Sie sind auf Seite 1von 6

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 5, May 2014)

Comparison of Different Scheduling Algorithm for LTE


Ronak D. Trivedi1, M. C. Patel2
The throughput of a UE depends on the different factor
like scheduling algorithms, distance from eNodeB,
multipath environment, multiple antenna techniques and
UE speed. In this paper, we consider the effect of
scheduling algorithm with throughput performance. We
apply proportional fair (PF) scheduler, round robin and best
CQI for LTE in order to find best scheduler which provides
high-quality cell throughput and improved fairness. The
scheduler has to serve multiple users and also tries to meet
individual user requirements on bit rates and delays. The
fairness of the scheduler is a way to couple the scheduler
and it help to the weakest users is proposed.

Abstract In this paper we evaluate the performance of


Packet Scheduling (PS) for different packet scheduling
algorithms of 3GPP UTRAN Long Term Evolution (LTE)
Downlink. Packet scheduling is importance in 3G LTE,
because different types of traffic with different Quality of
Service requirements are competing of the resources. In this
paper, packet scheduler for LTE downlink is described.
Comparison of three basic packet scheduling algorithm with
their simulation result with different amount of fairness is
explained. This paper shows that by dividing the packet
scheduler into a time domain, frequency domain and also
utilizing those using different algorithms, the throughput
fairness between users can be effectively controlled.
KeywordsLTE, round robin, proportional fair, best
CQI, scheduling.

II. B ACKGROUND O F T HE INVENTION


LTE aims at ambitious goals such as e.g. [2], [3], [4]
LTE has goal about Peak data rate is 100 Mbps in
downlink and 50 Mbps in uplink
Increased cell edge throughput
Significantly improved spectral efficiency e.g. 2-4
times better than in 3GPP Release 6
User plane latency below 5 ms with 5 MHz or higher
spectrum allocation
Significantly reduced control plane latency e.g.
transition time of less than 100 ms from a campedstate to an active state
Scalable bandwidth from 1.25 MHz to 20 MHz.
To achieve the aim of LTE, advanced Radio Resource
Management functions have been defined. LTE contain
some algorithms for example Hybrid ARQ (HARQ), Link
Adaptation (LA) and Channel Quality Indication (CQI).
HARQ is utilizing for fast retransmissions of the packets
which are in correct. HARQ is use to keep the radio
interface delay minimum. User Equipment (UE) is use to
measures the received channel quality, e.g. SINR, and news
the channel dependent CQI reports in uplink. They give
information to the Radio Resource Management (RRM)
about time and frequency variant channel quality like PS
and LA. LA select different modulation and coding
schemes (MCS) based on CQI reports to maximize the
spectral efficiency. [5].

I. INTRODUCTION
Recently the increase of mobile data usage and
appearance of new application like mobile TV, Web2.0 and
other streaming contents forced the 3rd Generation
Partnership Project (3GPP) to develop the Long-Term
Evolution (LTE). Long term evolution (LTE) is a latest
radio access technology planned by the 3GPP in order to
provide a smooth journey towards fourth generation (4G)
wireless systems.
In the 3GPP LTE radio network architecture, there is
only one node between the user and the core network
known as eNodeB which is used to operate all radio
resource management (RRM) functions. Packet scheduling
is function of the RRM. Because of its smart selections of
users and transmission of their packets, the radio resources
are utilized efficiently and QoS (quality of service) is also
maintained.
Packet scheduling for wireless communications has been
an active research area in recent years, because there has
been rapidly increasing demands on data services with the
likely to explode progress of traffic such as Internet, Email,
multimedia. To support these packet data services, the
scare and limited wireless resource must be used in best
way to increase capacity and security QoS. Providing
priority or fairness is also an open issue in wireless system.
However, it is not simple to meet all of these requirements.

334

International Journal of Emerging Technology and Advanced Engineering


Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 5, May 2014)
A maximum scheduled users stricture defines the
maximum amount of users that can be scheduled in each
TTI. TD-PS schedule both the new transmissions and
pending HARQ retransmissions. HARQ retransmissions
can be prioritized in two ways. Either all users with
awaiting HARQ retransmissions are automatically chosen
for the SCS (i.e. before the TD scheduling) or the HARQ
retransmissions are prioritized for TD scheduled users (i.e.
after the scheduling).

Both time (TDD) and frequency division duplex (FDD)


modes can be used in LTE. In downlink the time is
separated into 1 ms Transmission Time Intervals (TTI) and
in frequency 180 kHz Physical Resource Blocks (PRB).
LTE is optimized for packet data transfer and the core
network is purely packet switched. In [5] the authors study
the spectral efficiency of LTE DL with different UE
receiver structures and with advanced SIMO receivers they
achieved about 1.25 bits/s/Hz. In [6] the authors show
about 1.56 bits/s/Hz spectral efficiency.
The time domain (TD) packet scheduler choose a subset
of all users linked to the base station (called evolved NodeB (eNB) in LTE ) and the FD scheduler does the real
frequency allocation for the users. This division is suitable
for two reasons: using different schedulers in both
scheduling domains provides scheduling flexibility, since
both domains can be independently configured.
The scheduling in both the TD and FD is through
algorithm-specific scheduling priority metrics. A priority
metric generally provide the function of obtaining a certain
general characteristic of the scheduling algorithm. These
characteristics can be e.g. certainty of regular scheduling,
fairness among users or highest possible spectral
efficiency. Most scheduling metrics can be used for both
TD and FD scheduling and they can be also a grouping of
different metrics.
Packet scheduler cooperates with CQI manager, link
adaptation and throughput measurement (TPM). CQI
calculation gives us PRB dependent channel quality
information for the use of PS metrics and link adaptation
[7]. Inner loop LA (ILLA) selects the best MCS for the
user depending on the effective CQI of the allocated RBs
and amount of data for the UE in question in Evolved
Node-B (i.e. base station) buffer. ILLA provide
instantaneous throughput estimates for PF metrics. Outer
loop LA (OLLA) aims at controlling the user average
BLER for the first transmissions in order to fix the HARQ
operating point to optimal value. Throughput measurement
calculate the past user throughput by using a recursive
averaging filter [8].

B. Frequency domain packet scheduling


The purpose of the frequency domain packet scheduler is
to allocate PRBs for the users in the SCS provided by TDPS. However, it should be distinguished that users in the
SCS are consider only candidates, since FD-PS does not
necessarily guarantee that all users are being allocated
frequency resources. A user may be given any number of
PRBs, and the PRBs do not need to be consecutive. The
algorithms particular priority metrics are taken into
account in PRB selection.
III. DOWNLINK RESOURCE ALLOCATION
OFDMA is used for downlink transmission in LTE. Data
is allocated to the UEs in terms of Resource Blocks (RB).
In time, the length is 0.5 ms of a RB is one slot in frame.
With 15 kHz sub-carrier spacing, in normal cyclic prefix
the number of symbols in one slot is 6 and for extended
cyclic prefix 7. The length of a RB is 180 kHz, in terms of
frequency. The number of sub-carriers in the 180 kHz span
is 12 for 15 kHz sub-carrier spacing.
The eNodeB allocate different RBs to an exacting UE in
either localized or distributed way. The eNodeB uses DCI
format 1, 1A, 1B, 1C, 1D, 2, 2A or 2B on PDCCH to
transmit the resource allocations on PDSCH for the
downlink transmission.
The scheduler at eNodeB attempts for appropriate
allotment of the resources among UEs. The UE reports CQI
(Channel Quality Indicator) which helps eNodeB to
approximate the downlink channel quality. By the use of
CQI report about the whole downlink bandwidth or about
information about sub-band, the eNodeB can organize. CQI
reporting for different sub-bands needs more uplink
resources.
The channel dependent scheduling requires maintaining
some fairness among the users and on the other hand it
would like to provide higher cell throughput in their
resource allocations. There is a connection between
fairness and cell throughput.

A. Time domain packet scheduling


The purpose of time domain packet scheduling is sharing
out of all users requesting frequency resources. The choice
is done based on calculated priority metrics, based on e.g.
L2 buffering delay, throughput or current channel
conditions. Note that in TD-PS we require to utilize
average full band CQI, since TD- PS does not consider
actual PRB allocations. The users with the higher priority
metric are appended, its called Scheduling Candidate Set
(SCS), which is then passed to the FD packet scheduler.

335

International Journal of Emerging Technology and Advanced Engineering


Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 5, May 2014)
Different scheduling methods are shown below in order
to address this trade off.
1. Round Robin (RR): The scheduler provides resources
cyclically to the users without considering channel
conditions into account. Its a simple procedure giving
the best fairness. But it would propose poor performance
in terms of cell throughput. RR meets the fairness by
providing an equal share of packet transmission time to
each user. In Round Robin (RR) scheduling the
terminals are assigned the resource blocks in turn (one
after another) without considering CQI. Thus the
terminals are equally scheduled. However, throughput
performance degrades significantly as the algorithm does
not rely on the reported instantaneous downlink SNR
values when determining the number of bits to be
transmitted.
2. Proportional fair (PF): Main purpose of Proportional
Fair algorithm is to balance between throughput and
fairness [8] among all the UEs. It tries to maximize total
[wired/wireless network] throughput while at the same
time it provides all users at least a minimal level of
service. PF was originally developed to maintain NRT
service in code division multiple access high data rate
(CDMA-HDR) system. The scheduler can affect
Proportional Fair (PF) scheduling by allocating more
resources to a user, comparatively with better channel
quality. This is done by giving each data flow a
scheduling priority that is inversely proportional to its
anticipated resource consumption. This gives high cell
throughput as well as fairness satisfactorily. Thus,
Proportional Fair (PF) scheduling may be the best
option.
3. Best CQI: This scheduling algorithm is used for strategy
to assign resource blocks to the user with the best radio
link conditions. The resource blocks assigned by the
Best CQI to the user will have the highest CQI on that
RB. The MS must feedback the Channel Quality
Indication (CQI) to the BS to perform the Best CQI. In
order to perform scheduling, terminals send Channel
Quality Indicator (CQI) to the base station (BS).
Basically in the downlink, the BS transmits reference
signal (downlink pilot) to terminals. These reference
signals are used by UEs for the calculation of the CQI. A
higher CQI value means better channel condition.

The main simulation parameters utilized are determined


by the 3GPP simulation cases and Performance with round
robin, proportional fair and best CQI scheduling. It has
been observed for five UEs at various distances from the
eNodeB and mapping of UEs and eNodeB is depicted in
Table I.
A. Mapping of UE and eNodeB within the Cell

Figure 1: The comparative distance between eNodeB and UE. Red dot
represents eNodeB and black dot represents UE.
TABLE 1
SIMULATION PARAMETERS

IV. S IMULATION
5 UEs are placed randomly in three sector of three
eNodeB.

336

Parameters

Assumptions

Transmission bandwidth

2.0GHz

Inter-site distance

5MHz

Receiver noise figure

9dB

Simulation length

100 TTI

UE speeds of interest

5km/hr

Fair

Thermal noise density

Uplink delay

3 TTIs

Scheduler

Round Robin, Proportional fair, Best CQI

eNodeB TX power

43dBm

International Journal of Emerging Technology and Advanced Engineering


Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 5, May 2014)
Figure 2 shows simulation for Round Robin algorithm.
Here throughput of sector 1 and 2 is 0.55Mbps where
sector 3 is 0.56Mbps.

V. RESULTS
Results of three different algorithms for LTE of LTE
simulator are described in figure 2, 3 and 4, which shows
throughput of third eNodeB for stream 2 for all three
algorithms.

Figure 2: Simulation Results for Throughput (Mbps) vs TTI (sec) for


RR (eNodeB 3, Stream 2).
Figure 3: Simulation Results for Throughput (Mbps) vs TTI (sec)
for PF (eNodeB 3, Stream 2).

337

International Journal of Emerging Technology and Advanced Engineering


Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 5, May 2014)
Figure 3 shows simulation for Proportional Fair
algorithm. Here throughput of sector 1 and 2 is 0.55Mbps
where sector 3 is 1.01 Mbps. The throughput of sector 3 of
proportional fair algorithm is batter then round robin
algorithm.

Figure 4 shows simulation for Best CQI algorithm. Here


throughput of sector 1 and 2 is 0.55Mbps where sector 3 is
0.98 Mbps. Throughput of sector 3 in Best CQI is greater
than Round Robin but lover then Proportional Fair. In the
same way, throughput of other eNodeB for different sector,
streams are shown in table 2.
Table 2 shows the throughput of different eNodeB by
using different sector using Round Robin (RR),
Proportional Fair (PF), and Best CQI scheduling
algorithms. Position of different UEs is shown in Fig 1.
TABLE 2
Simulation Results for Throughput (Mbps).

eNodeB

Stream
no

Sector Throughput in Mbps

1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3

RR
0.92
1.07
0.55
0.98
1.08
0.55
1.08
0.91
1.1
0.78
0.91
1.09
0.55
0.55
0.56
0.55
0.55
0.56

PF
0.55
1.2
0.55
0.55
1.2
0.55
1.05
0.99
0.79
1.1
0.98
0.79
0.55
0.55
1.01
0.55
0.55
1.01

Best CQI
0.55
0.72
0.55
0.55
0.71
0.55
0.79
0.96
0.82
0.77
1
0.82
0.55
0.55
0.99
0.55
0.55
0.98

Table 2 gives information of throughput of different


eNodB for different algorithm. Average throughput is
given in table 3.
TABLE 3
AVERAGE THROUGHPUT (MBPS).

eNodeB
1
2
3

Figure 4: Simulation Results for Throughput (Mbps) vs TTI (sec) for


Best CQI (eNodeB 3, Stream 2).

338

RR
0.8584
0.9784
0.5534

Average throughput
PF
Best CQI
0.7667
0.605
0.95
0.86
0.7034
0.695

International Journal of Emerging Technology and Advanced Engineering


Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 5, May 2014)
Table 3 used to find average throughput, for RR is 0.79
and Best CQI is 0.72, where PFs average throughput is
0.8067.
The above figures and table represent the comparative
throughput performances for three scheduling algorithms.
In RR average throughput of eNodeB 1 and 2 is good but
eNodeB 3 is very less, same in Best CQI where in
proportional fair, overall throughput is good. Proportional
fair provides the UEs close to the eNodeB with higher
throughput. The data rate is will fairly high in most cases
for proportional fair and the overall cell throughput is also
expected to be better with proportional fair.

REFERENCES
[1]
[2]
[3]

[4]

[5]

VI. CONCLUSION
We analyzed the performances of round robin,
proportional fair and Best CQI scheduling methods for
downlink transmission modes in LTE from this paper. It is
found that proportional fair will give very good data rate in
most cases. Round robin provides the UE with good
fairness but proportional fair maintain a balance between
fairness and throughput and so, proportional fair may still
be a better choice. Also results shows that proportional fair
provide good result than Beat CQI and Round Robin.

[6]

[7]

[8]

339

H. Holma and A. ToskalaI, LTE for UMTS - OFDMA and SCFDMA Based Radio Access, John Wiley and Sons, Ltd. 2009.
Physical Layer Aspects for Evolved UTRA, 3GPP Technical
Report 25.814, version 7.1.0, September 2006.
A. Toskala, H. Holma, K. Pajukoski, and E. Tiirola, UTRAN Long
term Evolution in 3GPP, in Proceedings of IEEE Personal Indoor
and Mobile Radio Communications Conference (PIMRC06),
September 2006.
H. Ekstrm, A. Furuskr, J. Karlsson, M. Meyer, S. Parkvall,
J.Torsner, and M. Wahlqvist, Technical Solutions for the 3G Long
Term Evolution, IEEE Communications Magazine, vol. 44, pp. 38
45, March 2006.
N. Wei, A. Pokhariyal, C. Rom, B. E. Priyanto, F. Fredriksen, C.
Rosa, T. B. Sorensen, T. E. Kolding, and P. E. Mogensen, Baseline
E-UTRA Downlink Spectral Efficiency Evaluation, in Proceedings
of the IEEE Vehicular Technology Conference (VTCF06),
September 2006.
K. Higuchi, T. Kawamura, Y. Kishiyama, Y. Ofuji, and M.
Sawahashi,
System-Level
Throughput
Evaluations
in
EvolvedUTRA, in Proceedings of the International Conference on
Communication Systems, October 2006.
T. Kolding, F. Frederiksen, and A. Pokhariyal, Low-Bandwidth
Channel Quality Indication for OFDMA Frequency Domain Packet
Scheduling, in Proceedings of the International Symposium on
Wireless Communication Systems (ISWCS06), September 2006.
A. Jalali, R. Padovani, and R. Pankaj, "Data Throughput of CDMAHDR a High Efficiency-High Data Rate Personal Communication
Wireless System," in IEEE 51st Vehicular Technology Conference
Proceedings, Tokyo, 2000, pp. 1854-1858.

Das könnte Ihnen auch gefallen