Service Disciplines Performance for Best-Effort Policies in Packet-Switching Wireless Cellular Networks

Transcription

1 Service Disciplines Performance for Best-Effort Policies in Packet-Switching Wireless Cellular Networks Wessam Ajib and Philippe Godlewski cole Nationale Sup rieure des T 1 communications, Computer Sciences and Networks Department 46, Rue Barrault, Paris Cedex 13, France Tel. (+33) Fax. (+33) address: (wajib&godlewski)@infies.enst.fr ABSTRACT - In this paper, we investigate the performance of service disciplines for packet-switched wireless networks in the context of General Packet Radio Service (GPRS). The GPRS is a new additional GSM service, which provide packet switched data services over GSM network and a packet access to data networks. Using either best effort service or guaranteed performance service, the service discipline affects considerably the MAC layer performance. To provide best possible performance for best-effort traffic in packet switching wireless networks, several service disciplines are proposed. We describe their mechanisms, and differences. These policies performance is studied in terms of throughput, delay and loss rate by simulating various traffic applications in GPRS context. KEY WORDS: multiple access technology, wireless data communications, GPRS, GSM. I. Introduction The system GPRS is a new additional data bearer service of GSM, which is currently being standardized by ETSI in order to satisfy the increasing demand for mobile data communications and recent developments of mobile data applications. The physical channels, available in a GSM cell, are dynamically shared between GPRS and other GSM services and the ones associated with GPRS are called Packet Data Channels (PDCHs). The radio interface of GPRS system is defined in [6] and an overview is given in [ The higher layer is called SubNetwork Dependent Convergence Protocol (SNDCP). It maps network level characteristics onto the ones of the underlying network. Under the SNDCP, the Logical Link Control (LLC) layer provides a highly reliable ciphered logical link between the tow entities. The lower layers of GPRS radio interface are Radio Link Control (RLC), Medium Access Control (MAC) (defined in [l]) and the physical layers. The RLC layer performs the segmentation and re-assembly of LLC- PDUs into RLC data blocks. Besides, it provides an acknowledgment mechanism based on selective repeat ARQ protocol [5]. In order to ameliorate this mechanism performance, an efficient modification is proposed in [3-71. The procedures defined at MAC layer enable multiple Mobile Stations (MS) to share a common transmission medium, which may consists of one or several physical channels i.e., PDCHs. The physical layer provides services for information transfer over physical channels between the MS and the network. In this paper, we centralise our studies on MAC layer procedures and particularly on the scheduling issue. The basic radio packet in GPRS is the RLC/MAC data block, called simply data block. It uses a sequence of four timeslots on a PDCH, called block period, to be transmitted [I]. A Temporary Block Flow (TBF) is a physical connection used to support the transfer of a number of data blocks. To each TBF, a Temporary Flow Identity (TFI) is assigned. By including the TFI in each data block, the multiplexing of blocks destined for (or originated from) different MSs on the same PDCH is performed. New types of packet data logical channels are defined and mapped dynamically onto a 52- multiframe [6]. They include Packet Broadcast Channels (PBCCH), Packet Common Control Channels (such as: Random Access (PRACH) and Access Grant (PAGCH)) and traffic channels. The material in this paper is organized as follows. The next section contains a rapid description of MAC procedures for uplink and downlink transmissions; after which, we focus on the procedures that define the resource sharing. The third section analyzes the quality of service in GPRS and introduces some traffic applications and models. This section considers the QoS and applications described in ETSI documents [2-91. The forth section summarizes the scheduling point and proposes different mechanisms. The proposed algorithms depend on the medium access modes defined in GPRS for uplink and downlink traffic. They can be used for different packet switching wireless networks. The simulation model and results are given and discussed in section V, showing the interest of our propositions. Finally, conclusions are drowning. 11. The MAC layer procedures The MAC layer is situated between RLC layer and physical layer. The functions of this layer are related to the management of shared transmission resources (i.e., PDCHs) and control block receptions. One or more PDCHs can be assigned to a mobile station and one or /00/$ IEEE VTC2000

2 more MSs can use one PDCH. Two MAC modes are defined: the packet idle mode where no TBF exists and the packet transfer mode where radio resource providing a TBF is allocated to MS. The MAC procedures, in both idle and transfer modes, concern cell re-selection, system information broadcasting and paging procedures. The network may initiate a downlink transfer by paging procedures if the MS is in standby MM state. Once the MS is in ready MM state, the network sends a packet downlink assignment, on PAGCH, including the list of assigned PDCHs and then it transmits data blocks operating with the RLC acknowledgment mechanism. The uplink TBF establishment, considering only one phase access, begins when the MS transmits an access request, on PRACH, and performs the access persistence control mechanism. The parameters of this mechanism (Max-Retrans, S, and 7 ) are defined in [l] and outlined in [4]. A new mechanism (called Access request Retransmission Announced Protocol or AQRAP) is proposed in [4] toward an amelioration of MAC performance. The network responds to the access request, on PAGCH, by (i) a packet uplink assignment defining the medium access mechanism, (ii) packet queuing notification or (iii) packet access reject. Upon packet assignment receipt, the MS transmits the data blocks acting with the RLC acknowledgment mechanism. Three medium access modes, for uplink transmission, are supported: dynamic, extended dynamic and fixed. The MS, in dynamic allocation, monitors the Uplink State Flag (USF) field, contained in downlink blocks, in order to recognise its assigned uplink block periods. The extended dynamic allocation is a simple extension of the dynamic one adapted to deliver large volume data packets. Within this mode, a USF value indicates the assigned block periods on several PDCHs. In the case of the fixed allocation, a certain amount of assigned block periods are fixed at the establishment of the TBF. Handling with this mode, a packet uplink assignment is sent to the MS, when needed, to update the amount of assigned resources. III. The QoS and applications in GPRS The services that have to be offered by GPRS system are: Point-To-Multipoint (PTM) services and Point-To- Point (PTP) services, for either connectionless network protocols such as Intemet Protocol (IP) and Connection Oriented Network Protocols (CONP) defined in IS [l 13 i.e., X.25. Within the GPRS Phase 1, only the PTP service is defined as a bearer service type and the QoS parameters (service precedence, reliability, delay and throughput) are defined as a user application GPRS profile [2]. The service precedence (priority) indicates the relative priority of maintaining the service under abnormal conditions. The reliability indicates the transmission characteristics requested by an application i.e., the probability of loss of, duplication of, missequencing of or corruption of data units. The delay defines the maximum values for the mean delay and the 95-percentil delay to be incurred by data transfer through GPRS network. The throughput indicates the maximum bit rate and mean bit rate requested by the user. Firstly, the GPRS system will support principally best effort services. Therefore, only best effort policies are taking into account in these studies. The three performance criteria are the average value of global throughput, delay requested to transmit a packet and packet loss rate where a packet is considered lost if it is not received correctly. A packet corresponds to a network data unit. The GPRS enables the cost effective and efficient use of network resources for packet data applications that exhibits one or more of the following characteristics [2]: (i) intermittent, non periodic data transmissions, (ii) frequent transmissions of small volumes of data and (iii) infrequent transmissions of larger volumes of data. The GPRS network may support a set of additional services, which include: 0 accessing information stored in data base centers on demand only (e.g., Intemet s World Wide Web WWW or minitel-like), messaging services (e.g., ), conversational services (e.g., File Transfer Protocol (FTP) service or Intemet s Telnet application) and tele-action services de (e.g., credit carte validation, lottery transactions or surveillance system ) Three traffic models are presented in [9] to be used for the evaluation of proposed GPRS solutions. The first model, called Mobitex, can present tele-action services (uplink: 30 f 15 bytes and downlink: 115 k 57 bytes). This model shows the frequent transmission of small volume of data. Therefore, it needs relatively a large amount of signaling and therefore a considerable amount of signaling resources. Thus diminishes the amount of resources that can transmit data. The second model (Railway) is approximated by a negative exponential distribution function with an average packet length of 170 bytes truncated at 1000 bytes. This model can imitate the minitel-like service. It denotes the medium sized packets with medium arrival rate. The third traffic model is called FUNET and can be approximated by Cauchy (0.8, 1) distribution truncated at 10 Kbytes. It denotes the infrequent transmission of large volume of TBF and can present the conversational services of Internet (e.g., FTP) or service. IV. The scheduling algorithms In order to simplify the studies, only uplink traffic in GPRS context is considered but the proposed scheduling algorithms can be used also for downlink traffic as well as for varied packet switching wireless networks. The MAC layer arbitrates the access to the shared medium between a multitude of MSs and the network. This layer has to manage scheduling algorithms, which define how radio resources are shared among the MSs. Two scheduling algorithms have to be defined. The first concerns the distribution of PDCHs between different MSs. It is executed at the connection establishment. The second algorithm defines the distribution of block periods belonging to the same PDCH between MSs to which this PDCH is assigned. It is executed within the transmission of data blocks. The first scheduling algorithm is not the subject of this paper; so, the /00/$ IEEE VTC2000

3 following simple algorithm is used. Considering that the network knows the number of MSs assigned to each PDCH and a PDCHI is supposed more loaded than PDCH2 if the MSs assigned to PDCHl are more numerous than the ones assigned to PDCH2. When a MS, whose multislot class is x, wants to establish a connection (a TBF), the network allocate the least loaded x PDCHs. The second scheduling issue is also a network dependent choice and the scheduling algorithm depends on the access medium allocation mode i.e., dynamic, extended dynamic or fixed. In [lo], the guaranteed performance service disciplines are studied. Basing on it, we would prefer a scheduling algorithm to be efficient, protective, flexible and simple. Efficient. An algorithm is more efficient than another one if it meets the same performance under a heavier load of traffic. 0 Protective. It is not largely affected by abnormal traffic conditions such as load network fluctuations. 0 Flexible. It can be adaptable to different traffic applications. Simple. An algorithm has to be conceptually simple to allow tractable analysis and mechanically simple to allow high speed implementations. The first proposed scheduling algorithm is called First Come First Served or FCFS. It avoids the interference of blocks originated from distinct MSs on the same PDCH. An arriving set of blocks will be transmitted together and this user monopolizes the PDCH until the end of set transmission. The other sets of blocks (new arriving originated from other users or retransmitted blocks) have to delay in a FIFO queue according to their arriving. When the current user releases the PDCH, the next set of blocks is served. If some blocks are negatively acknowledged, a new set of blocks is formed and then treated as a new arriving set of blocks. This algorithm is compatible with the fixed mode of medium access. It is simple to be conceived, analyzed and implemented but intuitively, it is not especially efficient or protective. It is more adapted to frequent transmissions of small TBF than infrequent transmissions of larger TBF. Within the FCFS algorithm, the retransmitted blocks grows considerably the transmission delay of a TBF because it is usually a small set of blocks which has to await the transmission of larger set of blocks originated from other users. The second algorithm proposed (called FCFS with priority) uses approximately the same scheduling policy FCFS giving a transmission priority to the set of blocks. This priority depends only on how many times the block has been sent. This policy is simple and more efficient than the first one. The third scheduling algorithm proposed is called "dynamic". The objective of this policy is to divide the bloc periods of a PDCH equitably between all the users of this PDCH. The resource assignment is updated at each new user arrival and at least at each bloc period. This mechanism operates like a set of FIFO queues (one queue per user) with a Round Robin server whose allocation cycle is one bloc period. This algorithm can be seen as a realistic implementation of the Fluid Fair Queuing (FFQ) scheduling mechanism [12]. It is more protective and flexible than the ones presented before (FCFS and FCFS with priority); just it is more complicated to be implemented. Moreover, it needs an important quantity of signaling. It can be used only with dynamic allocation mode in uplink trafic. The forth algorithm is called "FCFS with windows". It is characterized by an important parameter (bitmap length noted W). This mechanism reallocates the resources (bloc periods of a PDCH) equitably among the users of this PDCH each "allocation cycle" (G W/x bloc periods, where x is the MS multislot class). In the beginning of each allocation cycle, the next bloc periods (within an allocation cycle) are reserved equitably for the different users. The new arriving users within an allocation cycle await the beginning of next allocation cycle to depart the transmission. The parameter W can be adapted to the application characteristics and to traffic conditions (signaling quantity and channel quality). This mechanism can be used with the three allocation modes. It is adapted to infrequent transmission of large TBF and intermittent transmissions. It seems protective, efficient and simple to be implemented. V. Simulation results Length of a LLC frame: 1520 bytes LLC acknowledgment mode: Ack Max number of retransmission of a LLC frame: 3 RLC acknowledgment mode: Ack Max number of retransmission of a data block: 7 Coding channel scheme used: CS-2 Max active MSs number in the cell: 32 Max MSs assigned to a PDCH: 16 0 Block error rate (BLER) = 10% Multislot class of MSs: 2 0 Total PDCH number: 7 S = 16, T = 8, and Max-retrans = 7 The simulator input is the load in the cell, which increases by increasing the MSs number in the cell. The performance criteria used in these simulations are: (i) the average global throughput in the cell, which is the quantity of data received correctly, (ii) the delay requested to transmit a TBF and (iii) the packet loss rate where a packet is supposed lost if it is not received correctly. The throughput and delay considers only the TBFs correctly received. Figures I, 2 and 3 use the Mobitex traffic model where a TBF length is at most two data blocks. These simulations consider only the three first algorithms since they are not adapted to the forth one. The following values are used in Mobitex simulation. 0 MSs number: ad Average IoadIMS: 250 bids. 0 There are 4 traffic PDCH, 3 PDCH carrying PRACH and 4 PDCH carrying PAGCH. We notice that the FCFS with priority algorithm gives the best performance and dynamic algorithm gives the worst performance. In the case of Mobitex model, the scheduling does not affect notably the performance /00/$ IEEE VTC2000

4 . I --. FCFS,CfSlnms...;... I., /-q Figure 1. The throughput vs MSs number (mobitex) 00,,,,..... I.....,-.- cffs,,,, eo-. :-Mr.'..._I_.: ( do,,usam5&d, Figure 4. The throughput vs MSs number (Railway) 25 I I I I I I I, I 4 ; ;,....?or. w...:...:...;./.:...; I... ; ;I.: I ;..., I ;/ Figure 3. The loss rate vs MSs number (mobitex) Figures 4, 5 and 6 employ the Railway traffic model. Average load per MS: 1000 bids, MSs number: There are 6 traffic PDCH, 1 PDCH carrying PRACH and 1 PDCH carrying PAGCH. When the MSs number is less than 50 (i.e., global input load<50kbit/s), the scheduling algorithms, classified according to their decreasing performance, are FCFS with priority, FCFS with windows, FCFS and dynamic. When the MSs number is higher than 50, the classification of scheduling algorithms becomes FCFS with priority, FCFS, dynamic, and FCFS with windows. The FCFS with priority algorithm gives the best possible performance when the traffic application contains only small volume TBFs. Figure 5. The delay vs MSs number (railway).o ,, Figure 6. The loss percent vs the MSs number (Railway) The Funet traffic model is used in figures 7, 8 and 9. Average load per MS: 1000 bids, MSs number: There are 6 traffic PDCH, 1 PDCH carrying PRACH and 1 PDCH carrying PAGCH /00/$ IEEE VTC2000

5 The analysis done for this simulation is similar to the one done in Railway traffic case. When global input load is less than 50kbit/s, the scheduling algorithms, classed according to their decreasing performance, are dynamic, FCFS with windows, FCFS with priority, and FCFS. When the MSs number is higher than 50, the scheduling algorithms order becomes FCFS with priority, FCFS, dynamic, and FCFS with windows. The FCFS with priority furnishes the best performance possible for frequent transmission of small volume TBFs. It is a simple algorithm but not flexible. The performance dynamic algorithm are the best when the traffic application contains infrequent transmission of large volume of data. In the context of uplink traffic in GPRS network, the FCFS with priority mechanism can be used with the three access modes. The dynamic scheduling mechanism operates only with the dynamic access mode. In contrast, the FCFS with windows operate with the three access modes. VI. Conclusion Figure 7. The throughput vs the MSs number (FUNET).,,. Figure 8. The delay vs the MSs number (FUNET) "ro1us.1"~.0*i Figure 9. The loss percent vs the MSs number (FUNET) In this paper, we discuss varied scheduling algorithms and their performance is analyzed by simulation methods. The definition of applications used in GPRS network and these applications modeling are discussed. Our simulation results show the benefits of the proposed algorithm and compare the performance of the different scheduling algorithm in GPRS context. The WWW application can give a valuable model of infrequent transmission of large volume of data. From this cause, the rule of service disciplines on GPRS performance using WWW application is under study. In this paper, the best effort service was studied. However, the introduction of guaranteed performance services in GPRS network requires more complicated studies of the scheduling issue. REFERENCE [l] ETSI Doc. Draft EN, GSM 04.60: GPRS, MS-BSS interface, RLCMAC Protocol, ver , [2] ETSI Doc. Draft EN, GSM 02.60: GPRS, Service Description Stage1, ver , 04/1999. [3] W. Ajib and P. Godlewski, "Acknowledgment Procedures at Radio Link Control level in GPRS", proc. of ACM MSWiM'99, Seattle, WA. Aug [4] W. Ajib and P. Godlewski, A proposal of an Access persistence Protocol over Data Wireless Networks, proc. of IEEE IPCCC 00, Phoenix, AR, Feb [5] G. Brasche and B. Walke, Concepts, Services and Protocols of the New GSM phase 2+ (GPRS), IEEE Comm. Mag. pp , August [6] ETSI Doc. TS, GSM 03.64: Overall description of the GPRS radio interface, ver , [v W. Ajib and P. Godlewski, "Acknowledgment operations in the RLC layer of GPRS", proc. IEEE MoMuC'99, San Diego, CA. Nov [8] J. Cai and D.J. Goodman, "General Packet Radio Service in GSM", IEEE Comm. Mag. Oct. 1997, pp [9] ETSI Doc. Evaluation Criteria for the GPRS Rdaio Channel, SMG2 GPRS ad-hoc, Tdoc 29/96, 2/1996. [lo] Hui Zhang, Service Disciplines for Guaranteed Performance Service in Packet-Switching Networks, Proceedings of IEEE, Oct [ 113 IS0 8208: Information processing systems -data communications -X.25 packet level protocol for data terminal equipment. [12] L. Kleinrock. Queuing systems Volume 11: Computer Applications. Wiley, /00/$ IEEE VTC2000