Upload
phammien
View
227
Download
2
Embed Size (px)
Citation preview
LTE eNB Traffic Analysis and Key Techniques
Towards 5G Mobile Networks
T. B. Iliev*, Gr. Y. Mihaylov*, T. D. Bikov*, E. P. Ivanova*, I. S. Stoyanov** and D. I. Radev*** *University of Ruse, Department of Telecommunication, Ruse, Bulgaria
**University of Ruse, Department of Electrical and Electronic Engineering, Ruse, Bulgaria ***University of Telecommunications and Posts, Department of Telecommunication, Sofia, Bulgaria
e-mail addresses: [email protected], [email protected], [email protected], [email protected],
[email protected], [email protected]
Abstract - With the fast growth of mobile data services, rich
services deliver a brand new experience to end users, and
also bring about new opportunities for operators. Some
applications, such as vehicular and industrial applications,
demand a level of reliability that wireless communication
systems typically are not able to guarantee. This paper
provides a framework that enables these applications to
make use of wireless connectivity only if the transmission
conditions are favorable enough. In this paper, we will
partially compare the fourth and future fifth generations of
mobile networks and their applicability and were carried
out measurements of data traffic amount and quality
characteristics.
Keywords: Ultra-Reliable Communication, Fifth-
Generation (5G), Mobile communication systems
I. INTRODUCTION
Few past years, there was been a tremendous surge in the demand for mobile and wireless connectivity, which is forecasted to grow exponentially within the near future. However, the Fifth-Generation (5G) mobile communication system [1] that is currently under discussion will not only have to cope with an increasing demand of traffic volume, but also provide a wider range of applications with new requirements in terms of reliability, availability and efficiency. Instead of integrating new radio access concepts such as Device-to-Device (D2D), Vehicle-to-Vehicle (V2V) communications, Massive Machine Type Communications (MMC), or Moving Networks (MN), the support of Ultra-Reliable Communication (URC) is seen as a key enabler for 5G communication systems [2].
Link reliability is the ability of a radio link to transmit and receive a certain amount of information successfully within a predefined deadline. System reliability is the ability of a system to accurately indicate the absence of link reliability to the application, and at the same time, to ensure the presence of link reliability as often as possible when required by the application [3, 4]. URC services such as road safety applications require very high rates within low deadlines circa 100 ms. Due to the sensitive nature of these applications, it is from very high importance to warn the application about the lack of link reliability according to the specific requirements of each application. Wireless communication systems are not
designed to provide reliability at all times and in every reception scenario, as this would result in an overdesigned system with a very ineffective air interface in terms of data rate and power consumption. With this approach may harm the acceptance of URC services and restrict their usage. In this paper we define the URC concept in a general manner, i.e., without relying on details of any air interface requirement or radio transmission scheme. This method is motivated by the fact, that definition from the applications point of view is required in order to allow URC services to be deployed in a wide variety of scenarios. From this point of view, the implementation details related to the wireless communication system are not included in the suggested URC concept.
The rest of the paper is organized as follows. Section 2 gives a brief overview and theoretical representation of the system concept for Ultra-Reliable Communication. Section 3 presents the conducted experimental measurements of data traffic amount and quality characteristics of two Bulgarian mobile operators. Section 4 concludes the paper.
II. KEY ENABLERS FOR 5G COMMUNICATION
SYSTEMS
Despite the fact that it is practically impossible to ensure error-free wireless communication - it is feasible to derive boundary conditions for the transmission success. As illustrated in Fig. 1, the system concept for URC is based on a “Reliable Transmission Link” (RTL) that is set to transmit packets successfully and within a predefined deadline, and an “Availability Estimation and Indication” (AEI) mechanism that is able to reliably calculate the availability of the RTL under the corresponding conditions. Availability Indicator (AI) signals the outcome of the AEI to the application. An application requests an RTL by sending an Availability Request (AR) to the AEI. Depending on the implementation characteristics, the AR contains information such as the packet size, the maximum acceptable delay until successful reception or the maximum allowable error probability.
For the availability estimation, the AEI needs to monitor the channel conditions, e.g., by evaluating the Signal-to-Noise and Interference Ratio (SINR) and/or the ACK/NACK statistics of the retransmission protocols used at link level.
624 MIPRO 2017/CTI
Service/Application
Availability Estimation and Indication (AEI)
Reliable Transmission Link (RTL)
Availability Request (AR)
Availability Indicator (AI)
Figure 1. System concept of URC
Typically, the AI is a binary value, i.e., either RTL available (AI=1) or unavailable (AI=0). After indicating the RTL availability, the application will be able to use it by transmitting data packets over the RTL (it is not shown in Fig 1).
A. Mathematical representation
In this section we formulate mathematically the URC concept by adopting a simple time-slotted model, in which each time slot (τ) corresponds to the time interval [t, t + DAR), where DAR is the maximum delay tolerated by the application. According to this definition is defined:
successfulnot ison transmissi,0
successful ison transmissi,1
RTL
For the availability indication is used a simple binary signaling format per time slot:
0
1AI
where AI refers to availability of RTL for time slot τ (AI(τ)=1) and the non – availability of RTL for time slot τ (AI(τ)=0).
On Fig. 2 is shown the URC state transition probabilities divided into two stages. The principle of the URC concept is that an application should rely on the wireless communication only in those situation in which the link reliability is guaranteed with a certain probability. The AI indicates the availability of a RTL for time slot τ to the application, the probability of successful transmission for the time slot τ must be over a certain value, PUR, according to the application requirements. This criteria is refer as the ultra-reliable requirement, which can be explained by (3):
1AI1RTLPrP11
where P1|1 ≥ PUR. The idea of the URC system design is to maximize the availability of the RTL under the predefined requirement:
URPP
11
1
s.t
1AIPrPmax
AI=1
AI=0
RTL=1
RTL=0
Availability Request
P1
P2
P0|0
P1|1
P1|0
P0|1
Figure 2. URC State Transition Probabilities
According to Bayes’ rule and from Fig.2, we have:
11
1
P
1RTL1AIPr1RTLPr
1AIPrP
It can be concluded that there are two possibilities in order to improve the URC concept:
improving the transmission model (Pr(RTL(τ) = 1));
improving the estimation of the AI (Pr(AI(τ) = 1|RTL(τ) = 1)).
The first possibility can be achieved by more effective modulation and forward error correction (FEC) schemes, whereas the second possibility can be accomplished through more precise channel estimation and prediction methods.
It is interesting to note that the two probabilities P0|1 and P1|0 in Fig. 2 correspond to the Type I and Type II errors [5] in statistical probability analysis:
1AI0RTLPrP 10
0AI1RTLPrP 01
B. URC system concept
In this section we will explain the URC concept by using a simple example based on the predicted SINR. We
assume that the SINR for time slot is () and define
that the transmission is successful for time slot if the corresponding SINR is larger than or equal to a given
threshold ι (() ≥ ι), which is given by the used modulation and FEC scheme including the use of retransmissions. The AEI signals to the application AI=1
if the predicted SINR for time slot , p(), is larger than
or equal to the threshold ι (p() ≥ ι). We can formulate the optimization problem of the URC system model as:
lpl
l
P
Prs.t
PrPmax
11
1
where the Type I and Type II error can be expressed as:
MIPRO 2017/CTI 625
lp l10PrP
lp l01PrP
In this model, it would be possible to improve the
URC concept by decreasing ι by means of more robust modulation and FEC schemes, or by optimize the
computation of p() in order to predict more accurately the availability of the RTL.
III. DATA USAGE AND TRAFFIC ANALISYS
In this section we will compare two Bulgarian Mobile operators and their data traffic amount and quality characteristics, will describe the benefits of developing and implementing next generation mobile network standards and what are the key point which eventually needs improvement. To highlight the differences, the research was done in two different locations as follows:
Service Provider 1 – Mid-size City with approximately 76 000 citizens
Service Provider 2 – Capital City
Used for the study Base Stations (BS), from both Service Providers (SP) are from one and the same manufacturer. This approach, to compare two base stations from two different cities has selected to highlight the traffic usage, throughput and TBF Drop Rates in 2G/3G and LTE standards used in the selected BSs.
So far for Service Providers, the ultimate goal of network optimization is the improvement of user experience and enhance user satisfaction. Nonetheless, network optimization does not perfectly matches Quality of Service (QoS) improvement. Traditional network Key Performances Indicator (KPI) optimization largely deals with signaling, such as the Temporary Block Flow (TBF) establishment success rate and attach success rate. Although operators have been paying attention to them, these indexes cannot accurately reflect QoS. QoS indicators like “slow page loading”, “slow download rate” and “failure to open page” are indexes on the data service layer. As mentioned above, the TBF drop is KPI which is related to 2G data called GPRS/EDGE in GSM Network. TBF drop indicate how often the services is disconnected when GPRS or EDGE data Services are used [6].
A. 3G/UMTS
On September 1995, officially first GSM network in Bulgaria took place. It has been started working with six base stations covering a few central parts of the Capital City. The total number of active 3G/UMTS base stations, twenty years later in this country is nearly 5000. The 3G coverage reaches 96% of the areas in the country and 99% of its population.
The diagrams of current TBF drop rates in downlink and respectively uplink direction are shown on Fig.3 and Fig.4.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829
TBF
Dro
p R
ate
Time (Days)
Service Provider 1 Service Provider 2
Figure 3. Measured TBF Drop Rates in Downlink direction for a
month
As mentioned above, the statistics are from BSs, physically located in two different geographical locations.
As shown in Fig.3, the second Service Provider`s BS is either not optimally utilized or very precisely planned and implemented in the selected region, because of the low volume of TBF Drop Rate in downlink direction. Another parameter, which is from major importance for consumers and in which Service Providers are working to improve, is throughput. Large number of different throughput values for 3G/UMTS standard are known.
However, on Fig.4 and Fig.5 are depicted diagrams of currently measured throughput in Mbps.
The metrics are measured in uplink and downlink direction for a period of one month. In both figures, Service Provider 2 notably have wider throughput. In downlink, SP2 has peak value of 164 Mbps, where SP1 has just 20,5 Mbps and in uplink, SP2 has maximum value of 73,7 Mbps where SP1 has almost constant 12 Mbps with peak from only 14 Mbps.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
0
20
40
60
80
100
120
140
160
180
Time (Days)
Thro
ug
hp
ut
(Mb
ps)
Service Provider 1 Service Provider 2
Figure 4. Throughput, measured in Megabits/s in Downlink direction
for a month
0
10
20
30
40
50
60
70
80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Thro
ug
hp
ut
(Mb
ps)
Time (Days)
Service Provider 1 Service Provider 2
Figure 5. Throughput, measured in Megabits/s in Uplink direction on
monthly basis
626 MIPRO 2017/CTI
0
200
400
600
800
1000
1200
1400
1600
1800
1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829
Da
ta T
raff
ic (
MB
)
Time (Days)
Service Provider 1 Service Provider 2
Figure 6. Data Traffic amount measured in Megabytes for one month
Last, but not least, Fig.6 represent graphical results of studied and analyzed Data Traffic from the two Service Provider`s Base Stations. Measured data traffic volume is for period of one month in Megabytes.
B. Long Term Evolution
Since the summer of 2014 the fourth generation mobile internet (4G) started to be offered in Bulgaria. As of September 2015, Bulgarian customers using Long Term Evolution (LTE), which is Fourth Generation (4G) wireless broadband technology, are 0.3% of all Internet using mobile users. According to offered plans, the high-speed internet access can provide download speeds of up to 75 Mbps and upload speeds of up to 25 Mbps. The bandwidths used by the fourth generation mobile networks are 900, 1800, and 2100 MHz. In 2015, the 900 MHz band is divided among the largest three operators in the country and in early September 2015 the Commission for Regulation of Communications (CRC) announced that it will hold a tender for permits for using the 2,50-2,69 GHz band.
With high data speed, the LTE can provides downlink peak rates of 300 Mbps. Some of the advantages of LTE over 3G/UMTS are High Speed, High Capacity, better Spectral Efficiency and others [7, 8]. To compare the two standards and to present the advantages of using the LTE, we analyze the data from one Base Station running LTE and represent it graphically (Fig.7 and Fig.8).
0
5
10
15
20
25
30
35
1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930
Thro
ug
hp
ut
(Mb
ps)
Time (Days)
Chart Title
Throughput DL Throughput UL
Figure 7. Throughput, measured in Megabits/s in Uplink and
Downlink direction on monthly basis
0
1000
2000
3000
4000
5000
6000
7000
8000
1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930
Da
ta T
raff
ic (
MB
)
Time (Days)
Data Traffic DL Data Traffic UL
Figure 8. Data Traffic amount measured in Uplink and Downlink
direction for one month
On both charts above is shows that the throughput and the Data Traffic amount, passed through the specified BS running LTE, for a period of one month is significantly higher than those passed through the same BS, using the 3G/UMTS standard.
IV. CONLUSION
Comparing the future fifth-generation mobile networks with current and the previous generations of mobile services were not so extremely sensitive in manners like delay. Expected characteristics for 5G networks are data rates of tens of megabits per second for tens of thousands of users; data rates of 100 megabits per second for metropolitan areas and significantly enhanced spectral efficiency, compared to 4G. To meet these specific requirements contemporary mobile networks should be improved in signaling efficiency, better coverage and to offer lower latency.
The expectations of fifth generation mobile networks are the mobile subscribers to grow even more, including D2D, V2V communication and Internet of Things (IoT), which demands a lot of improvements.
REFERENCES
[1] “METIS - Mobile and wireless communications Enablers for the Twenty-twenty Information Society.” [Online]. Available:https://www.metis2020.com/
[2] D. Gozalvez Serrano, H. D. Schotten, P. Fertl, R. Sattiraju and Z. Ren, “Availability Indication as Key Enabler for Ultra-Reliable Communication in 5G”
[3] F. B. F. Bai and H. Krishnan, “Reliability Analysis of DSRC Wireless Communication for Vehicle Safety Applications,” 2006 IEEE Intelligent Transportation Systems Conference, 2006.
[4] A. Birolini, Reliability Engineering: Theory and Practice, 2010.
[5] S. Kay, Fundamentals of Statistical Signal Processing: Detection theory, ser. Prentice Hall Signal Processing Series. Prentice-Hall PTR, 1998
[6] Sadinov S., D. Koleva, Software Research the Quality of Signals in Cellular Network with the Software ASSET, International Scientific Conference Unitech’15, Gabrovo, Bulgaria, Vol.2, ISSN 1313-230X, pp.142 ÷ 146
[7] J. F. Monserrat, , H. Droste, Ö. Bulakci, J. Eichinger, O. Queseth, M. Stamatelatos, H.Tullberg,V. Venkatkumar, G. Zimmermann, U. Dötsch, A. Osseiran, “Rethinking the Mobile and Wireless Network Architecture: The METIS Research into 5G”, EuCNC 2014, June 23-26, Bologna, Italy, 2014
[8] I. Silva, G. Mildh, M. Säily, S. Hailu, “A novel state model for 5G radio access networks“, 5G RAN design Workshop, IEEE ICC 2016, 27 May 2016
MIPRO 2017/CTI 627