Performance comparison of dynamic elastic optical networks with optical regeneration

h and first fit algorithm (called First Fit) [21]
and the spectrum-least RSA algorithm which has been
developed in [22] (denoted as Least Spectrum).
Moreover, two typical network topologies, that are (i)
National Science Foundation network (NSF) consisting of
14 nodes and 22 links, and (ii) US backbone network
(USNET) including 24 nodes and 43 links (shown in Fig-
ure 6) are used for numerical experiments. The regeneration
capable node number is assumed to be limited at 4 and 7
for NSF and USNET respectively and each case will be
tested with 20 random scenarios. We also use following
parameters for the numerical simulation. Each fiber link
can carry up to W spectrum slots (W is fixed at 128)
and the slot bandwidth is assumed to be 12.5 GHz. The
networks can flexibly and dynamically set up and release
optical paths (also called lightpaths). Modulation format
of each lightpath is distance-adaptively assigned. Lightpath
requests arrive sequentially and follow Poisson distribution.
Average arrival rate of ligthpaths is λ (requests per time
unit). Distribution of lightpath holding time is assumed to
be a negative exponential one with the mean hold time of
1/µ (time units). Consequently, the given network traffic
load in Erlangs is λ/µ. Here, the capacity, C, of each
requested lightpath between node pairs is also randomly
assigned between 50 and 100 Gbps following a uniform
distribution. The assumed modulation formats are BPSK,
QPSK, 8-QAM and 16-QAM. The slot bandwidth and the
corresponding transparent reach of BPSK, QPSK, 8-QAM,
and 16-QAM optical signals are given in Table II [21, 22].
As mentioned above, we evaluate the network perfor-
mance in three comparative regeneration scenarios of elastic
optical networks: (i) without regeneration capability (named
47
Research and Development on Information and Communication Technology
(a) No regeneration scenario
(b) 3R regeneration scenario
(c) 4R regeneration scenario
Fig. 5. Regeneration network scenario comparison.
(a) NSF
(b) USNET
Fig. 6. Experimental network topologies (NSF and USNET).
No regeneration), (ii) 3R regeneration with two different
RSA algorithms including First-fit and Least-spectrum al-
gorithms (so called First fit w/ 3R and Least spectrum w/
3R correspondingly), and (iii) 4R regeneration also with
TABLE II
SUMMARY OF KEY PARAMETERS
Parameter Value
Spectrum slot number per link 128
Slot bandwidth 12.5 GHz
M
od
ul
at
io
n
fo
rm
at
BPSK
Slot capacity 12.5 Gbps
Transparent reach 9600 km
QPSK
Slot capacity 25 Gbps
Transparent reach 4800 km
8-QAM
Slot capacity 37.5 Gbps
Transparent reach 2400 km
16-QAM
Slot capacity 50 Gbps
Transparent reach 1200 km
Capacity of requested lightpaths 50-100 Gbps
Connection mean hold time 1000
First-fit and Least-spectrum RSA algorithms (denoted by
First fit w/ 4R and Least spectrum w/ 4R respectively).
The performance of the comparative network scenarios, in
terms of the blocking probability, is shown severally in
Figures 7 and 8 for NSF and USNET network topologies
when the traffic load ranges from 100 to 1500 Erlangs.
These figures verify that the network scenarios using optical
regeneration, i.e. 3R or 4R regeneration, can dramatically
decrease the blocking probability in comparison with that
without regeneration. That is because implementing optical
regeneration can help to resolve the spectrum collision,
save the spectrum resources with higher modulation formats
and enhance the optical reach to improve the network
utilization. Moreover, the results also show that applying
48
Vol. 2019, No. 1, September
Fig. 7. Blocking probability for NSF network.
Fig. 8. Blocking probability for USNET network.
4R regeneration offers better performance than that of 3R
regeneration. Such performance improvement is obtained
by re-modulating the optical signal into the optimal mod-
ulation format for exploiting the distance-adaptive feature
of elastic optical networks.
Furthermore, as being verified in Figures 7 and 8, the
performance of the networks also strongly depends on the
applied RMSA algorithm due to the effect of selecting the
regeneration nodes and the modulation format. Optimizing
the regenerating node position and assigning suitable modu-
lation formats play an important role to enhance the overall
spectrum utilization efficiency. It is confirmed that the effect
of optical regeneration is enhanced with larger network.
The main reason is that high-order modulation levels can
be assigned to shorter lightpaths and consequently, help
to lessen the number of spectrum slots required. On the
other hand, larger network which contains longer average
length of lightpaths requires more regeneration resources,
especially for the lightpaths with high-order modulation
format due to the transparent reach limitation.
Moreover, Figure 9 demonstrates the comparison of
relative accepted traffic volumes among the experimented
scenarios, that are No regeneration, First fit w/3R, First
fit w/ 4R, Least spectrum w/ 3R and 4R, with the given
blocking probability of 10−5. The results obtained by No
Fig. 9. Accommodated traffic volume comparison.
Fig. 10. Average number of regenerating nodes per lightpath for NSF
network.
regeneration is used as the benchmark, so its graph is 1.
The figure describes that, with the same network condi-
tions, using optical regeneration provides higher accepted
traffic volume thanks to the signal regeneration. We can
attain at least 29.8% (45.6%) higher traffic volume with
NSF (USNET) network. Implementing more-efficient RSA
algorithm (i.e. spectrum least) even can further improve
the network performance. However, note that it may results
in higher network CAPEX. In fact, we must consider the
cost of regeneration devices carefully to satisfy the tradeoff
between the CAPEX and the network performance.
We have also estimated the average number of regenerat-
ing nodes and that of spectrum slots per lightpath in the ex-
perimental network topologies to determine the requirement
of necessary regeneration resources. Figures 10 and 11
illustrate the obtained numerical results for NSF topology
and Figures 12 and 13 show that of USNET respectively.
Because of the use of same regenerating node selection
strategy, both First fist w/ 3R and First fit w/ 4R scenarios
need almost the same average number of regenerating nodes
for each lightpath. However, the number of regenerated
spectrum slots is decreased with the use of 4R regeneration
thanks to modulate the signals distance-adaptively. The
graphs also verify that, with the Least spectrum algorithm,
the improvement of the network performance costs more
regenerating resources.
49
Research and Development on Information and Communication Technology
Fig. 11. Average number of regenerated spectrum slots per lightpath
for NSF network.
Fig. 12. Average number of regenerating nodes per lightpath for
USNET network.
Finally, Figures 14 and 15 illustrate the modulation
efficiency, which is the ratio of the average spectrum slot
number per established lightpath of each compared scenario
to that of No regeneration one, of both NSF and USNET.
The spectrum slot number of No regeneration scenario is
used as a benchmark and consequently, its normalized value
is one. It demonstrates that although First fit strategy offers
better modulation efficiency (less ratio) for established
lightpaths, it does not consider the spectrum collision an
as a result, its performance is poorer than that of Least
spectrum strategy.
IV. CONCLUSION
In this paper, we have investigated the impact of various
optical regeneration techniques including 3R and 4R on
the performance of dynamic elastic optical networks. We
firstly discussed and clarified the differences among optical
regenerating usage scenarios without regeneration and with
regeneration (both 3R and 4R regeneration). We have, then,
evaluated and compared the dynamic elastic optical network
performance, in term of the blocking probability and the
total accepted traffic volume, etc. for three comparative
network scenarios including (i) no regeneration, (ii) with
3R regeneration and (iii) with 4R regeneration capabilities.
Fig. 13. Average number of regenerated spectrum slots per lightpath
for USNET network.
Fig. 14. Modulation efficiency of NSF network.
Numerical simulation results prove that using optical
regenerators, especially 4R regenerators, can help to ex-
ploit elastic optical networking and enhance the dynamic
network performance for provisioning bandwidth-flexible
lightpath services. It was demonstrated that at least 29.8%
(or 45.6%) more traffic volume can be attained for NSF
(USNET) network. However, optical regenerating devices
are expensive and hence, the trade-off between the required
network performance and the necessary regenerating re-
source cost should be considered carefully in order to cre-
ate future cost-effective, spectrum-efficient and bandwidth-
flexible optical networks.
ACKNOWLEDGMENT
This research is funded by Vietnam National Foundation
for Science and Technology Development (NAFOSTED)
under grant number 102.02-2015.39.
REFERENCES
[1] Cisco, “Cisco visual networking index: Forecast
and methodology, 2014–2019.” [Online]. Available:
provider/ip-ngn-ip-next-generation-network/white paper
c11-481360.pdf
[2] K. Sato and H. Hasegawa, “Optical networking technologies
that will create future bandwidth-abundant networks,” J.
Optical Com. Net., vol. 1, pp. A81–A93, 2009.
50
Vol. 2019, No. 1, September
Fig. 15. Modulation efficiency of USNET network.
[3] G. Zhang, M. De Leenheer, A. Morea, and B. Mukherjee,
“A survey on OFDM-based elastic core optical networking,”
IEEE Com. Surveys & Tutorials, vol. 15, pp. 65–87, 2012.
[4] A. Jukan and J. Mambretti, “Evolution of optical networking
toward rich digital media services,” Proceedings of the IEEE,
vol. 100, no. 4, pp. 855–871, 2012.
[5] M. Jinno, H. Takara, B. Kozicki, Y. Tsukishima, Y. Sone,
and S. Matsuoka, “Spectrum-efficient and scalable elastic
optical path network: architecture, benefits, and enabling
technologies,” IEEE Com. Mag., vol. 47, pp. 66–73, 2009.
[6] O. Gerstel, M. Jinno, A. Lord, and S. B. Yoo, “Elastic optical
networking: A new dawn for the optical layer?” IEEE Com.
Mag., vol. 50, 2012.
[7] B. C. Chatterjee, N. Sarma, and E. Oki, “Routing and
spectrum allocation in elastic optical networks: A tutorial,”
IEEE Communications Surveys & Tutorials, vol. 17, no. 3,
pp. 1776–1800, 2015.
[8] A. Lord, P. Wright, and A. Mitra, “Core networks in the
flexgrid era,” Journal of Lightwave Technology, vol. 33,
no. 5, pp. 1126–1135, 2015.
[9] M. Jinno, B. Kozicki, H. Takara, A. Watanabe, Y. Sone,
T. Tanaka, and A. Hirano, “Distance-adaptive spectrum
resource allocation in spectrum-sliced elastic optical path
network,” IEEE Com. Mag., vol. 48, pp. 138–145, 2010.
[10] A. Klekamp, R. Dischler, and F. Buchali, “Limits of spectral
efficiency and transmission reach of optical-OFDM super-
channels for adaptive networks,” IEEE Photonics Technology
Letters, vol. 23, no. 20, pp. 1526–1528, 2011.
[11] I. Cerutti, F. Martinelli, N. Sambo, F. Cugini, and P. Castoldi,
“Trading regeneration and spectrum utilization in code-rate
adaptive flexi-grid networks,” Journal of Lightwave Technol-
ogy, vol. 32, no. 23, pp. 3894–3901, 2014.
[12] A. Bocoi, M. Schuster, F. Rambach, M. Kiese, C.-A. Bunge,
and B. Spinnler, “Reach-dependent capacity in optical net-
works enabled by OFDM,” in Conf. Optical Fiber Commu-
nication, 2009, pp. 1–3.
[13] H.-C. Le, H. Hasegawa, and K.-i. Sato, “Impact of electrical
grooming and regeneration of wavelength paths in creating
hierarchical optical path networks,” Optics express, vol. 19,
no. 26, pp. B882–B894, 2011.
[14] M. Jinno, K. Yonenaga, H. Takara, K. Shibahara, S. Ya-
manaka, T. Ono, T. Kawai, M. Tomizawa, and Y. Miyamoto,
“Demonstration of translucent elastic optical network based
on virtualized elastic regenerator,” in National Fiber Optic
Engineers Conf. Optical Soc. America, 2012, pp. PDP5B–6.
[15] M. Jinno and H. Takara, “Elastic optical transponder and
regenerator: Toward energy and spectrum efficient optical
transport networks,” in 2012 International Conference on
Photonics in Switching (PS), 2012, pp. 1–3.
[16] H.-C. Le, N. T. Dang, and S. Vu-Van, “Dynamic multi-
domain elastic optical networks with 4R regeneration capa-
ble border nodes,” in National Foundation for Science and
Technology Development Conference on Information and
Computer Science (NICS), 2015, pp. 169–173.
[17] P. Poggiolini, A. Carena, V. Curri, G. Bosco, and
F. Forghieri, “Analytical modeling of nonlinear propagation
in uncompensated optical transmission links,” IEEE Photon-
ics technology letters, vol. 23, no. 11, pp. 742–744, 2011.
[18] G. Bosco, P. Poggiolini, A. Carena, V. Curri, and
F. Forghieri, “Analytical results on channel capacity in
uncompensated optical links with coherent detection,” Optics
express, vol. 19, no. 26, pp. B440–B451, 2011.
[19] R.-J. Essiambre, G. Foschini, P. Winzer, and G. Kramer,
“Capacity limits of fiber-optic communication systems,” in
Optical Fiber Com. Conf. Optical Soc. America, 2009.
[20] J. G. Proakis and M. Salehi, Digital Communications, 4th ed.
McGraw-hill New York, 2000.
[21] Z. Zhu, W. Lu, L. Zhang, and N. Ansari, “Dynamic ser-
vice provisioning in elastic optical networks with hybrid
single-/multi-path routing,” Journal of Lightwave Technol-
ogy, vol. 31, no. 1, pp. 15–22, 2012.
[22] H.-C. Le, N. T. Dang, and N. D. Nguyen, “Dynamic RMSA
algorithm for multi-domain EONS with limited virtualized
elastic regenerators,” in International Conference on Com-
munications and Electronics (ICCE), 2016, pp. 46–51.
Le Hai Chau received the B.E. degree
in Electronics and Telecommunications En-
gineering from Posts and Telecommuni-
cations Institute of Technology (PTIT) of
Vietnam in 2003, and the M.Eng. and
D.Eng. degrees in Electrical Engineering
and Computer Science from Nagoya Uni-
versity of Japan in 2009 and 2012, respec-
tively. From 2003 to 2006, he worked at PTIT as a lecturer. He
was a researcher in Nagoya University from 2012 to 2014. He
then became a research scientist at University of California, Davis.
Now, he is a lecturer at PTIT. He is an IEEE member.
Dang Hoai Bac received the B.E. de-
gree from the Hanoi University of Tech-
nology ,Hanoi, Vietnam, in 1997, and both
the M.E. degree and the Ph.D. degree in
Electronics and Telecommunications from
the Posts and Telecommunications Institute
of Technology (PTIT), Hanoi, Vietnam in
2004 and 2010, respectively. He is currently
an Associate Professor/Vice Director at PTIT. His current research
interests include the area of communication theory with a partic-
ular emphasis on modeling, design, and performance evaluation.
51