International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2859
ISSN 2229-5518
Packet scheduling in LTE mobile network
Noha Sheta ,Fayez W. Zaki and Sherif Keshk
Abstract
Selection of proper packet scheduling algorithm is crucial component in QoS provisioning over next generation networks, especially LTE networks which
is expected to provide higher throughput with better QoE ( Quality of Experience ) among the users. A key factor affects the QoS & QoE of LTE
networks, is the packet scheduling algorithms used in them.
In this paper ,the performance of most famous packet scheduling algorithms such as Priority Queuing (PQ), Weighted Fair Queuing (W FQ), Round
Robin (RR) and First In First Out (FIFO) in LTE mobile networks had been studied.
OPNET modeler 17.1 simulator used to compare the results in different scenarios which contain different applications.
Key Words: Packet scheduling, LTE mobile networks, QoS, OPNET
—————————— ——————————
other through the X2 interface and connected to the core
network (EPC "Evolved Packet Core") through the S1
The demand for IP packet service is growing rapidly in
mobile communication in recent years. IP services enabling
multimedia retrieval, web browsing, data download, and
ecommerce, etc., will play a more and more important role in the future. Generally, the features of those services require high-speed data rate. Thus 3GPP introduced the
3GPP long term evolution (LTE) standards to the mobile data markets. 3GPP LTE provides high speed wireless
communications based on universal mobile terrestrial systems (UMTS)/high speed packet access (HSPA) network technologies on the way towards 4G mobile networks.
However, unfortunately the radio resource is limited and thus needs sophisticated management schemes.
So, there is strong motivation and challenge beyond scheduling procedure and resource allocation in order to improve system performance by increasing spectral efficiency of the wireless interface and hence improve overall network capacity.
The reminder of this paper is organized as follows: Section II shows important components of LTE .Section III network presents QoS in LTE network. Section IV gives brief description of most famous scheduling algorithms.. Section V presents simulation result of this paper study. Finally, conclusions & future works are discussed in Section VI.
LTE has introduced a number of new technologies when compared to the previous cellular systems. They enable LTE to be able to operate more efficiently with respect to the use of spectrum, and also to provide the much higher data rates that are being required.
Number of nodes reduced from 4 nodes used in W-CDMA (which are NodeB, RNC, SGSN & GGSN) to only 2 nodes (e-NodeB and SAE-GW).
The radio network consists of the E-Node B’s which are
responsible for scheduling and are interconnected to each
interface.
OFDM technology has been incorporated into LTE because it enables high data bandwidths to be transmitted
efficiently while still providing a high degree of resilience
to reflections and interference. In OFDM, the data stream is
distributed over many subcarriers. Each subcarrier will
thus have a slow symbol rate and correspondingly, a long
symbol time. This means that the Inter Symbol Interference (ISI) is reduced. SC-OFDM for the uplink since the power amplifier in the UE can be manufactured at a lower cost then.
As a result of introducing new applications in mobile network and as it is not limited to voice and SMS applications, each application has its own QoS demands to guarantee satisfaction for both of user and mobile
operators. In LTE, the QoS is provided by means of a bearer
which is responsible for the priority that is given to a packet
flow across the LTE network. Bearers are established after the successful authentication and registration of the user in the LTE network.
IJSER © 2013 http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2860
ISSN 2229-5518
GBR requires resources to be reserved for its applications.
Bearers are used for services that do not have strict QoS
constraints. It is useful for background, interactive traffic
packets, where packets are grouped in flows and each flow
has its own weight.
The rate of each flow is determined based on below
equation
classes such as email, web browsing, file download, etc.
W(i) R(i) = R × ∑ W
(1)
is very important in LTE as it defines packet forward treatment at each node, such as scheduling weight, admission control priority, queuing threshold, etc.
This supports to determine relative priority of a bearer over other bearers in case of congestion.
which identifies maximum bit rate supported by bearer.
Below table has example for values of above mentioned factors.
It does not take number of packets in each flow into
consideration, just care about weight of the flow.
which depends only on priority of class the traffic belongs to without taking into consideration fairness between users
FIFO is the simplest scheduling algorithm. Packets coming from all the input links were en-queued into a FIFO memory stack, and then they were de-queued one by one on to the output link. So it simply queues processes in the
order in which they arrive in the ready queue. Since context
switches only occur upon process termination, and no
reorganization of the process queue is required, scheduling
overhead is minimal. Throughput turnaround time, waiting
time and response time can be low. No prioritization occurs, thus this system has trouble meeting process deadlines. The lack of prioritization does permit every process to eventually complete, hence no starvation.
LTE network was composed in OPNET which contains 6 cells, each has one NodeB & 5 users. NodeBs connected to Evolved Packet Core (EPC) through E1. EPC connected to IP cloud which connected to router. The router connected to server which supports HTTP applications. Simulation run for 7 minutes.
Applications used by users in this simulation are:
1. HTTP: in “Best Effort” QoS level = 0
2. Voice: in “Interactive Voice” QoS level = 6
3. Video: in “Streaming Multimedia” QoS level =4.
Table 1
Packet scheduling is the process of resolving contention for bandwidth. A scheduling algorithm has to determine the allocation of bandwidth among the users and their transmission order. One of the most important objectives of a scheduling scheme is to satisfy the Quality of Service (QoS) requirements of its users while efficiently utilizing
the available bandwidth
It is a generalization of Fair Queuing (FQ). WFQ is a data packet scheduling technique that is used for various size
IJSER © 2013 http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2861
ISSN 2229-5518
Fig.2 ( Voice received byte rate )
WFQ: Green grapgh PQ: Red grapgh FIFO: Blue grapgh
QoS was assessed by the following parameters in simulation:
Fig.1 shows that WFQ has the highest rate
As shown in Fig.2, all scheduling algorithms resulted in same received voice traffic
MOS provides numerical measurements of user Quality of
Experience in voice telecommunications
MOS score | Mapping score to user(s) satisfaction |
4.3 - 5 | Very much satisfied |
4 - 4.3 | Satisfied |
3.6 - 4 | Many users satisfied |
3.1 - 3.6 | Many users dissatisfied |
2.6 - 3.1 | Nearly all users dissatisfied |
Table 2
As shown in Fig.3, the 3 algorithms has different values starting from the 5th minute, with PQ silghtly higher than WFQ & FIFO has the lowest values.
Fig.3 (Average MOS value of Voice)
IJSER © 2013 http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2862
ISSN 2229-5518
As shown in Fig.4 PQ has the lowest delay, this is because
Voice has the highest priority & PQ take into its
consideration only priority of traffic
Fig.6 (Average received packet rate of HTTP)
Fig.4 (Time average E2E delay of Voice)
Jitter is the maximum difference in one way delay of packets over a particular time interval.
Fig.5 (Time average of jitters rate for voice)
As shown in fig.5, almost all scheduling algorithms results in same jitter
As shown in fig.7, WFQ has the highest received byte rate
Fig.7 (Time average of received bytes)
Fig.8 ( Time average for E2E delay )
As shown in fig.8, PQ has the lowest E2E delay as Video has the 2nd higher priority in simulation, while WFQ has
IJSER © 2013 http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2863
ISSN 2229-5518
the highest value as video load is very high & the main
target of WFQ to achieve fairness between users.
Special thanks to NTRA (National Telecommunication Regulatory Authority) in Egypt and their Sr.Manager Mohannad El-Megharbel.
After analysis and simulation of the three scheduling algorithms in LTE network, it is found that WFQ is the best algorithm to use to achieve both highest downlink load and fairness between users using different applications.
In future work, effect of changing weights of WFQ to be
dynamic instead of static & constant weights will be
studied. The weights changing criteria should take into
consideration Operator revenue as well as QoE from users, with flexibility to change degree of fairness based on above mentioned two important factors to achieve satisfaction of both Mobile operators & Mobile users.
[1] Huda Adibah Mohd Ramli,Kumbesan Sandrasegaran and Khaizuran Abdullah “A Study of Packet Scheduling with Mobile Cellular Channel Impairments”
1.
[2] C. Turyagyenda, T. O'Farrell and W. Guo “Long Term Evolution Downlink Packet Scheduling using A Novel Proportional-Fair-Energy Policy”
2.
[3] Alcatel-Lucent Strategic White Paper, "Options for Providing Voice over LTE and their Impact on the GSM/UMTS Network"; www.alcatel-lucent.com, August 2009.
3.
[4] Siomina, I.; Wanstedt, S.; , "The impact of QoS support on the end user satisfaction in LTE networks with mixed traffic," IEEE 19th International Symposium on Personal, Indoor and Mobile Radio Communications, pp.1-5, 15-18 Sept. 2008.
4.
[5] Li, X.; Toseef, U.; Weerawardane, T.; Bigos, W.; Dulas, D.; Goerg, C.; Timm-Giel,A.; Klug, A.; , "Dimensioning of the LTE S1 interface," Third Joint IFIP Wireless and Mobile Networking Conference (WMNC), vol., no., pp.1-6, 13-15 Oct. 2010.
5.
[6] A. Demers, S. Keshav, and S. Shenker "Analysis and simulation of a fair queueing algorithm", In Symposium proceedings on Communications architectures and protocols", ACM, New York,
NY, USA, 1989.
6.
IJSER © 2013 http://www.ijser.org