (13) and (14), respectively:
where \(P_{res,lk} \left( C_{lk} \right) \) refers to the CBP of service-class k calls in link l (l = 1, 2) and \(P_{res,{ l}} \left( j_{l} \ge \left\lfloor th_{1l} C_{l} \right\rfloor \right) \) refers to the probability that link l is not in support mode.
The values of \(P_{res,lk} \left( C_{l} \right) \) in (13) and (14) are determined by the formula:
where \(q(j_{l} )\) refers to the unnormalized values of the link occupancy distribution of link l (l = 1,2) while \(G_{l} =\sum _{j_{l} =0}^{C_{l} }q(j_{l} ) \) is the normalization constant.
In (15), the values of \(q(j_{l})\) can be recursively determined via the following formula:
where: \(a_{lk} =\lambda _{lk} /\mu _{k} \) is the total offered traffic-load of service-class k calls in link l.
As far as the values of \(P_{res,l} \left( j_{l} \ge th_{1l} C_{l} \right) \), in (13) and (14), are concerned they can be determined by:
where \({ q}(j_{l} )\) is determined via (16).
Finally, we propose the following formula for the total blocking probability of service-class k calls in the system of the two links:
4 Numerical Examples – Evaluation
In this section, we present an application example and provide analytical and simulation CBP results of the proposed model. Simulation results are mean values of 7 runs and are based on the Simscript III simulation language [47]. In each simulation run a total of ten million calls is generated. To account for a warm-up period, the first 5% of these calls are not taken into account in the CBP results. Furthermore, to increase the readability of figures we do not present reliability ranges. The latter are less than two order of magnitude.
As an application example, consider a two-link system of capacities \(\textit{C}_{1}\) = 24 b.u. and \(\textit{C}_{2}\) = 20 b.u. that accommodates K = 2 service-classes whose calls require \(\textit{b}_{1}\) = 1 and \(\textit{b}_{2}\) = 2 b.u., respectively. For the first link, let: = 9 calls/min and = 1 calls/min. Similarly, for the second link let = 7 calls/min and = 1 calls/min. Also let \(\mu _{1}^{-1} =\mu _{2}^{-1} =1.0\) min. As far as the values of the thresholds are concerned, we assume that the offloading thresholds are equal to \(\textit{th}_{21}\) = \(\textit{th}_{22}\) = 0.7 and consider two different sets of support thresholds: (1) \(\textit{th}_{11}\) = \(\textit{th}_{12}\) = 0.05 and (2) \(\textit{th}_{11}\) = \(\textit{th}_{12}\) = 0.25. Regarding the restricted accessibility blocking probability factors for each link, we consider two sets: (1) and (2) where l = 1, 2.
In the x-axis of Figs. 2, 3, 4 and 5, and increase in steps of 1.0 and 0.5, respectively. So, point 1 is: (\(\lambda _{11} =9.0,\, \lambda _{12} =1.0,\, \, \lambda _{21} =7.0,\, \lambda _{22} =1.0\)) while point 7 is: (\(\lambda _{11} =15.0,\, \lambda _{12} =1.0,\, \, \lambda _{21} =10.0,\, \lambda _{22} =1.0\)).
In Figs. 2 and 3, we present the CBP for the first service-class calls in each link, respectively. In Figs. 4 and 5, we present the corresponding CBP results for calls of the second service-class. Figures 2, 3, 4 and 5 show that the analytical CBP results: (a) are close to the simulation results especially when the values of the support thresholds \(\textit{th}_{11}\) and \(\textit{th}_{12}\) are kept within a reasonable level (e.g., 0.05 to 0.25). Depending on the two-link system, higher values of \(\textit{th}_{11}\) and \(\textit{th}_{12}\) may increase the error between analytical and simulation CBP results. This behavior can also be observed in the model of [27] and is anticipated due to the fact that (6), (7) and consequently (13), (14) imply that both links work independently from one another. (b) The choice of \(pb_{l,k} \left( j_{l} \right) \) greatly affects CBP. The higher values of set 1 result in much higher CBP compared to the lower values of set 2 .
5 Conclusion
We propose a new multirate loss model for the call-level analysis of a two-link system with restricted accessibility that accommodates Poisson arriving calls of different service-class with different bandwidth-per-call requirements. In the two-link system, each link may support and provide service to calls offloaded from the other link. The proposed model does not have a PFS for the steady state probabilities due to the offloading mechanism and the existence of restricted accessibility. However, we show that an approximate method does exist for the determination of CBP that achieves satisfactory accuracy compared to simulation. As a future work, we intend to use the proposed model as the springboard for the analysis of interference between the two links.
References
Stasiak, M., Glabowski, M., Wisniewski, A., Zwierzykowski, P.: Modeling and Dimensioning of Mobile Networks. Wiley, West Sussex, UK (2011)
Iversen, V.: Teletraffic Engineering and Network Planning. DTU Photonic, Denmark (2015)
Iversen, V.: Modelling restricted accessibility for wireless multi-service systems. LNCS, vol. 3883, pp. 93–102 (2006)
Stasiak, M., Glabowski, M.: A simple approximation of the link model with reservation by a one-dimensional Markov chain. Perf. Eval. 41(2–3), 195–208 (2000)
Moscholios, I., Logothetis, M., Kokkinakis, G.: Connection dependent threshold model: a generalization of the Erlang multiple rate loss model. Perform. Eval. 48(1–4), 177–200 (2002)
Glabowski, M., Stasiak, M.: Point-to-point blocking probability in switching networks with reservation. Ann. Telecommun. 57(7–8), 798–831 (2002)
Rácz, S., Gerő, B., Fodor, G.: Flow level performance analysis of a multi-service system supporting elastic and adaptive services. Perform. Eval. 49(1–4), 451–469 (2002)
Glabowski, M., Stasiak, M.: Multi-rate model of the limited availability group with finite source population. In: Proceedings of the 10th Asia-Pacific Conference on Communication, Beijing, China (2004)
Moscholios, I., Logothetis, M.: Engset multirate state-dependent loss models with QoS guarantee. Int. J. Commun. Syst. 19(1), 67–93 (2006)
Vassilakis, V., Moscholios, I., Logothetis, M.: Call-level performance modelling of elastic and adaptive service-classes with finite population. IEICE Trans. Commun. E91-B(1), 151–163 (2008)
Huang, Q., Ko, K.-T., Iversen, V.: Approximation of loss calculation for hierarchical networks with multiservice overflows. IEEE Trans. Commun. 56(3), 466–473 (2008)
Stasiak, M., Sobieraj, M., Weissenberg, J., Zwierzykowski, P.: Analytical model of the single threshold mechanism with hysteresis for multi-service networks. IEICE Trans. Commun. E95-B(1), 120–132 (2012)
Moscholios, I., Vardakas, J., Logothetis, M., Koukias, M.: A quasi-random multirate loss model supporting elastic and adaptive traffic under the bandwidth reservation policy. Int. J. Adv. Netw. Serv. 6(3&4), 163–174 (2013)
Yan, S., Razo, M., Tacca, M., Fumagalli, A.: A blocking probability estimator for the multi-application and multi-resource constraint problem. In: Proceedings of the ICNC, Honolulu, Hawaii (2014)
Moscholios, I., Logothetis, M., Vardakas, J., Boucouvalas, A.: Performance metrics of a multirate resource sharing teletraffic model with finite sources under both the threshold and bandwidth reservation policies. IET Netw. 4(3), 195–208 (2015)
Huang, Y., Rosberg, Z., Ko, K., Zukerman, M.: Blocking probability approximations and bounds for best-effort calls in an integrated service system. IEEE Trans. Commun. 63(12), 5014–5026 (2015)
Moscholios, I., Logothetis, M., Vardakas, J., Boucouvalas, A.: Congestion probabilities of elastic and adaptive calls in Erlang-Engset multirate loss models under the threshold and bandwidth reservation policies. Comput. Netw. 92(1), 1–23 (2015)
Glabowski, M., Sobieraj, M.: Analytical modelling of multiservice switching networks with multiservice sources and resource management mechanisms. Telecommun. Syst. 66(3), 559–578 (2017)
Widjaja, I., Roche, H.: Sizing X2 bandwidth for inter-connected eNBs. In: Proceedings of the IEEE VTC Fall, Anchorage, Alaska, USA (2009)
Stasiak, M., Zwierzykowski, P., Parniewicz, D.: Modelling of the WCDMA interface in the UMTS network with soft handoff mechanism. In: Proceedings of the IEEE GLOBECOM, Hawaii (2009)
Renard, B., Elayoubi, S., Simonian, A.: A dimensioning method for the LTE X2 interface. In: Proceedings of the IEEE WCNC, Shanghai, China (2012)
Stasiak, M., Parniewicz, D., Zwierzykowski, P.: Traffic engineering for multicast connections in multiservice cellular network. IEEE Trans. Ind. Inform. 9(1), 262–270 (2013)
Moscholios, I., Kallos, G., Katsiva, M., Vassilakis, V., Logothetis, M.: Call blocking probabilities in a W-CDMA cell with interference cancellation and bandwidth reservation. In: Proceedings of the IEICE ICTF, Poznan, Poland (2014)
Moscholios, I., Kallos, G., Vassilakis, V., Logothetis, M.: Congestion probabilities in CDMA-based networks supporting batched Poisson input traffic. Wirel. Pers. Commun. 79(2), 1163–1186 (2014)
Khedr, M., Makki Hassan, R.: Opportunistic call admission control for wireless broadband cognitive networks. Wirel. Netw. 20(1), 105–114 (2014)
Machado de Medeiros, A., Yacoub, M.: BlockOut: blocking and outage in a single performance measure. IEEE Trans. Veh. Tech. 63(7), 3451–3456 (2014)
Burger, V., Seufert, M., Hossfeld, T., Tran-Gia, P.: Performance evaluation of backhaul bandwidth aggregation using a partial sharing scheme. Phys. Commun. 19, 135–144 (2016)
Moscholios, I., Vassilakis, V., Logothetis, M., Boucouvalas, A.: A probabilistic threshold-based bandwidth sharing policy for wireless multirate loss networks. IEEE Wirel. Commun. Lett. 5(3), 304–307 (2016)
Vassilakis, V., Moscholios, I., Logothetis, M.: Uplink blocking probabilities in priority-based cellular CDMA networks with finite source population. IEICE Trans. Commun. E99-B(6), 1302–1309 (2016)
Vassilakis, V., Moscholios, I., Logothetis, M.: Quality of service differentiation of elastic and adaptive services in CDMA networks: a mathematical modelling approach. Wirel. Netw. 24(4), 1279–1295 (2018)
Wang, Z., Mathiopoulos, P., Schober, R.: Performance analysis and improvement methods for channel resource management strategies of LEO-MSS with Multiparty Traffic. IEEE Trans. Veh. Tech. 57(6), 3832–3842 (2008)
Yiltas, D., Zaim, A.: Evaluation of call blocking probabilities in LEO satellite networks. Int. J. Satell. Commun. 27(2), 103–115 (2009)
Wang, Z., Mathiopoulos, P., Schober, R.: Channel partitioning policies for multi-class traffic in LEO-MSS. IEEE Trans. Aerosp. Electron. Syst. 45(4), 1320–1334 (2009)
Vardakas, J., Moscholios, I., Logothetis, M., Stylianakis, V.: An analytical approach for dynamic wavelength allocation in WDM-TDMA PONs servicing ON-OFF traffic. IEEE/OSA J. Opt. Commun. Netw. 3(4), 347–358 (2011)
Deng, Y., Prucnal, P.: Performance analysis of heterogeneous optical CDMA networks with bursty traffic and variable power control. IEEE/OSA J. Opt. Commun. Netw. 3(6), 487–492 (2011)
Vardakas, J., Moscholios, I., Logothetis, M., Stylianakis, V.: On code reservation in multi-rate OCDMA passive optical networks. In: Proceedings of the CSNDSP, Poznan, Poland (2012)
Vardakas, J., Moscholios, I., Logothetis, M., Stylianakis, V.: Performance analysis of OCDMA PONs supporting multi-rate bursty traffic. IEEE Trans. Commun. 61(8), 3374–3384 (2013)
Casares-Giner, V.: Some teletraffic issues in optical burst switching with burst segmentation. Electron. Lett. 52(11), 941–943 (2016)
Guan, Y., Jiang, H., Gao, M., Bose, S., Shen, G.: Migrating elastic optical networks from standard single-mode fibers to ultra-low loss fibers: strategies and benefits. In: Proceedings of the Optical Fiber Communication Conference, Los Angeles, USA (2017)
Mamatas, L., Psaras, I., Pavlou, G.: Incentives and algorithms for broadband access sharing. In: Proceedings of the ACM SIGCOMM Workshop on Home Networks, New Delhi, India (2010)
Psaras, I., Mamatas, L.: On demand connectivity sharing: queuing management and load balancing for user-provided networks’. Comput. Netw. 55(2), 399–414 (2011)
Moscholios, I., Vassilakis, V., Logothetis, M., Boucouvalas, A.: State-dependent bandwidth sharing policies for wireless multirate loss networks. IEEE Trans. Wirel. Commun. 16(8), 5481–5497 (2017)
Iversen, V., Stepanov, S., Kostrov, A.: Dimensioning of multiservice links taking account of soft blocking. LNCS, vol. 4003, pp. 3–10. Springer (2006)
Kaufman, J.: Blocking in a shared resource environment. IEEE Trans. Commun. 29(10), 1474–1481 (1981)
Roberts, J.: A service system with heterogeneous user requirements. In: Pujolle, G. (ed.) Performance of Data Communication Systems and Their Applications, North Holland, Amsterdam, pp. 423–431 (1981)
Pantelis, S., Moscholios, I., Papadopoulos, S.: Call-level evaluation of a two-link single rate loss model for Poisson traffic. In: Proceedings of the IEICE ICTF, Poznan, Poland, pp. 4–6 (2017)
Simscript III. http://www.simscript.com. Accessed July 2019
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Keramidi, I.P., Moscholios, I.D., Sarigiannidis, P.G., Logothetis, M.D. (2020). Call-Level Analysis of a Two-Link Multirate Loss Model with Restricted Accessibility. In: Choraś, M., Choraś, R. (eds) Image Processing and Communications. IP&C 2019. Advances in Intelligent Systems and Computing, vol 1062. Springer, Cham. https://doi.org/10.1007/978-3-030-31254-1_29
Download citation
DOI: https://doi.org/10.1007/978-3-030-31254-1_29
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-31253-4
Online ISBN: 978-3-030-31254-1
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)