key: cord-0525240-84ejmt6n
authors: Dao, Hai
title: Network Coding in Photonic-land: Three Commandments for Future-proof Optical Core Networks
date: 2021-05-03
journal: nan
DOI: nan
sha: 360969a1009d526eff79171975ddb514f33cd3e5
doc_id: 525240
cord_uid: 84ejmt6n

The digital transformation has been underway, creating digital shadows of (almost) all physical entities and moving them to the Internet. The era of Internet of Everything has therefore started to come into play, giving rise to unprecedented traffic growths. In this context, optical core networks forming the backbone of Internet infrastructure have been under critical issues of reaching the capacity limit of conventional fiber, a phenomenon widely referred as capacity crunch. For many years, the many-fold increases in fiber capacity is thanks to exploiting physical dimensions for multiplexing optical signals such as wavelength, polarization, time and lately space-division multiplexing using multi-core fibers and such route seems to come to an end as almost all known ways have been exploited. This necessitates for a departure from traditional approaches to use the fiber capacity more efficiently and thereby improve economics of scale. This paper lays out a new perspective to integrate network coding (NC) functions into optical networks to achieve greater capacity efficiency by upgrading intermediate nodes functionalities. In addition to the review of recent proposals on new research problems enabled by NC operation in optical networks, we also report state-of-the-art findings in the literature in an effort to renew the interest of NC in optical networks and discuss three critical points for pushing forward its applicability and practicality including i) NC as a new dimension for multiplexing optical signals ii) algorithmic aspects of NC-enabled optical networks design iii) NC as an entirely fresh way for securing optical signals at physical layers

The Internet of the future will be evolving to keep pace with the proliferation of digital innovations in all shapes and sizes [1] . On one hand, from the user sides in broadest senses, there has been accelerated development of Internet usage and bandwidth-intensive services including such as autonomous vehicles, remote robotic surgery, tele-presence, to name but a few and of course the future unknown ones. A case for illustration is the operation of an autonomous car will generate about 5 Terabytes of data and it is no less than a current supercomputer in terms of generating and transmitting a colossal amount of data. On the other hand, thanks to the convergence of technologies, the massive adoption for new services will take more and more shorter time and it means that what is rare or unimaginable today can become widely adopted tomorrow. Such trend can be observed from the entire migration of working, learning and several other conventional physical activities to the Internet platform during the COVID-19. In this context, Internet traffic have been exploding and will continue to rise dramatically [2] , [3] .

Behind the scenes to support such exponential Internet traffic growth is the fiber-optics networks. Indeed, fiberoptic communications make up the backbone of Internet infrastructure, providing ultra-high capacity, low-latency and highly secured communication channels and hence, enabling the coming into availability of digital society and digital future [1] , [4] . For many years since its birth in 1966, fiber-optical communications have undergone remarkable progresses in capacity, spectrum efficiency and ultimately cost to keep pace with increasing demand in traffic growth and this is owning to many ground-breaking scientific and technological advances in the field of both electronics, photonics and digital signal processing [5] - [16] . In a span of roughly 30 years from 1990 to 2020, a 400 times increase in data rate has been recorded for a typical single wavelength channel, a spectacular rise from 2.5 Gb/s in around 1990 to 1 Tb/s in 2020 [17] . In addition to increase in data rate per channel, the number of channels per fiber is also multiplied via various multiplexing techniques to further improve overall transmission capacity. In particular, four physical dimensions of a light wave traveling along a fiber has been exploited including amplitude and phase, polarization, wavelength and lately space-division multiplexing by using multiple single-mode fibers in a same fiber cable or multi-core/few-modes fibers [18] . Continuing the direction of improving multiplexing techniques and/or increase bit-rate per channel then appears to be increasingly difficult as hard limitation has been almost reached [19] . On the constraint of finite spectrum capacity, innovative techniques and non-conventional ideas are needed to break from the norm in achieve higher capacity at greater economic of scales.

Network coding (NC), originally invented in [20] , has soon become a revolutionary technique in networking to attain greater throughput, security and capacity. The core idea is that intermediate nodes, instead of simply storing and broadcasting data as in conventional networking paradigm, is allowed to manipulate data and then forward such (non-) linearly combined data to its output. The successes of network coding has been remarkable and therefore has been a de facto in future wireless networks. However, with radically different transmission conditions, the wisdom from wireless networks could not be directly applied to optical networking realms. In the past due to the immature of photonic signal processing technologies, the proposal of exploiting NC in optical networks remains inadequately addressed due to the limited capability if network coding is performed in electronic domain. Nevertheless, the considerable advances in photonic signal processing and enabling technologies animates the interest in leveraging photonic network coding to re-define optical networking realm towards greater cost and energy-efficiency. Indeed, recent works in [21] - [33] have renewed the interests and potential benefits of adding network coding layer to optical networks. The use of digital all-optical physical-layer network coding has also been extended to mm-wave radio-over-fiber networks [34] , in visible light communications [35] , [36] , and in passive optical networks (PON) [37] . Data-center networking has been an active area for applying NC to reduce traffic [38] - [40] . Physical layer encryption with NC has been also addressed in [41] , [42] .

In this paper, we argue that Network Coding and its implementation in photonic domain thanks to the maturing of enabling technologies could be a new venue to push forward the capacity boundary of a conventional optical fiber. We lay out a framework via an illustrative example to highlight potential uses cases of NC in optical networks relying on the core idea of transforming the functionalities of intermediate nodes from traditional tasks of regeneration and/or switching to encoding/decoding. The final point we raise is about the algorithmic aspects when it comes to network design and planning with network coding-enabled realm to reach its full potential gains. In supporting our arguments, we include some recent research results highlighting the impact of re-designing optical transport networks with NC compared to the conventional ones.

Let us consider an exemplary case highlighting how network coding could be spectrally beneficial to optical networking. Assuming that there are two demands with dedicated protection from node A and node B to node C as shown in Fig. 1 . Their working signals and protection signals are indicated as A w , B w and A p , B p respectively.

Supposing that the illustrative optical networks in Fig. 1 adopting the opaque architecture and thus, the protection signals from node A and B being routed via node X are undergone optical-electrical-optical conversion before forwarding on fiber link XC. In the conventional operation, the elements composing of node X is illustrated in Fig.  2 . It can be seen that two optical transponders are needed to handle the protection signal A p (B p ) when it is routed over node X. From the functionality perspective, node X simply performs regeneration functions on individual input signals and forward such regenerated signals to proper output links. We now turn attention to the new scenario where the intermediate node (node X) is powered by the new capability of mixing/encoding input signals and this is highlighted in Fig. 3 . In this scenario, instead of simply processing input signal individually, the intermediate node X performs XORencoding between protection signal A p and B p to create an encoded one A p ⊕ B p . Such encoded A p ⊕ B p signal is then modulated by simply using one transponder instead of two transponders as in traditional case and is forwarded on link XC. Furthermore, on fiber link XC, rather than using two wavelength units to support A p and B p , only one wavelength unit is needed to accommodate the encoded A p ⊕ B p . At the destination node C, three signals are received including A w , B w and A p ⊕ B p and it is important to note that in case of any single link failures, the receiver is fully capable of recovering the lost signal from the two remaining ones. The decoding process is shown in Fig. 4 . What the aforementioned example tells us is that network coding opens up new opportunities to re-define the optical networking realm radically from design and planning to operation and management. Among potential benefits, as clearly revealed in the example, the capital cost associated with transponders and spectrum usage could be remarkably reduced. Furthermore, as a consequence of hardware and spectrum savings, operational expenditure could also be reduced by turning off un-used hardwares including such as transponders and/or in-line amplifiers. The first commandment for applying network coding could be therefore stated that photonic network coding should be viewed as a new venue for multiplexing optical signals, in addition to well-known dimensions including wavelength, time, polarization and space and exploiting this new dimension could potentially heralds future-proof core networks that can sustain explosive traffic growths.

With respect to the use of XOR coding, it has to be noted that there has been remarkable progresses in photonic XOR which permits fully optical encoding/decoding up to Terabit/s [43] and moreover, such operations could be optically performed between different modulation format signals [44] . For the possible concern on time overhead of encoding and decoding processes, the fact that the encoding and decoding are all performed in the photonic domain constitutes thus negligible impact to the recovery time. This is radically different with other network coding schemes proposed in the literature that is performed in electronic domain in conjunction with the opaque network architecture. Next, let's move to a scenario where the introducing of network coding functions at optical nodes could drive the coming into availability of new services. Fig. 5 illustrates the case where there are two traffic demands from node A and node B to node C respectively. However, different from traditional requests, demand B is required an added service that the signal transmitted from node B to node C must be protected against physical eavesdropping and/or attacking. In offering such services, the optical signal on the route from node B to node C must be physically encrypted so that the tapping of such signal at intermediate nodes or in the middle way could not extract meaningful information. In this context, photonic network coding could be leveraged at node B to create a physically encrypted signal, that is, A ⊕ B and transmit such encrypted one all the way to the receiving node C. By transmitting an encrypted version, the traffic from node B to node C could not be eavesdropped or decoded by any means. Only at the receiving node, the original signal is decoded with the key from the complementary signal, that is, B = A⊕(A⊕B). This example highlights a radically original approach for providing security services at the physical layer without overhead in transmission as still only one wavelength unit is needed. Moreover, it has to be noted that such secured transmission is performed at a scale of a wavelength capacity which could be potentially up to Tb/s. In acknowledging this unique opportunity, the second commandment for application of photonic network coding is that there is a tremendous opportunity in upgrading the security of optical networks in a cost and energy-efficient manner and that could be performed at a scale of wavelength granularity.

It has become clear that the incorporation of photonic network coding paves the way for re-defining the optical networks to achieve greater efficiency in terms of capital, operational and security aspects. However, it should be noted that in order to realize such benefits, in addition to the technological upgrades, the algorithmic factors are of crucial importance. It has been well-known that algorithms involving all phases of a network from designing, planning to operation and management. Clearly a good set of algorithms result in better network utilization, even optimal one which in turn translates to considerable gain in capital and operation revenues. Conversely, poor algorithms could under-utilize the resources and as a consequence, potential benefits are failed to realized [30] , [45] - [48] . Well-designed algorithms are therefore critical to optimize the resources in optical networks. Although network coding could be a potentially new venue for leveraging optical networks, it indeed adds a new layer of complexity when it comes to network designs and operations. The introduction of network coding functions in optical networks gives rise to the issue of network coding assignment in which the determination of pair of demands for encoding/decoding, the coding nodes and spectrum-related issues have to be efficiently solved. This observation leads us to a third commandment, that is, innovative algorithms including exact and heuristic ones should be developed to tackle the various shades in applying network coding and network coding-related design problems. Otherwise, the potential gain enabled by network coding may not be fully realized.

III. RE-DESIGNING OPTICAL TRANSPORT NETWORKS WITH NETWORK CODING: SOME RESULTS In this part, we report some of our recent results regarding to the re-design of optical transport networks with network coding operations. We compare the network performances measured by routing cost, spectrum cost and network throughput. The network topology under test is a realistic one, COST239 as shown in Fig. 6 and Table I summarizes its key characteristics. All the results comparing network codingenabled designs and conventional designs ones are based on solving optimally the network design formulation, often based on mixed integer linear programming models. Such comparison based on optimal values of both designs are to guarantee the fairness of benchmarking. Three technologies for optical core networks are covered including opaque WDM networks, transparent WDM ones and the recently proposed Elastic Optical Networks. 

The opaque WDM networks have been remained widely deployed in optical core networks thanks to, on one hand, the historical factor, and on the other hand, its merit of regeneration capability at intermediate nodes. For accommodating traffic demands, the central issue to be solved is the routing problem and if network coding is enabled, a new problem arises, called, routing and network coding assignment problem (RNCA). In particular, in addition to the tradition task of identifying the optimal route for each demand, new opportunities and also challenges emerge with respect to determining pair of demands for encoding and decoding to reduce the overall traffic in the network. In [22] , [29] , we provided the optimal design framework for solving the RNCA problem and Table II highlights the comparison between the coding-aware designs and non-coding ones. It has been clearly seen that applying network coding brings about remarkable improvements in network performances including routing cost and network throughput. Besides, it has to be noted that such performance gain comes at a cost of increasing complexity for network designs.

Transparent WDM networks have long been progressing with the advances in transmission and switching technologies, allowing long-haul and ultra-high bit-rate operations. Optical core networks have then adopted this paradigm to partly replace the legacy opaque ones. Provisioning traffic demands in transparent WDM networks involves solving the routing and wavelength assignment (RWA) which is known as computationally hard. The arrival of network coding nevertheless adds a new layer of complexity for network designs, that is, the network coding assignment. In addition to the determination of coding nodes, pair of demands for encoding, the critical issue of wavelength assignment for a pair of demand before and after encoding must be taken into account. In [21] , [23] , an optimal framework for solving such issue (RWNCA) was proposed and 

Elastic optical networks (EONs) represent a future-proof framework as adaptive traffic provisioning depending on the specific bit-rate requirement and/or transmission quality of a demand is permitted. In doing so, they enables the use of fiber capacity more efficiently, therefore avoiding over-provision or under-utilizing the fiber spectrum. A central issue to be tackled in EONs is the routing and spectrum assignment (RSA) problem, that is, a more complicated version of RWA problem in WDM networks. In an analogy to the RWNCA problem, in network coding-enabled elastic optical networks, the routing, spectrum and network coding assignment (RSNCA) arises and have to be solved optimally to maximize the network coding benefits. Table IV compares the network throughput in traditional EONs and in network coding-enabled EONs and encouraging improvement up to roughly 30% has been observed in our studies.

This paper has discussed about the use of photonic network coding to transform the entire operations of existing [28] optical networks infrastructure towards a greater capacity and security efficiency. This is enabled by upgrading the intermediate nodes functions with encoding/decoding capabilities in photonic domain. Three critical points have been raised addressing various shades of applying photonic NC including NC as a new dimension for multiplexing optical signals, NC as a radically efficient encryption tool at scale and algorithms aspects lie at the heart of realizing NC benefits.

Optical transport for industry 4.0

On the spectrum-efficiency of qos-aware protection in elastic optical networks

Qosaware protection in elastic optical networks with distanceadaptive and reconfigurable modulation formats

Contribution to Flexible Optical Network Design: Spectrum Assignment and Protection

Combining heuristic and exact approaches for solving the routing and spectrum assignment problem

An efficient networkside path protection scheme in ofdm-based elastic optical networks

On the routing and spectrum assignment with multiple objectives

On the advantages of co-ofdm transponder in network-side protection

On the usage of flexible transponder in survivable transparent flex-grid optical network

A novel adaptive operation of multi-line rate transponder for dedicated protection in wdm network

Multi-objective genetic algorithm for solving routing and spectrum assignment problem

An efficient genetic algorithm approach for solving routing and spectrum assignment problem

On the efficient use of multi-line rate transponder for shared protection in wdm network

On the spectrum-efficiency of transparent optical transport network design with variable-rate forward error correction codes

An efficient column generation approach for solving the routing and spectrum assignment problem in elastic optical networks

On development of efficient data acquisition systems and parameter extraction technique for dfb lasers

Fiber-optic transmission and networking: the previous 20 and the next 20 years

Breakthroughs in photonics 2014: Optical physical-layer network coding, recent developments, and challenges

Communication networks beyond the capacity crunch

Network information flow

Leveraging the survivable all-optical wdm network design with network coding assignment

An optimal design framework for 1+1 routing and network coding assignment problem in wdm optical networks

On optimal designs of transparent wdm networks with 1+1 protection leveraged by all-optical xor network coding schemes

A bi-objective integer linear programming model for the routing and network coding assignment problem in wdm optical networks with dedicated protection

On solving the 1 + 1 routing, wavelength and network coding assignment problem with a bi-objective integer linear programming model

Re-designing dedicated protection in transparent wdm optical networks with xor network coding

A priority-based multiobjective design for routing, spectrum, and network coding assignment problem in network-coding-enabled elastic optical networks

Network coding for improving throughput in wdm optical networks with dedicated protection

The consolation of network coding and partial protection techniques to optical transport networks in data, data, data era

Performance modeling of diversity coded path protection in obs/ops networks

Survivable virtual network link protection method based on network coding and protection circuit

Cost comparison of 1+1 path protection schemes: A case for coding

Digital alloptical physical-layer network coding for 2gbaud dqpsk signals in mm-wave radio-over-fiber networks

Phase-aligned physical-layer network coding in visible light communications

Modulated retroreflector-based physical-layer network coding for space optical communications

Implementation of network-coding approach for improving the ber performance in nonorthogonal multiple access (noma)-pon

Physical-layer network coding for passive optical interconnect in datacenter networks

Exploiting parallelism with random linear network coding in high-speed ethernet systems

Survivable inter-datacenter network design based on network coding

Confidentiality meets protection in elastic optical networks

Providing confidentiality in optical networks: Metaheuristic techniques for the joint network coding-routing and spectrum allocation problem

1 tb/s all-optical xor and and gates using quantum-dot semiconductor optical amplifier-based turboswitched mach-zehnder interferometer

All-optical modulation format conversion and applications in future photonic networks

Three shades of partial protection in elastic optical networks

The achilles heel of some optical network designs and performance comparisons

Development of dynamic qot-aware lightpath provisioning scheme with flexible advanced reservation for distributed multi-domain elastic optical networks

A pragmatic approach for designing transparent wdm optical networks with multi-objectives

ACKNOWLEDGMENT This research is funded by the Vietnam National Foundation for Science and Technology Development (NAFOSTED) under the grant number 102.02-2018.09 for the project entitled, Network Coding for Spectrum-Efficient and Reliable Optical Core Networks (CEO)