Content uploaded by Enrique Rodriguez-Colina
Author content
All content in this area was uploaded by Enrique Rodriguez-Colina on Feb 22, 2014
Content may be subject to copyright.
Telecommun Syst
DOI 10.1007/s11235-013-9764-4
A bit error rate analysis for TCP traffic over parallel free space
photonics
Enrique Rodriguez-Colina ·Diego Gil-Leyva ·
Jose L. Marzo ·Víctor M. Ramos R.
© Springer Science+Business Media New York 2013
Abstract Inter-satellite links (ISL) are a useful technology
to transmit data to space stations and to communicate be-
tween satellites. However, there are serious limitations due
to long delays and poor channel performance, resulting in
high bit error rates (BER). In this paper, parallel transmis-
sion and the scaling of the Transport Control Protocol (TCP)
window in free space optics (FSO) communications are an-
alyzed in order to overcome these disadvantages in optical
inter-satellite links. Latency and BER are the dominant ef-
fects that determine link performance. Thus, a physical, link,
network and transport cross-layer analysis for FSO over ISL
is presented in this paper. This analysis shows the advan-
tages and disadvantages of using optical parallel transmis-
sion and TCP window scaling for free space optical links
between stations and satellite constellations. The key contri-
bution of this work is to simulate the effects of the BER and
Part of this work has been done when Enrique Rodriguez-Colina was
with the Photonics Systems Team of the Engineering Department
of the University of Cambridge, and later with the Institute
of Informatics and Applications of the University of Girona.
E. Rodriguez-Colina (B)·V.M. Ramos R.
UAM-Iztapalapa, San Rafael Atlixco 186 Col. Vicentina,
Iztapalapa, 09340 México City, Mexico
e-mail: erod@xanum.uam.mx
V.M. Ramos R.
e-mail: vicman@xanum.uam.mx
D. Gil-Leyva
Sener Ingeniería y Sistemas, Severo Ochoa 4—Parque
Tecnológico de Madrid, 28760 Tres Cantos, Madrid, Spain
e-mail: dg263@cantab.net
J.L. Marzo
University of Girona, EPS Edifici P-IV (D.208), Campus
de Montilivi, 17071 Girona, Spain
e-mail: joseluis.marzo@udg.edu
to link the results to packet error rate (PER) to determine the
goodput for TCP transmissions by using a cross-layering ap-
proach. The results give evidence that wavelength division
multiplexing (WDM) can mitigate the effects of long delay
and high BER for a FSO communication using TCP.
Keywords Free space optics (FSO) ·TCP window scale
option (WSO) ·Long fat pipes ·Bit error rate (BER) ·
Packet error rate (PER)
1 Introduction
Free space optics (FSO) has advantages for inter-satellite
links (ISL) over radio frequencies (RF) because optics al-
lows high bit rates to be achieved [8]. In addition, the opti-
cal transmission performance through space as a communi-
cation medium is not affected by atmospheric factors such
as fog, dust, sand, and heat that can easily cause signifi-
cant degradation or even disruption of terrestrial FSO links.
However, factors that affect the transmission over the long
distances of ISLs are attenuation of the signal, power bud-
get and sensitivity of the receptor.
Satellite-based networking is developing quickly in or-
der to provide diverse services for space stations, satellite
constellations and for Earth data-networks. Satellite constel-
lations expand the use of limited ground-space communica-
tions [18], providing higher overall network capacity and the
use of free space optics. The development of network proto-
cols that can satisfy the high demands of ISL requires new
designs in several networking layers of the ISO/OSI Refer-
ence Model [49] and the TCP/IP model as well.
The traffic transmitted over inter-satellite links has re-
strictions due to long distances between the satellites and
as a consequence long propagation delays and poor channel
E. Rodriguez-Colina et al.
performance, resulting in high bit error rates (BER). BER
measurements are useful in order to determine the degrada-
tion of the transmission. However, BER does not clearly in-
dicate how the goodput is affected [27]—i.e., a relationship
between bit error rate and the amount of useful data success-
fully received per time unit for the application layers is not
trivial [34]. In that work, the authors find that, in this free
space optical communication scenario, a linear relationship
is useful to investigate the effects of BER on the packet error
rate (PER). This was calculated considering the worst case
scenario.
In this work we consider that the amount of data success-
fully transmitted and useful for application layers are the
goal to reach and not just to increase the bit-rate. The TCP
protocol plays an important role in retransmitting segments
of data with errors and hence we consider it in this analy-
sis. The system under investigation consists in end-to-end
optical communications between satellites varying the dis-
tance. TCP flows are used to generate traffic at 1 Gbps and
4 Gbps using the SSFNet simulator [12].1Parallel transmis-
sion using WDM is also simulated to estimate the TCP per-
formance for four optical links each at 1 Gbps rate which
generate an aggregated rate of 4 Gbps.
Other multilayer analysis has been presented in previ-
ous work but the analysis shown here is new in the context
of FSO and considering a BER/PER relationship for TCP
flows [26]. This work, to the best of the authors’ knowledge
is the first of its kind where a multidisciplinary approach is
proposed. So, this is one of the main contributions of this
paper.
A cross-layer analysis is presented in order to understand
fundamental problems with the long delays and high bit er-
ror rates incurred in the inter-satellite communications. The
use of multiple wavelengths in WDM to overcome these
limitations is also presented in this paper. In addition, the
use of the window scale option (WSO) for TCP is evaluated
by simulation.
This paper is organized as follows. In Sect. 2, an over-
view of satellite communications and the use of TCP in ISL
are presented. In Sect. 3the concept of WDM, wavelength
striping and the problems faced by TCP in long fat pipes
are described. Section 4shows the methodology used to cal-
culate the BER and the cross-layer analysis in order to cal-
culate the TCP performance. In Sect. 5, the performance of
TCP and BER using direct and coherent optical detection are
assessed through simulation. The results of the TCP perfor-
mance with 1 Gbps, 4 Gbps and 4 ×1 Gbps WDM are pre-
sented in Sect. 6. Finally, in Sect. 7, conclusions are given.
1Scalable Simulation Framework (SSF) is a discrete simulator created
by Dartmouth College, USA under a free license distribution. SSFNet
is part of the SSF simulator and it is built in the Java Language. Part
of the code for the congestion window was modified by the authors of
this paper. Please see the link at, http://www.ssfnet.org.
2 Related work
Arthur C. Clarke’s paper in Wireless World in 1945 [11],
introduced the idea of a constellation of three geostationary
satellites to provide full Equatorial coverage of the Earth,
using the geostationary earth orbit (GEO). At present, GEO
systems approximate to Clarke’s idea and bring to the world
a data communications infrastructure for both developing
countries and the more technically advanced services of de-
veloped countries. Low-earth orbiting (LEO) and medium-
earth-orbiting (MEO) satellite constellations, using orbits
lower than the geostationary orbit have been developed, as
have highly elliptical orbit (HEO) constellations. These give
full global or targeted coverage of the Earth [47]. Commu-
nication between different orbits, referred to as multi-layer
satellite network, has also been studied and in this work we
name it as multi-orbit satellite network so as to avoid confu-
sion with the ISO/OSI multi-layer analysis presented in this
paper. The multi-orbit satellite network (MOSN), which is
more complex than a single-orbit network and with longer
distances between satellites, has become an important re-
search field in satellite communications since the 1990’s.
Kimura proposed a double orbit satellite constellation made
up of LEO and MEO satellites [30]. Li at al., and Akyildiz
each designed different multi-orbit satellite network hav-
ing different spacing ISL topologies and studied the routing
problem in MOSN [2,30,32]. The long round trip times
(RTT) observed in MOSN due to the long distances be-
tween satellites are significant and require an in-depth analy-
sis [43]. Examples of distances between satellites are shown
in Fig. 1.
In [2], Akyildiz points out the communication problems
of TCP arising from the RTT and with the congestion win-
dow which represents the amount of unacknowledged data
Fig. 1 Examples of FSO links and ISL distances representing long
RTTs for TCP communications
A bit error rate analysis for TCP traffic over parallel free space photonics
that the sender can have in transit to the receiver [3,23,46].
Latency affects TCP adversely since the growth of the win-
dow size depends on the RTT between the fragments of
data packets and their acknowledgments. Other work sug-
gests several TCP variants in order to improve performance
in satellite communications [1,3,4,19].
The TCP congestion management is composed of two
important algorithms. The slow-start and congestion avoid-
ance algorithms allow TCP to increase the data transmission
rate without overwhelming the network. They use a vari-
able called the congestion window (CWND). TCP’s con-
gestion window is the size of the sliding window used by
the sender. TCP cannot inject more than CWND segments
of unacknowledged data into the network. The general be-
havior of TCP’s algorithms is referred to as Additive In-
crease Multiplicative Decrease (AIMD). For example the
TCP Reno variant increases the congestion window by one
packet per window of data acknowledged, and halves the
window for every window of data containing a packet drop
[6,20,23,24].
HighSpeed TCP for Large Congestion Windows was in-
troduced by Floyd [17] as a modification of TCP’s con-
gestion control mechanism to improve performance of TCP
connections with large congestion windows. Modified Re-
sponse Function HighSpeed TCP introduces a new relation
between the average congestion window and the steady-
state packet drop rate. The HighSpeed TCP response func-
tion maintains the property that the response function gives
a straight line on a log-log scale, this is also the response
function for regular TCP in the presence of low to moderate
congestion [13]. Work on large windows suitable for long
fat networks (LFNs, pronounced ‘elephants’) [7,25] has
been designed to overcome implementation problems with
communications that exhibit high bandwidth-delay prod-
ucts, such as ISL [1,17,38,48].
It has been observed in [31] that random losses lead to
significant throughput deterioration when the product of the
loss probability and the square of the bandwidth-delay prod-
uct is larger than one. In addition, the authors showed that
for multiple connections sharing a bottleneck link, TCP is
unfair to connections with high round-trip delays. It was
concluded that changes in both the transport and the network
layers are required to provide good end-to-end performance
over high-speed networks.
Although TCP has been carried over ISL successfully, the
long propagation delay to satellites in geostationary earth or-
bit (GEO) has limitations on data applications. In this work,
we propose the use of the window scale option for TCP in
order to transfer data efficiently when the bandwidth-delay
product is greater than 64 kbytes. Additionally, we consider
in our analysis the use of WDM to send information in par-
allel.
3 Context: parallel communications and long latency
effects
It is important to consider the effects of latency and degra-
dation in FSO-ISL communications due to the attenuation
of the signal and noise. TCP and 10 Gigabit Ethernet with
64B/66B scrambler coding are used for the analysis pre-
sented.
A single channel link with a data rate of 4 Gbps is an-
alyzed using two methods of signal detection, direct detec-
tion, for distances of less than 20,000 km, and coherent de-
tection, for distances of more than 20,000 km and no more
than 85,000 km. These results have been compared with the
calculation of a 4 wavelengths WDM link at 1 Gbps data rate
(4×1 Gbps) which corresponds toa4Gbpsaggregated data
rate. Thus, a comparative analysis of 4 Gbps and 4×1 Gbps
for TCP is presented in this paper.
3.1 Wavelength division multiplexing and wavelength
striping
The Internet Protocol (IP) over WDM is used for high ca-
pacity networks. While IP provides a common platform for
services, WDM offers a large bandwidth and throughput.
In optical communications, wavelength-division multiplex-
ing (WDM) is a technology which multiplexes several op-
tical carrier signals on a single optical fiber or in a free
space propagation medium by using different wavelengths
of laser light (i.e., coherent light) to carry different sig-
nals. WDM is equivalent to frequency division multiplexing
(FDM), which is the term often used in applications with fre-
quencies out of the light spectrum where different frequen-
cies are used to carry different signals. WDM is therefore
a technique that can reduce the impact of the high speed
inequalities, by partitioning huge bandwidth into multiple
wavelengths, each operating at slower speeds and more suit-
able with the network interface [10,15]. Although one prob-
lem with WDM is the long latency during the switching
time, it can be useful for point-to-point links and for sys-
tems with longer switching time than the propagation time.
An alternative to this is the use of wavelength-striping also
known as WDM-Bit-Per-Wavelength (BPW). This technol-
ogy reduces the latency in the switching time, sending the
information in parallel and switching in parallel. Other dif-
ferences from the WDM technique are that no parallel-to-
serial conversion is needed and that parallel pulses are laun-
ched simultaneously on different wavelengths [35]. For ex-
ample, in a microprocessor the equivalent would be to send
eight bits where each bit is transmitted in a different sin-
gle line, thus if the eight lines are observed simultaneously
the result is that a byte is transmitted in a single time-slot.
This concept of sending bits per wavelength, or fragments
of packets per wavelength per time-slot is described in this
work as “wavelength-striping”.
E. Rodriguez-Colina et al.
Fig. 2 Wavelength-striping with semi-synchronous time-slots
This paper evaluates WDM in order to compensate long
propagation delays and their use in future ISL implementa-
tions. Figure 2illustrates how information (packets) coming
from different sources are distributed and transmitted using
four wavelengths. Thus, if more wavelengths with packets
are added in the transmission, a higher aggregate through-
put will be achieved.
When wavelength-striping is used and the packets ar-
rive at the switch fabric, all the wavelengths are switched
per time-slot. This time-slot and the stripe-size are related.
This is because the number of bits allocated in the stripe
is restricted by the time-slot. When the switching is per-
formed with fixed time-slots, the stripes of data can be al-
located into the time-slots. These time-slots can be adjusted
to the size of the stripes plus a guard-band gap. The switch-
ing time in time-slotted networks can be reduced by the use
of wavelength-striping when the stripe-size is short, but this
has an adverse effect on long distance communications.
Figure 2shows stripes allocated into the wavelengths in
time-slots. For example, the stripe-size distributed over the
four wavelengths can be as short as a bit. However, the pro-
cess to insert those bits into a time-slot requires complex
electronic control to perform the high speed switching, and
this results in an impractical bit level time-slot for the up-
per layer. On the other hand, if the stripe-size is excessively
long the advantage of fast switching is compromised and
the system may not have any advantage over WDM. Con-
sequently, the optimum stripe-size depends on the control
characteristics and the requirements of the switch fabric and
applications. For ISL it is required to have stripe-sizes as big
as possible to compensate long RTT values. In addition, the
switching time required for ISL is not significant in compar-
ison with the huge propagation time and thus, in this case, it
is advantageous to use WDM instead of wavelength-striping
with short stripes. The optimal stripe size can be calculated
as a trade-off between the link capacity, i.e. the bandwidth-
delay product and the switching performance required.
This paper investigates the use of multiple-wavelengths
which gives several advantages that improve the perfor-
mance of FSO-ISL. Although wavelength-striping is recom-
mended for links with transmission-time propagation-
time (e.g. short reach interconnects), the use of TCP is
impractical when transmission-time propagation-time,
since the slow-start in the congestion window is directly af-
fected by the RTT. The TCP’s slow-start function limits the
growth of the TCP window and therefore, when WDM is
used it is possible to have a higher aggregated throughput.
High transmission frequencies and data rates have detec-
tion problems and are more sensible to noise effects. WDM
splits the data rate in different wavelengths and hence the
data rate is a fraction of the original data rate. As a conse-
quence, noise effects and detection problems are reduced.
Latency affects TCP adversely because the growth of the
window size depends of the RTT. For example, an RTT be-
tween Earth and a geostationary satellite station is in the or-
der of 500 ms, and for ISL it can vary between 100 ms and
533 ms; these long latencies of the RTT have a significant
impact in the performance of the communication, and as a
consequence, the goodput is reduced considerably.
The ISL cannot be used at its full capacity by a single
TCP connection for a single high bandwidth application, due
to small buffer spaces at the sender and receiver. This lim-
its the use of satellites for fluent internetworking. There is,
however, no limit to the number of concurrent TCP flows
that can be multiplexed together to fill the satellite chan-
nel [47]. Section 5describes the TCP window scaling-up
and the increase of the buffers at the sender and receiver and
Sect. 6shows the results of the simulation. In addition, it
has been suggested that by having multiple parallel connec-
tions, the throughput increases making better use of existing
ISL [5]. The BER as a function of distance is assessed for the
FSO model and the results are used to simulate the packet
error rate (PER) during the transmissions. These results are
used in order to obtain the goodput (i.e., the amount of use-
ful data successfully transmitted) of the 4 Gbps link and for
the 4 ×1 Gbps link for long distances. It is shown in the
next sections that WDM improves the goodput of the TCP
transmission.
4 Methodology
The system under investigation consists in end-to-end op-
tical communications between satellites. The satellite dis-
tances are different as shown in Fig. 1and hence, examples
of different distances between satellites are simulated. TCP
flows are used to generate traffic at 1 Gbps and 4 Gbps and
the communications are simulated with the SSFNet simula-
tor [12]. Parallel transmission using WDM is also simulated
in order to calculate the TCP performance for four optical
links each at 1 Gbps rate.
4.1 Comparison of single channel 4 Gbps and 4 ×1 Gbps
WDM links
A comparison of a single channel at 4 Gbps and 4 chan-
nels at 1 Gbps WDM links is shown to prove that there are
A bit error rate analysis for TCP traffic over parallel free space photonics
significant advantages by increasing the number of wave-
lengths rather than increasing the bitrate of a single channel.
This comparison is assessed with a corresponding BER for
the Inter-satellite links and for the optical detection methods
used as shown in next section.
BER as a function of data rate and distance for optical co-
herent and direct detection
At the detector electronics, the instantaneous signal to
noise ratio (SNR) is given by:
SNR =(Res ×PR)2
i2
n
(1)
with ‘Res’ being the device’s responsitivity, as obtained
in [39],
Res =qeηλ
hc (2)
where ηis the quantum efficiency of the photo detector (p-
i-n), his the Planck’s constant and cis the speed of light in
vacuum.
Then, let the power emitted by the transmitter be given
by PT. Thus, the instantaneous received optical power PR,
asshownin[42], is given by
PR=PTηRηTGRGTλ
4πz2
(3)
where,
zdistance between two satellites
ηRoptical efficiency of the receiver
ηToptical efficiency of the transmitter
GRreceiver telescope gain
GTtransmitter telescope gain.
For a case in which the aperture size Aof both telescopes
is the same, the telescope gain is:
GT=GR=πA
λ2
.(4)
The noise signal iNin Eq. (1) arises mainly from two
sources: thermal noise and shot noise. The dominant type of
noise source depends strongly on the detection method cho-
sen. For a direct detection method the thermal noise is dom-
inant and therefore, restricts the ability of the link to operate
close to the photo-detector quantum limit. The thermal noise
signal then is given by
i2
Nth =4KBT
RBe,(5)
where KBis Boltzmann’s constant, Tis the operating tem-
perature, Ris the resistor’s value and Beis the receiver elec-
trical bandwidth. This last factor is adjusted according to the
bit rate of the incoming signal. For an On-Off keying (OOK)
modulation Beis twice the channel bandwidth (BW) to pre-
vent signal distortion [39].
On one hand, for a coherent detection scheme the noise
at the detector is given primarily by the shot noise of a lo-
cal oscillator. This enables the electronics receiver to oper-
ate close to the quantum limit. In an ISL where laser power
is limited, a coherent detection is an advantage due to the
long distances involved. On the other hand, the hardware re-
quired to implement this type of detection scheme is more
complicated than direct detection [39]. Assuming that the
requirements for a coherent detection scheme can be met,
the performance of an FSO-ISL is assessed.
The shot noise at the receiver is given by
i2
Nsh =2qeResPRBe,(6)
where Beis adjusted according to the bit rate of the incom-
ing signal.
With the use of previous equations the SNR is determined
as a function of the distance z. The BER as obtained in [39]
is related with the SNR as follows,
BER =1
√2πSNRe−SNR/2.(7)
In Eq. (7), we assume that the decision threshold is zero.
This means that for zero current, the logic value is ’0’. The
BER calculations as a function of the distance between satel-
lites are plotted later in Sect. 6.
4.2 Physical layer
The case scenarios investigated reflect the advanced technol-
ogy used in ISL, and they are considered for the calculations.
The factors that system designers wish to optimize on board
a satellite are: size, weight, power and cost. These factors
are used as basic reference points; examples of equipment
characteristics used in ISL are listed below [22,41].
– Data rate 0.5–1 Gbps
– Range >10,000 km
– Estimated power consumption 55–140 W
The literature indicates a range between 0.5 and 1 watt for
the transmission power. The parameters used for the simula-
tionareshowninTable1[8,44].
4.3 Considerations for the BER/PER cross-layer analysis
The worst case scenario is when a single bit damaged or lost
in the Physical Layer causes a whole packet to be damaged
in the upper layers. In this case, TCP answers with a negative
acknowledgment (NACK) to the sender and retransmission
takes place adding more delay to the transmission. Another
E. Rodriguez-Colina et al.
Table 1 Parameters for the BER calculation
Parameter Value
Power (PT)0.9W
Wavelengths 1550 ×10−9m
Receiver (Rx) optics efficiency 0.75 %
Apertures (Tx)0.15m
Apertures (Rx)0.15m
Distance (z) 20,000–85,000 km
Quantum detector’s efficiency 0.75 %
Planck’s constant 6.26 ×10−34 Js
Speed of light (c)3×108m/s
Boltzmann’s constant (KB)1.386504 ×10−23 J/K
Operating temperature of the detector (T) 500 K
Resistor load electronics (R)50×103ω
Bandwidth of the receiver electronics (Be)2×109and
8×109Gbps
Electron charge (qe)1.6×10−19 C
Fig. 3 Worst case scenario where each 1 bit-error causes a packet-er-
ror
significant consideration is that block coding impacts the in-
teraction between the Optical Physical Layer and the trans-
ported data, and this represents a cross-layer effect since it
adds encoding bits to the Physical Layer. As a consequence,
the sequences of ones and zeros are modified.
Figure 3shows how a single damaged bit causes a packet
error; this may affect the goodput because a whole packet
loss will require a TCP retransmission. Without this condi-
tion, is not trivial to find a relationship between BER and
PER.
Taking into consideration the worst case scenario, the lin-
ear relationship between BER and packet error rate (PER) is
expressed as:
PER =8×BER ×MTU ×66/64,(8)
where the MTU is the maximum transmission unit, and us-
ing the Ethernet standards it is set to 1500 bytes for the sim-
ulations and then the MTU is increased to improve perfor-
mance. A conversion from 8 bits to 1 byte is shown, and the
mapping for the 64B/66B line coding is applied and is equal
to 1.03125.
The block coding (n, k) which is used in the standard 10
Gigabit Ethernet has been adopted for the analysis in this
work in order to establish the relationship between TCP and
BER in the free space optical links. Block codes (n, k) are
data bits grouped into data words of length n. A set of code
words with kbits is chosen where n<kand the data words
are mapped onto the code words.
Block coding is a commonly used technique because it
reduces the ambiguity of the arbitrary transitions between
binary states and also helps to maintain timing information
when long strings of ones or zeros are present. Block codes
restrict the longest time interval in which no level transitions
occur and therefore help to recover the timing or synchro-
nization. The 64B/66Bscrambling, which is a one-to-one
mapping of the data stream provides data-whitening. Such
data-whitening can remove the random distribution of the
error data, and consequently the uniformity of the data errors
is improved. Scrambling produces a random data pattern by
the modulo-2 addition of a known bit sequence to the data
stream. The resulting randomness of scrambled data pro-
vides an adequate amount of timing information [29]. In ad-
dition, the 64B/66Bcode used in Ethernet at 10 Gbps [37]
also improves the homogeneity of the incidence of data er-
rors among the packets received by the Network Layer. An-
other advantage of scrambling is that additional bandwidth
is not required because it is a binary exclusive ‘OR’ opera-
tion (XOR) with a selected sequence of bits. Then the conse-
quence of using the 64B/66Bcode is the elimination of the
non-uniformity of data-errors [33–35]. We believe that mod-
ern coding techniques, like network coding [14,16,28,45],
may be also applicable to our approach, though we leave that
study for future work in this direction.
If the scrambling code has to work with a cyclic redun-
dant check (CRC) implemented in the Link Layer, then the
selected sequence could interact in a negative way. This is
because the ability of the CRC to detect errors is reduced
whenever its polynomial generator has factors in common
with the polynomial scrambler [9]. In this paper, the Link
Layer with error detection is considered but not error cor-
rection. The influence of the lower layers in the TCP per-
formance is considered as part of the payload of a TCP seg-
ment, thus it could be generalized that the encapsulation of
the bits into the lower layers affect TCP as indicated in Fig. 2
where the worst case scenario is shown and, where each bit-
error causes a packet error, otherwise it is not feasible to
assume a suitable cross layer analysis.
The ability of the 64B/66Bcode to distribute errors uni-
formly is commented in [34]. This is a significant function
which allows errors in the Physical Layer to be related with
errors in above layers. This is because the errors in the Phys-
ical Layer are uniformly distributed, which suggests that
A bit error rate analysis for TCP traffic over parallel free space photonics
packet errors are also uniformly distributed and therefore a
connection between BER and PER can be established.
5 PER as a function of delay (distance) and analysis of
traffic performance
5.1 Scaling of the window
The window scale of the TCP sending rate is a signifi-
cant limitation of standard current implementations of TCP.
However, the results in previous work suggest that by scal-
ing the TCP sending rate, the performance improves consid-
erably [36,40].
If the window scale option (WSO) [23] is not used, the
maximum value for the slow start threshold (SST) is 65 KB.
After this SST is reached, the scaling of the window size
becomes linear, a function of the congestion-avoidance al-
gorithm [21,25]. The conventional SST for all versions of
TCP congestion window is limited to 65 KB unless a scaling
option is used. The WSO allows exponential growth to 1.07
GB for the congestion window, i.e. 65,535 ×214, where
the number 214 derives from options fields within the TCP
header which have a window scaling factor of 3 bytes. The
last byte of the 3 in the window scaling factor is the shift
count and this is set to 0 when WSO is applied.
The role of window scaling was explored by implement-
ing a customized option for the SSFNet simulator [12], an
option not previously available; this new scaling implemen-
tation was developed in previous work and it will be referred
to in this paper as the “scaling up (Sup)” model [40].
The ‘Sup’ model considered the maximum congestion
window available (i.e. 1.07 Gbytes) in TCP for sending and
receiving rates. ‘Sup’ directly modifies the limit of conges-
tion window exponential growth without changing the TCP
header. Thus it differs from the original idea of WSO be-
cause the scaling is not announced between the receiver and
the server during the synchronization of the communication,
i.e. when the transmission is initiated. By modifying the
Congestion Window threshold and the Advertised Window
in the ‘Sup’ model, scaling is available for all transmissions
without being announced.
Even if the WSO and the proposed ‘scaling up’ are differ-
ent in the initial period of the communication (i.e., TCP syn-
chronization), the scaling of the congestion window is sim-
ilar during the slow-start growth. The exponential growth
is controlled by the value of the slow-start threshold (SST).
In the simulation model, the SST is expanded allowing the
window to grow always to 1 GB without modifying the TCP
header.
Figure 4shows the difference between scaling up a con-
ventional growth of the TCP window. The growth of the
window in a TCP connection is as follows:
Fig. 4 Congestion window vs time
if (cwnd < SSTS)
/*this opens the window
exponentially */
SSTS=1GB
else
/*do Congestion Avoidance increment
by 1 Laser channel is dedicated
and big enough so no congestion
avoidance is required in this
scenario */
cwnd += 1/cwnd;
5.2 TCP traffic scenarios
We use the SSFNet simulator to determine the TCP
throughput on the basis of the predicted PER which is re-
lated to the BER using Eq. (8). Several TCP concurrent
flows are generated and transmitted over the free space op-
tics both with a single link and with four wavelength link
between two satellites. The aim is to generate a consider-
able traffic flow; the sending of concurrent flows of data
provides full-pipe utilization for both 4 Gbps rate link ex-
amined. The transmission simulations are performed with
1 Gbps rate for each of the four wavelengths and therefore,
ideally, the aggregated transmission rate is 4 Gbps.
The goodput of the TCP connections is affected by the
link delay as shown in Fig. 5. Even though the data rate is
4 Gbps, the 1 Gbps link results in nearly the same goodput.
This is because the slow-start growth of the TCP window in-
creases as a function of the number of acknowledgments re-
ceived and the number of segments. These segments are set
to the maximum segment size (MSS) of 1500 bytes which
is a standard for Ethernet and TCP, and then the MSS is
doubled increased up to 12000 bytes which is close to the
simulation limit. This simulation shows when the propaga-
tion time is greater than the transmission time. Therefore,
the dominant delay of the transmission is due to the signal
transmitted over the distance at speed of light.
E. Rodriguez-Colina et al.
Fig. 5 Time vs. distance for
1 Gbps and 4 Gbps where the
transmission time
propagation time
Table 2 ISL distances and corresponding one-way propagation delays
Distance (km) Corresponding one way delay (sec)
16,000 0.0533
25,000 0.0833
50,000 0.1667
80,000 0.2667
The width of the bars in Fig. 5represents the transmission
delay and the slope of the diagonals represents the propaga-
tion delay (for simplicity, the slopes of the diagonals are not
proportional to the width of the bars).
Four case scenarios have been chosen for the analysis of
the delay of the ISL. These delays were assessed with the
distances between satellites as follows in Table 2.
6 Results
6.1 BER as a function of delay for FSO-ISL
We assume a fixed transmission power of 0.9 watts to calcu-
late BER as a function of delay for FSO-ISL. The parame-
ters used for the analysis are as shown previously in Table 1.
From Eq. (8), we find that direct detection results in a higher
BER than coherent detection. However, coherent detection
requires more complex hardware to be implemented. There-
fore, the use of direct detection is suggested for distances
below 20,000 km and coherent detection for distances above
20,000 km.
The BER/PER relationship described above is used to
perform the TCP traffic flow simulations. The BER results
in a PER that greatly increases the ISL with distances above
Fig. 6 BER vs distance—direct and coherent detection for 1 Gbps and
4 Gbps
20,000 km, as shown in Fig. 6. Thus the TCP simulations
of distances above 50,000 km include the BER effect. The
BER results of 1 Gbps and 4 Gbps are shown in Fig. 6.
With the use of coherent detection it is possible to trans-
mit longer distances than using direct detection. In both
cases, with direct and coherent detection the BER is higher
for transmissions at 4 Gbps. Thus, with the use of low data
rates is possible to keep a lower BER and for this reason it
is suggested that the data rate should be split into several
wavelengths.
These results have been assessed for delays correspond-
ing to the distances between satellites as shown previously
in Table 2. We consider PER for TCP from the cross-layer
analysis as described in Sect. 4.3 in this work, and we assess
it using Eq. (8).
A bit error rate analysis for TCP traffic over parallel free space photonics
Table 3 Aggregated goodput
with and without wavelength
striping and with two detection
methods
Delay
(sec)
Distance
(km)
PER Detection
method
Data rate
(Gbps)
Goodput
(Mbps)
Wavelength
stripped
Aggregated
goodput
0.0533 16,000 1.E−12 direct 1 5.769 yes 23.076
0.0533 16,000 1.E−12 direct 4 5.772 no 5.772
0.0833 25,000 1.E−08 direct 1 3.692 yes 14.768
0.0833 25,000 1.E−08 direct 4 3.693 no 3.428
0.1667 50,000 1.E−04 coherent 1 1.845 yes 1.845
0.1667 50,000 1.E−04 coherent 4 1.846 no 7.384
0.2667 80,000 1.E−03 coherent 1 0.291 yes 0.291
0.2667 80,000 1.E−03 coherent 4 0.294 no 1.176
Fig. 7 Goodput vs. line delay for 1 Gbps, varying the maximum seg-
ment size (MSS) and with and without scaling up of the TCP Window
(WSO)
6.2 Scaling-up of the window and maximum segment size
increase
We calculate the goodput for several serial links by varying
the distances of each link with the delays shown in Table 2.
A BER of 10−9is used for the calculations, and this corre-
sponds to the distances shown in Fig. 6for the direct and
coherent detection methods previously described. In addi-
tion, the maximum segment size (MSS) is increased to com-
pare the performance of the link transmission. The increased
MTU improves performance if the MSS is also increased as
shown in the simulation results. The goodput with and with-
out the scaling-up (Sup) of the TCP window size is simu-
lated and the results are shown in Fig. 7.
The goodput of the TCP connections is affected by the
link delay as shown in Fig. 8and Table 3. The WDM link
transmits the packet faster. This is because the slow-start
growth of the TCP window increases as a function of the
number of acknowledgments received. Therefore, the delay
in the RTT is the key factor to determine performance when
1 Gbps and 4 Gbps are compared.
Fig. 8 Goodput vs. distance for 1 Gbps, 4 Gbps single link and for
4×1 Gbps WDM
Because of the four fold increase in the aggregated data
rate, important implications can be deduced from the results.
The goodput decreases with distance but the four fold differ-
ence between 4 Gbps single link and the 4 Gbps wavelength-
striped link is maintained even with the increase in dis-
tance. Thus, the increment in the number of wavelengths
increases proportionally the aggregated throughput and it
helps to compensate the diminished performance of TCP
due to long delays. It is interesting to observe that the good-
put for 1 Gbps and 4 Gbps single links are practically the
same due to the propagation delay dominance.
Table 3shows that the goodput for 1 Gbps and 4 Gbps are
indeed the same due to the propagation delay dominance.
Thus a trade-off between cost to include more links and per-
formance should be considered.
7 Conclusions
We described in this paper how WDM can increase the
goodput of the transmission. A cross-layer analysis is estab-
lished using a relationship between BER and PER for TCP
flows for 64B/66B encoding. This relationship is not trivial
E. Rodriguez-Colina et al.
for most of the application case scenarios, here the analy-
sis of a worst case scenario is shown. So, our findings show
limits on this research direction.
This work provides strong multidisciplinary foundations
and in-depth analysis to show significant characteristics that
can improve the performance of FSO over ISL. The TCP
window is expanded with the use of a customized simulation
of the window scale option and the results show that for a
long propagation delay it is useful to scale the window size.
Simulations of the direct and coherent methods for optical
detection are assessed and the corresponding BER for the
Inter-satellite links are demonstrated.
The analysis shown in this paper suggests that WDM is
a promising solution to overcome latency related problems
in FSO over ISL. The advantages of increasing the number
of wavelengths rather than increasing the transmission data
rate are shown when the propagation delay is greater than
the transmission delay.
We believe that network coding may be applied to our ap-
proach with good results. Recent work has shown that using
such techniques for satellite systems is a promising solution.
So, our future work will look further on that direction.
References
1. Akan, O. B., Fang, J., & Akyildiz, I. F. (2002). Performance
of TCP protocols in deep space communication networks. IEEE
Communications Letters,6(11), 478–480.
2. Akyildiz, I. F., & Jeong, S. H. (1997). Satellite ATM networks:
asurvey.IEEE Communications Magazine,35(7), 30–43.
3. Akyildiz, I. F., Morabito, G., & Palazzo, S. (2001). TCP-Peach:
a new congestion control scheme for satellite IP networks.
IEEE/ACM Transactions on Networking,9(3), 307–321.
4. Akyildiz, I. F., Zhang, X., & Fang, J. (2002). TCP-Peach+: en-
hancement of TCP-Peach for satellite IP networks. IEEE Commu-
nications Letters,6(7), 303–305.
5. Allman, M., Kruse, H., & Ostermann, S. (1996). An application
level solution to TCP’s satellite inefficiencies. In International
workshop on satellite-based information services (WOSBIS) (pp.
100–107).
6. Allman, M., Paxson, V., & Stevens, W. (1999). RFC 2581: TCP
congestion control. http://www.ietf.org/rfc/rfc2581.txt.
7. Allman, M., Dawkins, S., Glover, D., Griner, J., Tran, D., Hender-
son, T., Heidemann, J., Touch, J., Kruse, H., Ostermann, S., Scott,
K., & Semke, J. (2000). RFC 2760: Ongoing TCP research related
to satellites. http://www.ietf.org/rfc/rfc2760.txt.
8. Begley, D. L. (2000). Laser cross-link systems and technology
free space laser communications. IEEE Communications Maga-
zine,38(8), 126–132.
9. Boudreau, P. E., Bergman, W. C., & Irvin, D. R. (1994). Perfor-
mance of a cyclic redundancy check and its interaction with a
data scrambler. IBM Journal of Research and Development,38(6),
651–658.
10. Brackett, C. A. (1990). Dense wavelength division multiplexing
networks: principles and applications. IEEE Journal on Selected
Areas in Communications,8(6), 948–964.
11. Clarke, A. C. (1945). Extra-terrestrial relays. Wireless World, 305–
308.
12. College, D. (2007). SSFNet: Scalable Simulation Framework
(SSFNET). http://www.ssfnet.org.
13. De-Souza, E., & Agarwal, D. (2003). A high speed TCP study:
characteristics and deployment issues (Tech. rep.). Lawrence
Berkeley National Laboratory, technical Report LBNL-53215.
14. Dong, W., Wang, J., Huang, M., Tang, J., & Zhou, H. (2011).
ATCS: an adaptive TCP coding scheme for satellite IP networks.
KSII Transactions on Internet and Information Systems,5(5),
1013–1027.
15. Dowd, P. W., & Sivalingam, K. M. (1994). A multi-level WDM
access protocol for an optically interconnected parallel computer.
In IEEE international conference on computer communications
(INFOCOM) (pp. 400–408).
16. Fausto, F. V., Shintre, S., & Barros, J. (2010). How feasible is net-
work coding in current satellite systems? In Proceedings of the 5th
advanced satellite multimedia systems conference and the 11th
signal processing for space communications workshop (pp. 31–
37).
17. Floyd, S. (2003). RFC 3649: High Speed TCP for large-scale con-
gestion windows. http://www.ietf.org/rfc/rfc3649.txt.
18. Gavish, B. (1997). LEO/MEO systems—global mobile communi-
cation systems. Telecommunications Systems,8(2), 99–141.
19. Giambene, G., & Miorandi, D. (2005). A simulation study of scal-
able TCP and highspeed TCP in geostationary satellite networks.
Telecommunications Systems,30(4), 297–320.
20. Hasegawa, G., Murata, M., & Miyahara, H. (2000). Fairness and
stability of congestion control mechanisms of TCP. Telecommuni-
cations Systems,15(1), 167–184.
21. Hellal, O. A., & Altman, E. (2000). Analysis of TCP vegas and
TCP reno. Telecommunications Systems,15(3), 381–404.
22. Hemmati, H. (2008). Free-space optical communications at
JPL/NASA (Tech. rep.). Jet Propulsion Laboratory, California In-
stitute of Technology, optical Communications Group.
23. Jacobson, V. (1988). Congestion avoidance and control. Computer
Communication Review,18(4), 314–329.
24. Jacobson, V. (1990). Modified TCP congestion control and avoid-
ance algorithms. End2end-interest mailing list.
25. Jacobson, V., Braden, R., & Borman, D. (1992). RFC 1323:
TCP extensions for high performance. http://www.ietf.org/rfc/
rfc1323.txt.
26. James, L., Moore, A. W., Wonfor, A., Plumb, R., White, I.,
& Penty, R. (2005). Packet error rate and bit error rate non-
deterministic relationship in optical network applications. In Op-
tical fiber communication conference, 2005. Technical digest.
OFC/NFOEC (Vol.4,p.3).
27. James, L. B. (2005). Error behaviour in optical networks.PhD
thesis, University of Cambridge.
28. Katti, S., Rahul, H., Hu, W., Katabi, D., Muriel, M. M., &
Crowcroft, J. (2006). XORs in the air: practical wireless network
coding. Computer Communication Review,36(4), 243–254.
29. Keiser, G. (2000). Optical fiber communications (3rd ed.). New
York: McGraw Hill.
30. Kimura, K., Inagaki, K., & Karasawa, Y. (1997). Double-layered
inclined orbit constellation for advanced satellite communication
network. IEICE Transactions on Communications,E80-B(1), 93–
102.
31. Lakshman, T. V., & Madhow, U. (1997). The performance of
TCP/IP for networks with high bandwidth-delay products and ran-
dom loss. IEEE/ACM Transactions on Networking,5(3), 336–
350.
32. Li, H., Zhang, Q., Zhang, N., Zhang, Y., & Xu, H. (2006). Adap-
tive routing strategy in multi-layer satellite communication net-
works. In Proceedings of the 7th international symposium on an-
tennas propagation & EM theory (ISAPE) (pp. 1–4).
33. Manchester, J., Anderson, J., Doshi, B., & Dravida, S. (1998). IP
over SONET. IEEE Communications Magazine,36(5), 136–142.
34. Moore, A. W., James, L. B., Glick, M., Wonfor, A., Plumb, R.
G., White, I. H., McAuley, D., & Penty, R. V. (2005). Optical
network packet error rate due to physical layer coding. Journal of
Lightwave Technology,23(10), 3056–3065.
A bit error rate analysis for TCP traffic over parallel free space photonics
35. Moore, A. W., James, L. B., Wonfor, A., White, I. H., Penty, R. V.,
Glick, M., & McAuley, D. (2005). Chasing errors through the net-
work stack: a testbed for investigating errors in real traffic on op-
tical networks. IEEE Communications Magazine,43(8), s34–s39.
36. Oueslati-Boulahia, S., Serhrouchni, A., Tohmé, S., Baier, S.,
& Berrada, M. (2000). TCP over satellite links: problems and
solutions. Telecommunications Systems,13(2), 199–212.
37. Raahemi, B. (2005). Error correction on 64/66 bit encoded
links. In IEEE Canadian conference on electrical and computer
engineering (pp. 412–416).
38. Rahmani, R., Kanter, T., & Åhlund, C. (2010). A self configuring
fuzzy active queue management controller in heterogeneous net-
works. In 17th international conference on telecommunications
(pp. 634–641).
39. Ramaswami, R., & Sirvarajan, K. N. (2001). Morgan Kaufmann
series in networking.Optical networks: a practical perspective
(2nd ed.). San Mateo: Morgan Kaufman.
40. Rodríguez-Colina, E., James, L. B., Penty, R. V., White, I. H.,
Williams, KA, & WMoore, A. (2006). TCP sending rate control
at terabits per second. In Proceedings of the 25th IEEE inter-
national conference on computer communications (INFOCOM),
Barcelona, Spain (pp. 1–5).
41. Satcom (2006). Optical communications and intersatellite links.
WTEC Hyper-Librarian.
42. Shlomi, A., Rotman, S. R., & Kopeika, N. S. (1998). Performance
limitations of a free-space optical communication satellite net-
work owing to vibrations: heterodyne detection. Applied Optics,
37(27), 6366–6374.
43. Taleb, T., Mashimo, D., Jamalipour, A., Kato, N., & Nemoto,
Y. (2009). Explicit load balancing technique for NGEO satellite
IP networks with on-board processing capabilities. IEEE/ACM
Transactions on Networking,17(1), 281–293.
44. Tirró, S. (1993). Satellite communication systems design.New
York: Plenum Press.
45. Vieira, F., & Barros, J. (2009). Network coding multicast in
satellite networks. In Proceedings of the 5th Euro-NGI conference
on next generation Internet networks, Piscataway, NJ, USA (pp.
18–23). New York: IEEE Press.
46. Wong, E., & Yum, T. (1997). Delay bounds for packet satellite
protocols. Telecommunications Systems,8(2), 277–291.
47. Wood, L. (2001). Internetworking with satellite constellations.
PhD thesis, School of Electronics, Computing and Mathematics,
University of Surrey, Centre for Communications Systems
Research, United Kingdom.
48. Zabir, S. M. S., & Kitagata, G. (2004). A proposal for efficient
TCP flow control over satellite networks. Telecommunications
Systems,25(3), 371–400.
49. Zimmermann, H. (1980). OSI reference model—the ISO model of
architecture for open systems interconnection. IEEE Transactions
on Communications,28(4), 425–432.
Enrique Rodriguez-Colina gained
his Ph.D. Degree in Photonics Com-
munications Systems in the Engi-
neering Department of the Univer-
sity of Cambridge, England. He
was research associate in the Broad-
band Communications and Dis-
tributed Systems (BCDS) Group at
the University of Girona, Spain in
2009. In 2003 he obtained his Mas-
ter Degree in Computer Science
from the Autonomous Metropolitan
University Azcapotzalco (UAM-A).
He received a Diploma Degree in
Management Skills from the Au-
tonomous Technology Institute of Mexico (ITAM) in 2001. In 1994 he
was awarded the B.Eng. in Electronics and Communications from the
Autonomous Metropolitan University, Iztapalapa (UAM-I). With more
than 15 years of experience in the industry, he worked as an engineer,
project manager and consultant for telecommunications companies.
His current research interests are in the area of high speed communi-
cation systems and optical networks; and protocols of communication
for high capacity networks including - inter-satellite links and cog-
nitive radio networks -. He is author of publications in international
conferences, seminars and journals and he is also a member of the
IEEE, Computer and Communications societies.
Enrique Rodriguez-Colina is Visitor Researcher and Lecturer at the
Electrical Engineering Department with the Networking and Telecom-
munications research team of UAM-I since August 2010.
Diego Gil-Leyva received the first
degree in Physics from the Insti-
tuto Tecnológico y de Estudios Su-
periores de Monterrey (ITESM),
Monterrey, Mexico, in 2000, fol-
lowed by the M.S. and Ph.D. de-
grees from Cambridge University,
Cambridge, UK, in 2002 and 2006,
respectively. His research interests
include free space optical inter-
connects, computer generated holo-
grams, and adaptive optics. Diego
Gil-Leyva was an optics research
and development engineer at Light
Blue Optics in England when this
work was done. Currently, he holds the position of optics consultant at
Sener Ingeniería y Sistemas in Madrid, Spain.
Jose L. Marzo is Professor at the
Computer Architecture and Tech-
nology Department at the Univer-
sity of Girona, Spain. From 1978
to 1991 he was with Telefonica.
In Telefonica, Dr. Jose L Marzo
had different responsibilities at the
province of Girona such as head
of the engineering department and
head of the planning and program-
ming office. His research interests
are in the fields of communication
networks, networking, optical net-
works control and management, ra-
dio cognitive networks (media ac-
cess) adaptive. Prof. Marzo leads a research group on broadband com-
munications and distributed systems (BCDS) at Informatics and Ap-
plications Institute (at the University of Girona). He coordinated the
participation of BCDS in national Spanish research projects. Prof. Jose
L Marzo is a member of the IEEE Communications Society. He has
participated to the technical program committees and chairing ses-
sions of several conferences, including SPECTS, IEEE Globecom, ICC
and Infocom. He serves in the editorial board of International Journal
of Communications Systems. He has co-authored several papers pub-
lished in international journals and presented in leading international
conferences.
E. Rodriguez-Colina et al.
Víctor M. Ramos R. is researcher
and lecturer at the Department of
Electrical Engineering with the Net-
working and Telecommunications
research team of Universidad
Autónoma Metropolitana-Iztapalapa
at Mexico City. He received in 2004
the Ph.D. degree in Computer Sci-
ence from the Université de Nice-
Sophia Antipolis, France working
within the Planète Research Team
of INRIA. In 2000, he received the
DEA Degree (the French equiva-
lent of the M.S. American degree)
in Networking and Distributed Sys-
tems from the same university. He got in 1995 the B.Eng. in Elec-
tronics and Communications Engineering degree from the Universidad
Autónoma Metropolitana, Mexico.
Prof. Víctor Ramos has been the Head of the Electrical Engineering
Department of UAM-Iztapalapa from 2007-2011, and from 2006-2007
the Head of the Networking and Telecommunications Research Team
from the same University. His research interests fall in performance
evaluation of communications protocols, multimedia applications and
routing protocols for wireless networks.
