GSM的切换业务的在GPRSGSM网络性能评估毕业论文外文翻译.docx

《GSM的切换业务的在GPRSGSM网络性能评估毕业论文外文翻译.docx》由会员分享，可在线阅读，更多相关《GSM的切换业务的在GPRSGSM网络性能评估毕业论文外文翻译.docx（31页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、外文资料Performance evaluation of GSM handover traffic in a GPRS/GSM networkJuan Ventura Agustina, Peng Zhang, Raimo KantolaNetworking LaboratoryHelsinki University of TechnologyOtakaari 5A, Espoo, FIN 02015, FinlandEmail: juavenag, pgzhang, kantolatct.hut.fiAbstract The introduction of GPRS services in

2、to GSM networks creates new challenges to network planning engineers. One critical challenge comes from the requirement for providing a certain quality of service for GPRS traffic without significantly degrading the performance of existing GSM services. In a GSM/GPRS integrated network, it becomes n

3、ecessary to reserve exclusive channels for GPRS in order to provide base-line QoS for GPRS users. On the other hand, the exclusive reservation obviously reduces the capacity of GSM traffic so that has significant impact on the performance of GSM traffic (especially GSM handover traffic). In this pap

4、er, we primarily evaluate the performance degradation of GSM handover traffic due to the introduction of GPRS in a GSM/GPRS network when various priority schemes for handover traffic over new call traffic are applied. A simplified case study of a GPRS/GSM network is simulated by using an event-drive

5、n simulator. The effect of an increasing GPRS penetration factor on the performance of existing GSM services is also studied. Our key results show that the performance of GSM handover traffic can be significantly degraded by the capacity reduction resulting from the introduction of GPRS but can be a

6、mended by using appropriate priority schemes.Keywords: GPRS, GSM, channel allocation, quality of service, handover handling1. IntroductionMany studies have predicted that within next few years there will be an extensive demand for mobile data services, specially wireless Internet 1. In order to addr

7、ess the inefficiencies of current circuit-switched mobile networks, such as GSM, for carrying bursty data traffic (typical Internet applications show such traffic behavior), packet switching techniques have emerged in mobile networks. GPRS is a new bearer service that greatly improves and simplifies

8、 wireless access to packet data networks, e.g., to the Internet. The basic GPRS concept is to utilize rest traffic channels unused by GSM traffic. In general, GSM traffic has higher priority than GPRS traffic when allocating channels, which means that an ongoing GPRS channel has to be terminated for

9、 a pending GSM traffic. On the other hand, in a GSM and GPRS integrated network high GSM traffic load may prevent GPRS traffic from achieving an acceptable quality guarantee if no channel is exclusively dedicated to GPRS. Therefore, given a number of channels for both GSM and GPRS, it is reasonable

10、that a fraction of the channels are exclusively assigned to GPRS and the rest are shared between GSM and GPRS while GSM traffic has higher priority over GPRS traffic. The combination of shared and dedicated traffic channels is so called partial sharing (PS) technique. When the exclusive reservation

11、of channels for GPRS takes place without allocating new spectrum, the GPRS PS implementation obviously reduces the capacity of existing GSM services. This reduction is especially critical in the case of GSM handover traffic because terminating a call in progress is clearly less desirable than blocki

12、ng a new call attempt. Therefore, methods for improving the handover performance become necessary when the capacity reduction is considerable. The performance of GPRS service has been extensively investigated in past years, but very few results have appeared regarding the effect of GPRS PS implement

13、ation on the existing GSM services 1-4. In 1, the performance loss of GSM services due to the introduction of GPRS is studied as an additional experiment, and no solution is given for counteracting the reduction of GSM capacity. In 4, the impact of GPRS on the quality of existing GSM services is ana

14、lyzed and a method for calculating the outage probability of a GPRS/GSM network is proposed, but the capacity reduction of GSM services is just mentioned.In this paper we introduce different handover priority-based channel allocation schemes in order to counteract the reduction of capacity of GSM se

15、rvices at the same time that prioritize handover traffic over new call attempts. We investigate the effectiveness of these handover prioritization schemes on improving the performance of handover traffic in a GPRS/GSM network. Particularly, we present the effect of an increasing GPRS penetration fac

16、tor on the performance of GSM traffic.The rest of the paper is organized as follows. Section 2 presents a number of handover prioritization schemes. In section 3, a simplified GPRS/GSM network is modeled by defining main features, assumptions and the models for the proposed schemes. Furthermore, the

17、 main performance parameters of interest are defined. In section 4, simulation results are shown and discussed. Finally, conclusions are drawn in section 5, as well as some future research guidelines. 2. Handover handling schemesWhen allocating a channel, a simple scheme employed by cellular technol

18、ogies handles both types of calls (new calls and handovers) without preference. This means that the probabilities of new call blocking and handover failure are the same. This scheme is referred to as the non-prioritized scheme (NPS). However, from the users point of view, the forced termination of a

19、n ongoing call is considered to be worse than blocking a new call attempt. Therefore, it becomes necessary to introduce methods for decreasing the probability of handover failure as well as new call blocking. Various handover prioritization schemes have been studied in the past 5-8. These schemes ca

20、n be sorted into four classes: Reserving a number of channels exclusively for handovers Queueing handover requests Sub-rating an existing call to accommodate a handover Combination of the above classesThese schemes are separately described in the following subsections.Note that these schemes take pl

21、ace on the basis of PS implementation. For example, given a fixed number of channels (Nch) in a GSM/GPRS cell, a fraction of channels are exclusively reserved for GPRS (Ngprs). The Ngprs is referred to GPRS penetration factor. Then the rest channels (Nshared) can be used for both GSM and GPRS, where

22、 these priorization schemes apply. 2.1. Reserved channel scheme (RCS)This scheme reserves a number of channels exclusively for handovers requests from Nshared. Then, the Nshared channels are divided into two different groups: a common channel group (Ncom) and a reserved channel group (Nho): the Ncom

23、 channels can be used by new calls as well as handovers, whereas the Nho channels can only be used by handovers. There are two types of reservation: Pre-reservation (RCS-pre)On the arrival of a handover request, a channel in Nho is allocated for the handover. If Nho is fully occupied by handover tra

24、ffic, the handover request contends with new call attempts for a channel in the Ncom pool. This ensures that certain minimal handover traffic will be admitted even under heavy load. Post-reservation (RCS-post)On the arrival of a new handover request, it contends with new call attempts for admission

25、into the Ncom pool. If Ncom is full, it will be allocated in the Nho pool. This post-reserved pool ensures that even under heavy loads extra priority is given to handovers.Reserving channels for handovers means fewer channels can be granted to new calls, so the blocking probability of new calls may

26、significantly increase. This disadvantage can be overcome by using a queueing scheme in the following subsection.2.2. Queueing priority scheme (QPS)On the arrival of a new handover request, if there is no free channel in the target cell for the request, the handover request is queued and the mobile

27、station (MS) continues to use the old channel in the current cell until a free channel becomes available in the target cell. The queueing can be performed at Base Transceiver Station (BTS)/Base Station Controller (BSC) or Mobile Switch Center (MSC) in GSM system. One new call in the target cell is s

28、erved only when a channel is available and no handover request exists in the queue. If any channel is released while handover requests are queued, the released channel is assigned to a handover in the queue. The “next” handover to be served is selected based on queueing policies, which also influenc

29、e the performance of the scheme: FIFO priority queueing(F-QPS)With the FIFO queueing discipline, if a handover request finds all channels occupied in the target cell, the request is queued according to a FIFO discipline, i.e., the last handover request joins the end of the queue and the first to be

30、served is the first in the queue (the earliest one to arrive in the queue). Measurement-based priority queueing (M-QPS)During the time interval the MS spends in the handover area, its communication with the current BTS degrades at a rate depending on various factors, such as its velocity and directi

31、on. This degradation rate is easily monitored by the means of radio channel measurements, usually taken by the MS and submitted to the network (Moblie Assisted Handoff - MAHO procedures of the GSM system). Then, the handover area can be viewed as a region marked by different ranges of values of the

32、power ratio. Then, the highest priority belongs to the MS whose power level is the closest to a receiver threshold. On the other hand, the MS that has just issued a handover request has the least priority. The power levels are monitored continuously, and the priority of an MS dynamically changes dep

33、ending on its power level. Obviously, the last comer joins the end of the queue, but the queue is dynamically reordered as new measurement results are submitted. When a channel is released, it is granted to the MS with the highest priority. 2.3. Sub-rating scheme (SRS)This scheme creates a new chann

34、el by sub-rating an existing call for a handover request when all channels are occupied in the target BTS. Sub-rating means that an occupied full-rate channel is temporarily divided into two channels at half the original rate: one to serve the existing call and the other to serve the handover reques

35、t. A protocol required to sub-rate a traffic channel is described in 8. 2.4. Hybrid schemes (HS)It is possible to combine the above handover prioritization schemes. For instance, Nho channels can be exclusively reserved for handovers while allowing queueing of handover requests. This results in shor

36、tening the queuing length of handover requests. Another possibility is to use the RCS-pre and the RCS-post schemes at the same time. A random number between 0 and 1 is generated in the BTS/BSC/MSC: if this number lies between 0 and 0.5, the RCS-pre is performed for handover calls; otherwise, the RCS

37、-post is performed. 3. Model description of the studied systemThe performance of a cellular network can be investigated by using either simulation or analytical study (or their combination). Simulation models are preferred when studying the behavior of a specific cellular system covering a given are

38、a. In this paper we carry out the performance study on the basis of computer simulation. An event-driven simulator has been implemented using a simulation library developed in C+ This simulator is available via http:/www.tct.hut.fi/pgzhang/GHS. The simulation library is based on a C+ class library c

39、alled CNCL developed in the Aachen University of Technology, which is available free of charge via anonymous ftpcomnets.rwth-achen.de . 3.1. System parametersThe simulation focuses on one cell in an overlaid GPRS/GSM cellular network. Its behavior is isolated from those of other cells. The simulatio

40、n concentrates on the uplink procedure where resource contention and resource reservation take place.The penetration factor of GPRS service is expected to be higher in urban/suburban areas, so the type of cell should be common in these areas. Therefore, we choose to study a microcell, that is, a cel

41、l with a relatively small size. In a microcell scenario the number of handovers per active call is higher than that in a macrocell 9. Common microcell radius is between 200m and 1km 10. Rmicro= 800m is assumed for the simulation. Furthermore, it is considered that users in cell border are always in

42、line-of-sight (LOS) in order to avoid street-corner effect 10. For simplicity, we employ a fixed frequency assignment strategy. Moreover, we consider that channels are allocated in a fixed manner, that is, no dynamic changing based on traffic load is implemented. In practice, a dynamic channel alloc

43、ation can be implemented which allows flexible adaptation to different traffic conditions. The number of TRXs in the cell is set to 4. Then, there are 32 physical channels (i.e., Nt = 32) available in the cell (every carrier is divided into 8 timeslots). We assume that 3 channels are reserved for ne

44、twork signaling (i.e., Nsig = 3). Thus, only 29 traffic channels are available for carrying users information (i.e., Nch = 29). The GPRS PS implementation exclusively dedicates Ngprs channels to GPRS, and the remaining channels Nshared = Nch Ngprs are shared by GSM and GPRS services. In the Nshared

45、channel pool, GSM traffic has higher priority with pre-emption over GPRS traffic so that the capacity of GSM traffic is the Nshared channels. 3.2. Traffic and mobility modelsThe arrival of GSM calls to the cell is the superposition of two processes corresponding to newly initiated calls within the c

46、ell (new calls) and calls handed over from the neighboring cells (handovers). It is assumed that new calls and handover requests are both generated according to a Poisson process with an arrival rate ln and lh respectively. Calls that are initiated in the cell can be divided into those that complete

47、 inside the cell and those that are handed over to other cells. In the same way, handovers can also be divided into calls that terminate in the cell and those that continue to other cells. Therefore, a channel could be occupied by the arrival of a new call or a handover, and it could be released eit

48、her by completion of the call or a handover to another cell. The time spent by a user on a particular channel in a given cell is defined as the channel holding (or occupancy) time.Although the channel holding time is taken equivalent to the call duration time in a fixed telephone network, it is ofte

49、n a fraction of the total call duration in a cellular mobile network. In general, the channel holding time is a random variable that is a function of system parameters such as cell size, user location, user mobility and call duration time 11. In 11 it is found that the distribution function of the channel holding time in a single cell follows a negative exponential