You are on page 1of 11

984 J. OPT. COMMUN. NETW./VOL. 9, NO. 11/NOVEMBER 2017 Mikaeil et al.

Performance Evaluation of XG-PON


Based Mobile Front-Haul Transport
in Cloud-RAN Architecture
Ahmed Mohammed Mikaeil, Weisheng Hu, Tong Ye, and Syed Baqar Hussain

Abstract—Time division multiplexing passive optical radio-over-fiber link known as front-haul [2] [e.g., Common
network (TDM-PON) technologies are viewed as an attrac-
Public Radio Interface (CPRI)]. To connect RRHs in multi-
tive solution for flexible and cost-efficient mobile front-
haul for cloud radio access network (C-RAN) architecture. ple locations to the BBU pool, a lot of high-capacity fiber-
However, it is a challenge for TDM-PONs to meet the based front-haul connections are required, which increases
strict latency requirement of mobile front-haul in C-RAN the deployment cost of C-RAN. A functional split of mobile
because they use a dynamic bandwidth allocation (DBA) baseband processing [3] provides a way to address the cost
mechanism to manage upstream traffic. The latency issue
of TDM-PON based mobile front-haul has been extensively issue of fiber-based front-haul, as it reduces the capacity re-
investigated in the literature with particular focus on IEEE quired for front-haul, allowing the usage of low cost and
TDM-PON (e.g., 10G-EPON). However, ITU TDM-PONs such shared front-haul transport technologies such as switched
as XG-PON and XGS-PON have not yet even been explored Ethernet [4,5], OTN architecture [6], and time division multi-
in the context of mobile front-haul. To cover this gap, we
plexing passive optical networks (TDM-PONs) [7]. In this
first evaluate the performance of two recently proposed
XG-PON-standard-compliant DBAs, namely, group-assured study we focus on TDM-PONs, as adopting the other two
GIANT (gGIANT) and round-robin DBA (RR-DBA), over technologies in the access network may need a long time
simulated mobile front-haul traffic. We conclude based on to become feasible.
our evaluation that neither RR-DBA nor gGIANT satisfies
the delay required for mobile front-haul. Therefore, we pro- TDM-PONs are a promising technology for realizing
pose an optimized DBA known as optimized round-robin cost-efficient mobile front-haul transport as they allow
(optimized-RR) that support front-hauling over XG-PONs. the sharing of optical fibers and transmission equipment.
The performance evaluation of our optimized DBA against
gGIANT and RR-DBA shows significant improvement in
However, the latency in TDM-PON upstream transmission
upstream delay and utilization as well as lower packet loss due to the dynamic bandwidth allocation (DBA) remains a
and jitter for front-haul uplink traffic transmitted via big challenge, as it is much higher than the front-haul la-
XG-PON in virtualized C-RAN architecture. tency tolerance in C-RAN (i.e., 300 μs [8]). There are many
DBA proposals in the literature for low-latency mobile
Index Terms—C-RAN; Dynamic bandwidth allocation;
Mobile front-haul; Performance evaluation; XG-PON; front-haul over IEEE TDM-PONs (e.g., 10G-EPON).
XGS-PON. However, ITU TDM-PONs (e.g., XG-PON) still remain un-
explored in this context. In this study, we evaluate in terms
of upstream delay the performance of some recently pro-
posed XG-PON-standard-compliant DBAs over simulated
I. INTRODUCTION mobile front-haul traffic. We also propose an optimized low-
latency DBA to accommodate dynamics of mobile front-
haul traffic in C-RAN architecture.

T he cloud radio access network (C-RAN) is one of the


key technologies for the next generation mobile net-
work 5G, as it simplifies network management and enables II. RELATED WORKS
coordination and pooling of mobile network radio resour-
ces. In C-RAN architecture the digital baseband processing
There are many proposals in the literature that have
units (BBUs) are moved from mobile base station sites to a
studied the latency issue of TDM-PON based mobile
central location known as the BBU pool, which serves a
front-haul. Among these proposals, a mathematical model
group of distributed radio units known as remote radio
for minimal latency mobile front-haul over TDM-PON has
heads (RRHs) [1]. Generally, RRHs are connected to the
been given in Ref. [9] based on queuing theory. For fixed
BBU pool using a high-speed and low-latency digitized
data-rate mobile front-haul traffic (i.e., CPRI), the authors
of Refs. [10,11] have presented performance evaluations of
Manuscript received June 14, 2017; revised August 16, 2017; accepted constant data-rate mobile front-haul over 10G-EPON us-
August 17, 2017; published October 17, 2017 (Doc. ID 297975). ing the fixed bandwidth allocation (FBA) method to over-
The authors are with The State Key Laboratory of Advanced Optical
Communications Systems and Networks, Shanghai Jiao Tong University,
come the upstream transmission latency issue. As for
200240 Shanghai, China (e-mail: ahmed_mikaeil@sjtu.edu.cn). dynamic data-rate mobile front-haul traffic, an approach
https://doi.org/10.1364/JOCN.9.000984 that utilizes mobile uplink scheduling information to

1943-0620/17/110984-11 Journal © 2017 Optical Society of America


Mikaeil et al. VOL. 9, NO. 11/NOVEMBER 2017/J. OPT. COMMUN. NETW. 985

minimize the upstream latency of 10G-EPON based mobile III. XG-PON BASED MOBILE FRONT-HAUL FOR
front-haul has been proposed in Ref. [12]. However, this ap- VIRTUALIZED SMALL-CELL C-RAN
proach requires an additional interface between BBUs and
the OLT to exchange mobile scheduling information.
A. Virtualized Small-Cell C-RAN Employing
Another study in Ref. [13] has introduced the usage of a
simple statistic function along with FBA to accommodate XG-PON Front-Haul Transport
dynamic data-rate front-haul traffic over 10G-EPON.
However, as this approach is based on periodical estima- In conventional small-cell base station (BS) architecture,
tion of RRH traffic, it is difficult for it to accurately capture a transport network known as backhaul [Fig. 1(a)] is
very low scale variation in front-haul traffic at every RRH. used to connect small cells to the mobile core network. By
A recent DBA proposal reported in Ref. [14] has also fo- introducing the concept of small-cell virtualization, the
cused on minimizing the upstream latency of 10G-EPON processing functionalities of a small-cell base station are
based front-haul by reducing the DBA cycle length through decomposed and shared between virtual BBU (vBBU) and
reduction of the DBA grant processing time. However, DBA RRH entities (e.g., MAC-PHY decomposition or the func-
algorithms for IEEE TDM-PONs are not fully compatible tional split between the MAC layer (L2) and the PHY layer
with ITU TDM-PONs due to the fact that IEEE TDM- (L1) of the LTE baseband) [Fig. 1(b)]. As a result, the
PONs are neither synchronous nor have a fixed frame front-haul interface is also introduced to transport mobile
length like ITU TDM-PONs [15]. On the other hand, traffic results from small-cell base station decomposition
ITU PONs have a brief transmission timing scale of 125 [Fig. 1(b)].
μs and support frame fragmentation, which allows them
The typical network topology of virtualized small-cell
to achieve a much lower level of transmission latency than
C-RAN architecture employing XG-PON front-haul trans-
in IEEE PONs [16].
port is illustrated in Fig. 2. In this architecture, RRHs are
All the above-mentioned reasons have motivated us to connected to ONUs, which are placed together with RRHs
investigate the feasibility of front-hauling over XG-PON. at the remote cell site. ONUs are connected using shared
Since there are no other works considering XG-PON, we optical fibers and a passive splitter to the OLT. The OLT is
initially evaluated the performance of two XG-PON-
standard-compliant DBAs over simulated mobile front-
haul traffic using the network simulator NS-3 [17]. The
first DBA is round-robin DBA (RR-DBA) [18], which is
the simplest DBA compliant with the XG-PON-standard
found in the literature, and the second one is group-assured
GIANT (gGIANT) DBA, which has been proposed in
Ref. [19] to optimize XG-PON to transport mobile backhaul
traffic. Based on our initial evaluation, we conclude that
(1) neither RR-DBA nor gGIANT satisfies the latency re-
quired for mobile front-haul and (2) in spite of the fact that
RR-DBA shows better delay performance compared to
gGIANT, it still suffers the drawback of poor utilization
of XG-PON upstream bandwidth under a bursty front-haul
traffic condition. To this end, we introduce optimized-RR
DBA to tackle this problem. Then, we evaluate its perfor-
mance against gGIANT and RR-DBA in a virtualized
small-cell C-RAN architecture. We further introduce an
application use case of optimized-RR DBA to support
5G-ultra-reliable low-latency communication (URLLC) ser- Fig. 1. Small-cell transport network: (a) conventional small-cell
vices over a 100 MHz 5G channel in distributed C-RAN BS with backhaul and (b) virtualized small cell with front-haul
architecture employing XGS-PON front-haul. and backhaul.

The rest of the study is organized as follows: the network


architecture of virtualized small-cell C-RAN employing
XG-PON front-haul transport is introduced in Section III.
In Section IV we introduce the bandwidth allocation meth-
ods for small-cell front-haul uplink traffic (e.g., gGIANT
and RR-DBA) and also introduce optimized-RR DBA to
accommodate the burst-ness of small-cell front-haul uplink
traffic. In Section V we present the simulation results of
the performance evaluation of optimized-RR DBA against
gGIANT and RR-DBA. In Section VI we present the
application use case of optimized-RR DBA in distributed
C-RAN architecture, and finally we give our conclusions Fig. 2. Architecture of XG-PON based mobile front-haul for
in Section VII. virtualized small-cell C-RAN.
986 J. OPT. COMMUN. NETW./VOL. 9, NO. 11/NOVEMBER 2017 Mikaeil et al.

then connected to the vBBU pool (GPP servers) to enable C. Statistical Multiplexing of Mobile Uplink Traffic
front-haul data to be exchanged between vBBU-pool and in XG-PON Based Front-Haul Transport
small-cell RRHs. The OLT and vBBU are placed at the cen-
tral office (in the macro cell base station site) and connected
In an XG-PON front-haul network connecting RRHs and
to the mobile core network via a backhaul network.
vBBUs (Fig. 2), the mobile downlink traffic from the vBBU
(XG-PON downstream traffic) is broadcasted by the OLT to
B. Traffic Model for Mobile Uplink Data That all RRHs via ONUs that share the same XG-PON network
Transmits via XG-PON Front-Haul Transport medium, while mobile uplink traffic (XG-PON upstream
traffic) from RRHs is multiplexed by the OLT using the
time division multiple access (TDMA) scheme. To avoid
Virtualized small-cell mobile front-haul data traffic upstream collisions between multiple RRHs transmitting
can be modeled using the same model for backhaul traffic at the same time, XG-PON uses the DBA mechanism to
when functional split architecture (e.g., MAC-PHY split) is control upstream transmission.
considered between BBUs and RRHs because in this case
front-haul data traffic varies dynamically according to the Due to the fact that XG-PON is very common in FTTH
LTE user’s traffic similar to backhaul traffic. The Next access networks, most XG-PON-compliant DBA algorithms
Generation Mobile Networks (NGMN) alliance has devel- were designed to accommodate near-deterministic traffic
profiles that depict fixed broadband last-mile users.
oped a model to estimate backhaul traffic in small-cell LTE
However, in case the last-mile user is a mobile base station
eNodeBs [20]. In this model, small-cell base station back-
(i.e., small-cell RRH), the mobile uplink traffic transmit-
haul traffic is characterized using two different loading
ting via XG-PON front-haul in the upstream exhibits a
conditions that depend on the spatial distribution of mobile
high degree of temporal variation or burst-ness (as we
users within the cell, which exhibits a heavy-tailed pattern
mentioned earlier). Therefore, an advanced bandwidth
[21]. These two loading conditions are busy times and quiet
allocation scheme is required for efficient statistical multi-
times. During busy times, there are many users sharing the
plexing of XG-PON upstream bandwidth between RRHs,
cell’s spectral resources. Meanwhile, during quiet times,
while accommodating the burst-ness of mobile front-haul
there may often be only one user accessing the cell and uti-
uplink traffic.
lizing the entirety of the cell’s spectral resources. In a
densely deployed small-cell C-RAN environment, the num-
ber of cells is large, and the backhaul traffic results from IV. DYNAMIC BANDWIDTH ALLOCATION OF FRONT-HAUL
aggregation of a large number of cells with busy times UPLINK TRAFFIC TRANSMITTING VIA XG-PON UPSTREAM
and quiet times, and formulates on–off processes with
heavy-tailed on-times or off-times. The DBA mechanism in XG-PON is used to assign an
It has been shown in Ref. [22] that the aggregation of a upstream transmission opportunity to the ONU queues
large number of on–off processes with heavy-tailed on- (i.e., transmission containers or T-CONTs). For QoS pur-
times or off-times results in long-range dependence. The poses, XG-PON defines four different bandwidth types
existence of long-range dependence in real-world LTE associated with T-CONTs, namely, fixed (T-CONT-1), as-
and LTE-advance base station backhaul traffic has also sured (T-CONT-2), non-assured (T-CONT-3), and best-effort
been confirmed in Ref. [23]. Reference [24] shows that (T-CONT-4). The DBA operations in XG-PON start by send-
the Poisson Pareto burst process (PPBP) can be used to ing initial grants in the downstream frame from the OLT to
generate a long-range dependent arrival process that the ONU/T-CONTs, allowing them to report their buffer oc-
matches statistical properties of real-life network traffic. cupancy. After receiving the buffer occupancy reports from
Therefore, in this study we adopt PPBP as a traffic source the ONU/T-CONTs, the DBA engine selects the T-CONT to
to simulate the burst-ness of front-haul data traffic trans- be served, determines the size of grant allocation for the T-
mitting between vBBUs and RRHs (Fig. 2) in both uplink CONT, calculates the start time for each grant allocation,
and finally broadcasts these decisions to the ONUs through
and downlink. Table I shows the parameters we use to gen-
a header in the downstream frame that is updated at regu-
erate the PPBP traffic that is used in our simulations in
lar time intervals, called the DBA cycle or maximum poll-
this study.
ing interval. Different DBA mechanisms use different
methods to assign the bandwidth for T-CONTs.
In the XG-PON mobile front-haul network (Fig. 2), the
TABLE I mobile uplink traffic that transmits via XG-PON front-haul
TRAFFIC SIMULATION PARAMETERS in the upstream direction is highly bursty, and accordingly
the bandwidth demand for ONUs/T-CONTs is very unpre-
Parameters Details dictable. Thus, evaluating the performance of XG-PON-
Number of bursts 5000 standard-compliant DBAs over such bursty traffic is highly
Mean burst time length 2 ms desirable. The following subsections present the perfor-
Poisson arrival rates [95, 110, 125–200 Mbps] mance evaluation of gGIANT and RR-DBA over simulated
Hurst parameter 0.8 bursty mobile front-haul traffic. Also, they introduce opti-
Pareto shape parameter 1.4 mized-RR DBA to accommodate mobile front-haul traffic
Packet size 1470 bytes
busrt-nesss.
Mikaeil et al. VOL. 9, NO. 11/NOVEMBER 2017/J. OPT. COMMUN. NETW. 987

A. Group-Assured GIANT DBA

The gGIANT DBA algorithm is based on the well-known


giga-PON access network DBA algorithm (GIANT) [25], in
which the assigned bandwidth for T-CONT is characterized
by two service parameters, the service interval (i.e., maxi-
mum service interval) and the allocation bytes (i.e., maxi-
mum allocation bytes). The service interval (in frames)
specifies how often the T-CONT gets served, while the
allocation bytes indicate how many bytes can be assigned
to the T-CONT. To assign an upstream transmission oppor-
tunity for the T-CONT, the DBA engine (at the OLT) keeps
a counter (per-bandwidth type, e.g., assured) that decreases
every upstream frame when the counter expires; the OLT
grants T-CONT with a grant allocation equal to the maxi-
mum allocation bytes over the period of maximum service
interval, and then resets the counter to its maximum ser-
vice interval value. In the gGIANT algorithm, the scheduler
Fig. 3. Delay performance comparison of gGIANT and RR-DBA
is capable of assigning group-assured bandwidth by shar-
over bursty mobile front-haul traffic.
ing unused capacity from individual assured bandwidth
with other T-CONTs of the same group. To do this, the 
assured bandwidth type of multiple T-CONTs is grouped Gjt  Rjt if Rjt ≤ W max : (1)
into a single group-assured bandwidth type in using a W max else
container known as (gT-CONT).
The DBA engine in gGIANT keeps a byte counter for each W max or the maximum allocation bytes limit is used to
T-CONT to track the amount of bytes shared by the individ- prevent heavily loaded T-CONTs from monopolizing the
ual T-CONTs at each upstream frame. The fixed and indi- upstream frame and delaying the transmission from other
vidual assured bandwidths are assigned as per the GIANT T-CONTs.
algorithm. When assigning individual assured bandwidth, Figure 3 shows the result of the upstream delay perfor-
if a particular T-CONT does not need reserved bytes in a par- mance comparison between gGIANT and RR-DBA over
ticular upstream frame, this amount is added to the group bursty mobile front-haul traffic simulated with a PPBP
byte counter, to know how many bytes are available to the traffic source (Table I) given the simulation parameters de-
group. After assigning individual assured bandwidth, scribed in Section V (Table II). From this figure we can see
the DBA engine will go through all the groups, checking that RR-DBA performs better than gGIANT in terms of the
how many bytes are available from the previous stage. upstream delay. The reason behind such a performance
Unused bytes are then assigned to a T-CONT from the same improvement is the fact that RR-DBA assures grant allo-
group in a round-robin fashion. After finishing the assign- cation to every T-CONT during every allocation cycle. This
ment of group-assured bandwidth, non-assured and best- is in contrast to gGIANT, in which the T-CONT has to wait
effort bandwidth follow similarly to the GIANT scheduler. after sending its queue report until its service interval
timer expires, then it is granted an allocation. Also from
the same figure we can notice that both gGIANT and
B. Round-Robin DBA RR-DBA fail to satisfy the 300 μs latency required for
mobile front-haul under the given simulation parameters
RR-DBA treats all T-CONTs in the network equally by and traffic load conditions (Tables I and II).
serving each T-CONT with an amount of bytes less than
or equal to a pre-determined fixed limit (i.e., maximum
allocation bytes W max ) in a circular manner. In RR-DBA, TABLE II
the upstream transmission opportunity for ONUs or DBA SIMULATION PARAMETERS
T-CONTs (assuming each ONU has only one T-CONT) is
Parameters Details
assigned as follows:
Application traffic type PPBP
After receiving the buffer occupancy report Rjt from the
Simulation time 10 s
jth T-CONT during the t upstream allocation cycle (where j T-CONT max allocation bytes 4 Kbytes
is the index of T-CONT, i.e., j  f1; 2; …; Kg, K is the total T-CONT max service interval (gGIANT) 2 frames
number of scheduled T-CONTs during one allocation cycle), Max polling interval (all DBAs) 125 μs
the DBA engine (at the OLT) allocates the jth T-CONT with Number of RRHs (ONUs) 10
grant allocation Gjt equal to the jth T-CONT report (Rjt ) if T-CONT per ONU 1
this report is less than W max ; otherwise the grant allocation Bandwidth per T-CONT (gGIANT) 248.8 Mbps
for this T-CONT will be equal to W max . The grant allocation Propagation delay 120 μs (RTT)
ONU queue size (T-CONT buffer) 1 Mbyte
with RR-DBA is written as
988 J. OPT. COMMUN. NETW./VOL. 9, NO. 11/NOVEMBER 2017 Mikaeil et al.

C. Optimized-RR DBA Algorithm to Accommodate


Mobile Front-Haul Uplink Traffic

We chose RR-DBA as a base for our optimization, due to


the relative delay performance superiority of RR-DBA
when compared to gGIANT as we mentioned earlier.
In virtualized small-cell mobile front-haul based on XG-
PON (Fig. 2), the instantaneous upstream traffic injected
to each ONU/T-CONT exhibits a high degree of temporal
variation. In such a situation, during every upstream
frame transmission cycle, some T-CONTs might be lightly
loaded and need grant allocations less than the maximum
allocation bytes limit W max . Meanwhile, some other
T-CONTs might be heavily loaded and need grant alloca-
tions larger than the maximum allocation bytes limit W max .
When using RR-DB with a fixed maximum allocation bytes
limit (W max ,), lightly loaded T-CONT excess bandwidth
cannot be utilized, which leads to an increased delay
and inefficient utilization of the XG-PON upstream band-
width. In order to overcome such drawbacks, we propose a
new DBA approach (i.e., optimized-RR DBA) that introdu-
ces the idea of using the dynamic maximum allocation
bytes limit (W jt ) to RR-DBA.
In this new approach, we calculate the total excess band-
width amount at the end of every upstream allocation
cycle. Then, we redistribute the calculated total excess
equally among heavily loaded T-CONTs in the upcoming
allocation cycle assuming that the T-CONT traffic state
will remain unchanged for at least one allocation cycle Fig. 4. Flow diagram of optimized-RR DBA algorithm.
ahead of the allocation current cycle. The flow diagram de-
picted in Fig. 4 illustrates the procedure of the optimized- cycle buffer occupancy reports Rjt from T-CONTs, the
RR DBA algorithm. The details of the parameters used in DBA engine (at the OLT) utilizes these reports and the ini-
the algorithm are given below: tial T-CONT’s updated limit W j0 to assign grant allocation
to each individual T-CONT as follows: Gj1  minRj1 ; W j0 .
BW: Total XG-PON upstream frame size in bytes (i.e., After that, the DBA engine determines the overloading
38.880 Kb). of T-CONT by comparing its report Rjt with the initial fixed
K: Total number of ONU/T-CONTs in the network. limit W max . (Note: This fixed limit is used to determine
W jt : T-CONT’s upcoming allocation cycle maximum overloading of T-CONTs during every allocation cycle.)
allocation bytes limit (i.e., T-CONT’s updated limit). In case the report is greater than W max , the DBA engine
Rjt : T-CONT’s buffer occupancy report in current updates the counter N to be used at the end of the alloca-
cycle t. tion cycle to calculate T-CONT’s excess shares amount,
Gjt : The amount of grant allocation for T-CONT in one and also updates the Boolean variable to B⩵True (i.e.,
allocation cycle. if Rjt > W max, B⩵True, N  ).
N: A counter for the number of heavily loaded
T-CONTs in one allocation cycle. In the next step the DBA engine calculates the current
B: Boolean variable to determine T-CONT over- T-CONT updated limit for the upcoming allocation cycle
loading in one allocation cycle (i.e., T-CONT is by checking on the Boolean variable B. If B⩵True (i.e.,
heavily loaded or not). T-CONT is heavily loaded), then the updated limit for this
j: A counter for index of scheduled T-CONT during T-CONT in the upcoming allocation cycle will be equal to
the allocation cycle. W j1  W max  Exj0 . Otherwise this T-CONT is lightly
t: Allocation cycle number. loaded and the updated limit for it remains unchanged
as W j1  W max . (Note: B is rested to “B⩵False” after
Extj T-CONT’s excess shares in one allocation cycle.
scheduling this T-CONT, to be used again to determine
overloading of the next scheduled T-CONT in the current
The procedure of the optimized-RR DBA algorithm that allocation cycle.)
is illustrated in Fig. 4 can be explained as follows: After scheduling all T-CONTs in the current allocation
In the first allocation cycle, we set the initial T-CONT’s cycle (using the same above-mentioned steps) and deter-
updated limit to W j0  W max  BW∕K, and T-CONT’s ex- mining the number of heavily loaded T-CONTs during it,
cess share to Exj0  0. After receiving current allocation the DBA engine calculates the excess shares for these
Mikaeil et al. VOL. 9, NO. 11/NOVEMBER 2017/J. OPT. COMMUN. NETW. 989

heavily loaded T-CONTs at the last step of the algorithm standard deviation of the difference in end-to-end delay
P
using Exjt  BW − K j
j0 W t ∕N, and finally formulates
between every successive packet’s reach at the OLT, ignor-
the bandwidth allocation map and broadcasts it back to ing lost packets; and (3) packet loss ratio.
the T-CONTs at the ONU side. (Note: After the first allo- Figure 5 shows the overall performance comparison of
cation cycle Exjt is no longer a zero, and also the excess the three algorithms when increasing the per-RRH/ONU
share calculated in this allocation cycle will be used to com-
pute the T-CONT updated limit of the upcoming allocation
cycle, i.e., W j2  W max  Exj1 , W j3  W max  Exj2 ; …; etc:)
In succeeding allocation cycles, similar to the first allo-
cation cycle, the DBA engine will assign grant allocation to
T-CONTs, then calculate the T-CONT updated limit for the
upcoming allocation cycle using the excess shares calcu-
lated from one cycle backward, and finally calculate the
excess shares for one cycle ahead.

V. RESULTS OF PERFORMANCE EVALUATION OF OPTIMIZED-


RR AGAINST GGIANT AND RR-DBA IN VIRTUALIZED
SMALL-CELL C-RAN

To evaluate the performance of our algorithm against


gGIANT and RR-DBA, we conducted multiple simulation
experiments in the network simulator NS-3 using the
XG-PON module [26]. In our simulations, we considered
a mobile front-haul network with 10 small-cell RRHs
(20 MHz/three sectors/four antennas) connected to 10
ONUs; each ONU had only one T-CONT with assured
bandwidth type (i.e., T-CONT2). We set the ONU (i.e.,
T-CONT) buffer size to 1 Mb. We used 120 μs as the round-
trip propagation delay (RTT) to represent 10 km front-haul
distance. (Note: For 10 km the RRT is 100 μs; the other
20 μs is an additional delay margin to tolerate the latency
from ONU/OLT processing and other front-haul elements.)
In order to generate small-cell RRH uplink traffic, we
injected each ONU with PPBP traffic (parameters are
summarized in Table I). We set the maximum transfer
data-rate for the PPBP traffic source to 2.048 Gbps (when
PPBP is in the ON-state) to allow utilizing 80% of the total
XG-PON upstream capacity and leave the remaining
capacity for network protection.
For a fair comparison between the three algorithms, we
used the same simulation parameters as mentioned above
to simulate the three DBA algorithms. Meanwhile, for
gGIANT we set the parameter allocation bytes (i.e., band-
width per T-CONT) to 248.8 Mbps and the parameter
maximum service interval (i.e., T-CONT service interval)
to two frames. To perform the bandwidth assignment for
T-CONTs at the OLT side, we let all three DBAs follow
the same DBA operations given in Section IV. The sum-
mary of the parameters that we used in our simulation
is given in Table II.
In our evaluation we compare the performance of opti-
mized-RR DBA against gGIANT and RR-DBA in terms
of (1) upstream delay (i.e., end-to-end packet delay), the
value of the difference in time between the arrival time
of a certain packet to the T-CONT buffer ingress at the
ONU side and the departure time of the same packet from Fig. 5. Performance comparison of gGIANT, RR-DBA, and
the OLT buffer egress, including one-way propagation de- optimized-RR over mobile front-haul traffic: (a) upstream delay,
lay for packet transmission; (2) jitter (i.e., packet jitter), the (b) packet loss ratio, and (c) jitter.
990 J. OPT. COMMUN. NETW./VOL. 9, NO. 11/NOVEMBER 2017 Mikaeil et al.

traffic load as follows: [95, 110, 125 ∼ 200 Mbps]. As we can window size to accommodate more data and efficiently
see, optimized-RR outperforms both gGIANT and RR-DBA utilize the upstream frames in every allocation cycle (i.e.,
in all performance evaluation metrics (upstream delay, exploit the unallocated remainder of the upstream frames
jitter, and packet loss ratio). in every allocation cycle). This decreases the waiting time
From Fig. 5(a), we can see that optimized-RR performs for the heavily loaded T-CONTs and reduces the total num-
ber of frames that are needed to serve all T-CONTs, as well
better than both gGIANT and RR-DBA in terms of up-
as the total number of allocation cycles that each heavily
stream delay. Such a performance improvement is due to
loaded T-CONT needs to clear its buffer.
the efficiency of optimized-RR in utilizing XG-PON up-
stream bandwidth compared to the two other DBAs. This Also from Fig. 5(a), we can notice that both gGIANT
is clear from Fig. 6, which shows the amount of improve- and RR-DBA fail to satisfy the delay required for mobile
ment in XG-PON upstream utilization that can be added by front-haul for the 10 RRH front-haul case, as they both
optimized-RR DBA in comparison with gGIANT and RR- show upstream delay performance higher than 300 μs when
DBA—and also from Fig. 7, which shows the amount of per-ONU (RRH) load is in the range of 100 to 140 Mbps.
achieved average throughput by each DBA. In addition, Meanwhile, optimized-RR DBA attains an upstream delay
the improvement in the delay performance by optimized- performance in the range of 290–310 μs for the same range
RR DBA can be more easily explained as follows: because of per-ONU traffic loads. (Note: The 20 MHz/three sectors/
of the usage of the dynamic maximum allocation bytes limit four antennas LTE base station requires 123.2 Mbps front-
by optimized-RR DBA, T-CONTs (i.e., heavily loaded haul uplink throughput when the MAC-PYH split is consid-
T-CONTs) are allowed to increase their transmission ered as a split point between the BBU and RRH [27].)
Figure 5(b) shows the packet loss ratio performance of
the three algorithms. As we can see, optimized-RR DBA
achieves the lowest packet loss ratio compared to RR-
DBA and gGIANT. This is due to the reduction of network
congestion that can be achieved by optimized-RR DBA, by
allowing T-CONTs to use wider transmission windows to
transmit data and alleviate the packet drop rate.
Figure 5(c) shows the performance comparison of the
achieved jitter by each algorithm. From this figure we
can notice that the lowest packet jitter performance is also
achieved by optimized-RR DBA because packet jitter is
mainly impacted by the network congestion (i.e., the delay
between two successive received packets varies instead
of remaining constant when the network congestion in-
creases). As we can see, when the network load increases
and the congestion increases, packet jitter increases
accordingly.
Figure 6 shows the amount of utilization improvement
(as a percentage) that can be added by optimized-RR
Fig. 6. Amount of utilization improvement (as a percentage) that DBA at different per-ONU traffic loads when compared
can be added by optimized-RR DBA in comparison with RR-DBA with RR-DBA and gGIANT. From this figure we can notice
and with gGIANT, respectively. that the amount of utilization improvement (%) that can be

Fig. 7. Comparison of the achieved throughput by each algorithm at different per-ONU load: (a) 95 Mbps, (b) 125 Mbps, and
(c) 200 Mbps.
Mikaeil et al. VOL. 9, NO. 11/NOVEMBER 2017/J. OPT. COMMUN. NETW. 991

added when compared with the RR-DBA “black line” is


much lower than when compared with the gGIANT “blue
line” (i.e., the utilization improvement seems to be mar-
ginal with respect to RR-DBA). This is because of the bal-
anced or homogeneous traffic that we use in our simulation
(i.e., all RRHs offer the same amount of traffic load at a
given time). In other words, the utilization improvement
gain in this case is only due to the burst-ness of ONU
(RRH) traffic. Therefore, we still believe that more im-
provement gain can be achieved with optimized-RR DBA
when considering non-homogeneous traffic similar to the
practical cellular mobile network traffic (e.g., 80% of the
base stations are loaded in the range of 1–100 MB per
day while 10% of the base stations are highly loaded [more
than 100 MB per day] as in Ref. [28]).
Figure 7 shows a comparison of the achieved average
throughput per second (i.e., total number of bytes received
at OLT  8∕1 s∕1024) by each algorithm at different per-
ONU loading conditions (i.e., 95, 125, and 200 Mbps). As we Fig. 9. Upstream delay performance of optimized-RR over 10 and
20 km mobile front-haul.
can see, optimized-RR DBA achieves the highest through-
put compared to gGIANT and RR-DBA; this is due to the
capability of optimized-RR DBA to exploit the unallocated (RRHs) connected to the XG-PON system. (Note: To gener-
capacity of upstream frames during every allocation cycle. ate this plot we set the per-ONU traffic load to 123.2 Mbps
Also from this figure we can notice that the achieved to represent mobile front-haul uplink traffic for the
throughput by optimized-RR DBA is very low compared [20 MHz/three sectors/four antennas] LTE base station em-
with that achieved by RR-DBA when the per-ONU load ploying a MAC-PHY split between the vBBU and RRHs.)
is 95 Mbps [Fig. 7(a)], acceptable when the per-ONU load As we can see from the plot, around 11 RRHs can be aggre-
is 125 Mbps [Fig. 7(b)], and high when the per-ONU load is gated over XG-PON upstream capacity by optimized-RR
200 Mbps [Fig. 7(c)]. This because in the case of low and compared to nine RRHs by RR-DBA and six RRHs by
moderate per-ONU loads the amount of traffic injected to gGIANT given the latency constraint of mobile front-haul
the majority of ONU/T-CONTs is small and moderate, in C-RAN architecture (i.e., 300 μs).
respectively. Thus, the probability of T-CONTs getting In Fig. 9, we evaluate the upstream delay performance of
overloaded in such cases is much lower than in the case optimized-RR DBA over a 10 and 20 km mobile front-haul
in which the per-ONU load is high (e.g., 200 Mbps). distance. It can be seen from the upstream delay perfor-
In Fig. 8 we clarify the maximum number of RRHs that mance in this figure that even with optimized-RR DBA only
can be aggregated over the XG-PON system by each algo- 10 km front-haul distance can be supported when the per-
rithm. In order to clarify this number for each algorithm, ONU traffic load is 123.2 Mbps and the latency required for
mobile front-haul is 300 μs.
we plot the upstream delay versus the number of ONUs

VI. APPLICATION USE CASE OF OPTIMIZED-RR DBA IN


DISTRIBUTED C-RAN EMPLOYING XGS-PON FRONT-HAUL

Figure 10 depicts the network topology of distributed


C-RAN. In this architecture, two levels of front-haul

Fig. 8. Upstream delay versus the number of RRHs that can be Fig. 10. Distributed C-RAN architecture with multilevel
aggregated over XG-PON system by each algorithm. front-haul.
992 J. OPT. COMMUN. NETW./VOL. 9, NO. 11/NOVEMBER 2017 Mikaeil et al.

Fig. 11. 3GPP functional split architecture of mobile baseband


processing.

are used [29,30]: front-haul-1, which connects RRHs to


distributed radio aggregation units (RAUs), and front-
haul-2 (i.e., mid-haul [31]), which connects RAUs to the
mobile radio cloud center (RCC). Currently, the IEEE
1914 working group [32] is discussing the usage of two
Fig. 12. Upstream delay performance versus the number of
levels of functional split in 5G distributed C-RAN architec- RAUs that can be aggregated over the XGS-PON system.
ture, in which 5G baseband processing functions are
divided such that some BBU processing functions are
centralized in the radio cloud center, while others are dis- configuration given in Table III, in which the required up-
tributed in RAUs and the rest are kept in RRHs. link throughput for mobile front-haul is around 3.024 Gbps
when option 2 or 3 of the 3 GPP functional split architec-
Under the assumption that the low-MAC processing func- ture [35] is considered (i.e., the split points illustrated
tions or the processing associated with mobile scheduling with a blue dashed line in Fig. 11). (Note: We did not
reside in the RAU part (e.g., option 6 ∼ 83 GPP spilt in simulate the traffic in the front-haul-1 interface in this
Fig. 11), the latency in the front-haul-1 interface will be simulation.)
impacted by mobile scheduling latency and restricted to
250–300 μs for LTE. For the 5G network, we assume a Figure 12 shows the upstream delay performance of each
short-range front-haul-1 interface with maximum latency algorithm (i.e., RR-DBA, optimized-RR, and gGIANT) as a
budget not exceeding 100–200 μs, where a low-latency trans- function of the number of supported RAUs over XGS-PON
port solution such as CPRI (i.e., option-10 line-rate [33]) can upstream capacity. From this figure we can see that the
be utilized to meet the latency constraint for this interface. maximum number of RAUs that can be aggregated over
As for front-haul-2, which utilizes XGS-PON transport, the the XGS-PON system is three RAUs with optimized-RR
latency is mainly impacted by 5G user-plane latency, which DBA and two with RR-DBA. Meanwhile, gGIANT fails
is around 4 ms for supporting 5G enhanced mobile broad- to satisfy the delay constraint of 300 μs even with two
band services (eMBB), and around 1 ms for supporting RAUs in spite of the burst-ness of the traffic that we use
5G URLLC services (with 0.5 ms latency budget for base sta- to produce this result (PPBP traffic, Table I).
tion delay, i.e., front-haul latency [34]). Within this 0.5 ms
latency budget, assuming the latency consumed by front-
haul-1 is in the range of 100–200 μs, we are only able to VII. CONCLUSION
support URLLC services if we keep the latency of front-
haul-2 in the range of 300 μs. For this reason, we conducted In this study, we presented, to the best of our knowledge,
a simple simulation experiment to clarify the maximum the first performance evaluation of XG-PON in the context
number of RAUs that can be supported over XGS-PON of mobile front-haul. We also proposed an optimized
by each algorithm under such a latency constraint. XG-PON-compliant DBA that assures the latency recom-
To carry out this simulation, we assumed the same traf- mended in some literature for mobile front-haul in
fic parameters as described in Table I. However, we change C-RAN architecture. The simulation results of the perfor-
the parameter Poisson arrival rates (i.e., per-ONU load) to mance evaluation of our proposed optimized DBA over si-
3.024 Gbps to adhere to the 5G channel mulated dynamic data-rate front-haul traffic have shown
the performance superiority of our proposed optimized
DBA algorithm in comparison to gGIANT DBA and RR-
TABLE III DBA algorithms. The results have also clarified the maxi-
5G CHANNEL PARAMETERS mum number of LTE RRHs (i.e., 20 MHz, three sectors, and
four antennas) that can be aggregated over XG-PON front-
Parameters Details
haul in virtualized small-cell C-RAN employing MAC-PHY
Channel bandwidth 100 MHz split, and the maximum number of RAUs that can be sup-
Modulation DL, 256QAM; UL, 64QAM ported over XGS-PON based mid-haul with option-2 3GPP
Number of MIMO layers 16 layers [8(DL/UL)]
split architecture.
Mikaeil et al. VOL. 9, NO. 11/NOVEMBER 2017/J. OPT. COMMUN. NETW. 993

ACKNOWLEDGMENT bandwidth-efficiency for TDM-PON,” in Optical Fiber


Communication Conf., 2017, paper M3I–2.
This work was supported by the National Natural Science
[15] R. A. Butt, S. M. Idrus, K. N. Qureshi, N. Zulkifli, and S. H.
Foundation of China (NSFC) (61431009, 61371082, and Mohammad, “Improved dynamic bandwidth allocation
61521062) and the National Science and Technology algorithm for XGPON,” J. Opt. Commun. Netw., vol. 9,
Major Project of the Ministry of Science and Technology no. 1, pp. 87–97, 2017.
of China (2015ZX03001021). [16] T. Orphanoudakis, E. Kosmatos, J. Angelopoulos, and A.
Stavdas, “Exploiting PONs for mobile backhaul,” IEEE
REFERENCES Commun. Mag., vol. 51, no. 2, pp. S27–S34, 2013.
[17] GNU GPLv2, “ns-3.25,” Mar. 2016 [Online]. Available: https://
[1] I. Chih-Lin, J. Huang, R. Duan, C. Cui, J. X. Jiang, and L. Li, www.nsnam.org/ns‑3‑25.
“Recent progress on C-RAN centralization and cloudifica- [18] J. A. Arokkiam, X. Wu, K. N. Brown, and C. J. Sreenan,
tion,” IEEE Access, vol. 2, pp. 1030–1039, 2014. “Experimental evaluation of TCP performance over
[2] A. Pizzinat, P. Chanclou, F. Saliou, and T. Diallo, “Things you 10 Gb/s passive optical networks (XG-PON),” in Global
should know about fronthaul,” J. Lightwave Technol., vol. 33, Communications Conf. (GLOBECOM), 2014, pp. 2223–2228.
no. 5, pp. 1077–1083, 2015. [19] P. Alvarez, N. Marchetti, D. Payne, and M. Ruffini,
[3] U. Dötsch, M. Doll, H.-P. Mayer, F. Schaich, J. Segel, and “Backhauling mobile systems with XG-PON using grouped
P. Sehier, “Quantitative analysis of split base station process- assured bandwidth,” in 19th European Conf. Networks and
ing and determination of advantageous architectures for Optical Communications (NOC), 2014, pp. 91–96.
LTE,” Bell Labs Tech. J., vol. 18, no. 1, pp. 105–128, June 2013. [20] “Further study on critical C-RAN technologies,” NGMN
[4] N. J. Gomes, P. Chanclou, P. Turnbull, A. Magee, and V. Alliance white paper, Mar. 2015 [Online]. Available: https://
Jungnickel, “Fronthaul evolution: from CPRI to Ethernet,” www.ngmn.org/uploads/media/NGMN_RANEV_D2_Further_
Opt. Fiber Technol., vol. 26, pp. 50–58, 2015. Study_on_Critical_C‑RAN_Technologes_v1.0.pdf
[5] K. Miyamoto, S. Kuwano, T. Shimizu, J. Terada, and A. Otaka, [21] Y. Zhou, R. Li, Z. Zhao, X. Zhou, and H. Zhang, “On the
“Performance evaluation of Ethernet-based mobile fronthaul α-stable distribution of base stations in cellular networks,”
and wireless CoMP in split-PHY processing,” J. Opt. IEEE Commun. Lett., vol. 19, no. 10, pp. 1750–1753, 2015.
Commun. Netw., vol. 9, no. 1, pp. A46–A54, 2017. [22] X. Liu and J. S. Baras, “Aggregation of heavy-tailed on-off
[6] V. Eramo, M. Listanti, F. G. Lavacca, P. Iovanna, G. Bottari, flows is multifractal,” in 8th Int. Conf. Communication
and F. Ponzini, “Trade-off between power and bandwidth Systems(ICCS), 2002, pp. 578–582.
consumption in a reconfigurable Xhaul network architec- [23] R. K. Polaganga and Q. Liang, “Self-similarity and modeling
ture," IEEE Access, vol. 4, pp. 9053–9065, Dec. 2016. of LTE/LTE-A data traffic,” Measurement, vol. 75, pp. 218–
[7] Y. Nakayama, K. Maruta, T. Shimada, T. Yoshida, J. Terada, 229, 2015.
and A. Otaka, “Utilization comparison of small-cell accommo-
[24] D. Ammar, T. Begin, and I. Guerin-Lassous, “A new tool for
dation with PON-based mobile front-haul,” J. Opt. Commun.
generating realistic internet traffic in NS-3,” in 4th Int.
Netw., vol. 8, no. 12, pp. 919–927, 2016.
ICST Conf. Simulation Tools and Techniques (ICST), 2011.
[8] N. Yoshimoto, “Operator perspective on next-generation
[25] H. C. Leligou, C. Linardakis, K. Kanonakis, J. D.
optical access for future radio access,” in IEEE Int. Conf.
Angelopoulos, and T. Orphanoudakis, “Efficient medium arbi-
Communications Workshops (ICC), 2014, pp. 376–381.
tration of FSAN-compliant GPONs,” Int. J. Commun. Syst.,
[9] N. P. Anthapadmanabhan, A. Walid, and T. Pfeiffer, “Mobile vol. 19, no. 5, pp. 603–617, June 2006.
fronthaul over latency-optimized time division multiplexed
[26] X. Wu, K. N. Brown, C. J. Sreenan, P. Alvarez, M. Ruffini, N.
passive optical networks,” in IEEE Int. Conf. Communication
Marchetti, D. Payne, and L. Doyle, “An XG-PON module for
Workshop (ICCW), 2015, pp. 62–67.
the NS-3 network simulator,” in 6th Int. ICST Conf.
[10] N. Shibata, T. Tashiro, S. Kuwano, N. Yuki, Y. Fukada, J.
Simulation Tools and Techniques. ICST (Institute for
Terada, and A. Otaka, “Performance evaluation of mobile
Computer Sciences, Social-Informatics and Telecommunications
front-haul employing Ethernet-based TDM-PON with IQ
Engineering), 2013, pp. 195–202.
data compression [Invited],” J. Opt. Commun. Netw., vol. 7,
no. 11, pp. B16–B22, 2015. [27] “Small cell backhaul requirements,” NGMN Alliance white
paper, June 2012 [Online]. Available: https://www.ngmn.org/
[11] N. Shibata, T. Tashiro, S. Kuwano, N. Yuki, J. Terada, and
uploads/media/NGMN_Whitepaper_Small_Cell_Backhaul_
A. Otaka, “Mobile front-haul employing Ethernet-based
Requirements.pdf.
TDM-PON system for small cells,” in Optical Fiber
Communications Conf. and Exhibition (OFC), 2015, [28] U. Paul, A. P. Subramanian, M. M. Buddhikot, and S. R. Das,
pp. 1–3. “Understanding traffic dynamics in cellular data networks,”
[12] T. Tashiro, S. Kuwano, J. Terada, T. Kawamura, N. Tanaka, S. in INFOCOM, 2011, pp. 882–890.
Shigematsu, and N. Yoshimoto, “A novel DBA scheme for [29] W. Hu, L. Yi, H. He, X. Yang, Z. Li, M. Bi, K. Zhang, H. Xin, Y.
TDM-PON based mobile fronthaul,” in Optical Fiber Liu, and W. Du, “Soft-stacked PON for soft C-RAN,” J. Opt.
Communications Conf. and Exhibition (OFC), 2014, pp. 1–3. Commun. Netw., vol. 8, no. 11, pp. B12–B20, 2016.
[13] T. Kobayashi, D. Hisano, T. Shimada, J. Terada, and A. Otaka, [30] I. Chih-Lin and J. Huang, “RAN Revolution with NGFI
“Bandwidth allocation scheme based on simple statistical (xHaul) for 5G,” in Optical Fiber Communications Conf.
traffic analysis for TDM-PON based mobile fronthaul,” in and Exhibition (OFC), 2017, pp. 1–4.
Optical Fiber Communication Conf., 2016, paper W3C–7. [31] T. Pfeiffer, “Next generation mobile fronthaul and midhaul
[14] S. Hatta, N. Tanaka, and T. Sakamoto, “Feasibility architectures [Invited],” J. Opt. Commun. Netw., vol. 7,
demonstration of low latency DBA method with high no. 11, pp. B38–B45, Nov. 2015.
994 J. OPT. COMMUN. NETW./VOL. 9, NO. 11/NOVEMBER 2017 Mikaeil et al.

[32] Next Generation Fronthaul Interface 1914 Working Group [34] G. Fettweis and S. Alamouti, “5G: Personal mobile Internet
[Online]. Available: http://sites.ieee.org/sagroups‑1914/ beyond what cellular did to telephony,” IEEE Commun.
p1914‑1/. Mag., vol. 52, no. 2, pp. 140–145, 2014.
[33] “Common Public Radio Interface (CPRI); Interface [35] “Study on new radio access technology; Radio access architec-
Specification,” Oct. 2015 [Online]. Available: http://www. ture and interfaces (Release 14),” Tech. Rep. 3GPP TR 38.801-
cpri.info/downloads/CPRI_v_7_0_2015‑10‑09.pdf. 200, 2017.

You might also like