Elastic optical networks (EONs) can help overcome the flexibility challenges imposed by emerging heterogeneous and bandwidth-intensive applications. Among the different solutions for flexible optical nodes, optical white box switches implemented by architecture on demand (AoD) have the capability to dynamically adapt their architecture and module configuration to the switching and processing requirements of the network traffic. Such adaptability allows for unprecedented flexibility in balancing the number of required nodal components in the network, spectral resource usage, and length of the established paths. To investigate these trade-offs and achieve cost-efficient network operation, we formulate the routing, modulation, and spectrum assignment (RMSA) problem in AoD-based EONs and propose three RMSA strategies aimed at optimizing a particular combination of these performance indicators. The strategies rely on a newly proposed internal node configuration matrix that models the structure of optical white box nodes in the network, thus facilitating hardware-aware routing of connection demands. The proposed strategies are evaluated in terms of the number of required modules and the related cost, spectral resource usage, and average path length. Extensive simulation results show that the proposed RMSA strategies can achieve remarkable cost savings by requiring fewer switching modules than the benchmarking approaches, at a favorable trade-off with spectrum usage and path length.

An approach to enhance core network survivability by utilizing dual-homing capabilities of the access network. Results reveal significant improvements in core network resource utilization and connection availability, suggesting benefits to network operators and service providers.

This paper studies the energy efficiency of exploiting dual-homing in access network to provide core network survivability. Simulation results show reduced number of utilized wavelengths, suggesting significant energy saving opportunities in the core network. © 2015 OSA.

Core network survivability affects the reliability performance of telecommunication networks and remains one of the most important network design considerations. This paper critically examines the benefits arising from utilizing dual-homing in the optical access networks to provide resource-efficient protection against link and node failures in the optical core segment. Four novel, heuristic-based RWA algorithms that provide dedicated path protection in networks with dual-homing are proposed and studied. These algorithms protect against different failure scenarios (i.e. single link or node failures) and are implemented with different optimization objectives (i.e., minimization of wavelength usage and path length). Results obtained through simulations and comparison with baseline architectures indicate that exploiting dual-homed architecture in the access segment can bring significant improvements in terms of core network resource usage, connection availability, and power consumption.

We propose an optimization model minimizing number of wavelengths in passive optical backbone networks and obtaining the same resource usage as in networks based on active switching while reducing both cost and power consumption.

This paper investigates the benefits of dynamic restoration with service relocation in resilient optical clouds. Results from the proposed optimization model show that service availability can be significantly improved by allowing a few service relocations.

A centralized network control and management plane, such as the one based on a path computation element (PCE), is highly beneficial in terms of resource optimization in wavelength division multiplexing optical networks. Benefits of centralized provisioning are even more evident when connection requests are provisioned in batches, i.e., they allow a better use of network resources via concurrent optimization. In this study, a dynamic bulk provisioning framework is presented with the objective of optimizing the use of network resources that also presents, as an additional benefit, the ability to yield a reduction of the control plane overhead. The rationale behind the proposed framework is based on a mechanism in which the PCE client is allowed to bundle and simultaneously send multiple labeled switch path (LSP) requests to the PCE where, in turn, several bundles can be concurrently processed together as a single bulk. From the network deployment perspective, a PCE-based network architecture is proposed to practically realize this approach. For dynamic bulk provisioning of optical LSP requests, a time-efficient integer linear programming (ILP) model (LSP BP ILP) is presented to minimize the request blocking, the network resource consumption, and the network congestion. In addition, a heuristic based on a greedy randomized adaptive search procedure (GRASP), namely LSP_BP_GRASP, is also proposed as a scalable alternative. The presented results demonstrate significant advantages of the proposed PCE bulk provisioning framework based on concurrent optimization in terms of reduced blocking probability and control overhead when compared with conventional dynamic connection provisioning approaches processing a single connection request at a time.

A time-efficient resource optimization model for dynamic concurrent provisioning of connection requests at PCE is proposed. It is shown that a significant performance improvement can be achieved without noticeable increase in connection setup-time.

The Path Computation Element (PCE) is a network entity utilized for network path computation operations, especially useful in optical networks based on wavelength division multiplexing (WDM). In the PCE paradigm, the communication between a node and the PCE is specified by the Path Computation Element Communication Protocol (PCEP). According to PCEP protocol, multiple LSP (Label Switched Path) requests can be bundled together before being sent to the PCE in order to reduce the control overhead. Multiple bundles received by the PCE can then be provisioned at once as a single bulk. Enabling bulk provisioning of LSP requests at the PCE in a concurrent manner can bring significant improvements in terms of higher network resource utilization and control plane overhead reduction. However, these advantages come at a cost of a longer connection setup-time and of an instantaneous increase in the network load, which may lead to a degradation of the network performance, e.g. blocking probability. In this study pros and cons of bulk provisioning are explored in shared path protection (SPP) by comparing sequential and concurrent path computation strategies. An efficient meta-heuristic named GRASP-SPP-BP (Greedy Random Adoptive Search Procedure for Shared Path Protection with Bulk Provisioning) is proposed for concurrent provisioning of primary and shared backup path pairs. GRASP-SPP-BP minimizes the backup resource consumption while requiring minimal path computation time. The presented results demonstrate that, in a SPP network scenario, a significant reduction in the PCEP control overhead, network blocking probability and backup resource consumption can be achieved via LSP bulk provisioning at the PCE with the proposed GRASP-SPP-BP approach.

Maximizing connection availability in WDM networks is critical because even small disruptions can cause huge data losses. However, there is a trade-off between the level of network survivability and the cost related to the backup resources to be provided. The 100% survivability can be achieved by dedicated path protection with multiple pre-reserved protection paths for each provisioned connection, i.e., DPP (1:N). Unfortunately, the blocking probability performance of DPP (1:N) is negatively affected by the large number of pre-reserved backup wavelengths standing-by unutilized. On the other hand, path restoration (PR)-based solutions ensure good blocking performance at the expense of lower connection availability.

The work in this paper aims at finding hybrid network survivability strategies that combine the benefits of both techniques (i.e., high availability with low blocking rate). More specifically, the paper focuses on a double link failure scenario and proposes two strategies. The first one, couples dedicated path protection DPP (1:1) with path restoration (referred to as DPP + PR) to minimize the number of dropped connections. The second scheme adds up the concept of backup reprovisioning (BR), referred to as DPP + BR + PR, in order to further increase the connection availability achieved by DPP + PR. Integer Linear Programming (ILP) models for the implementation of the proposed schemes are formulated. Extensive performance evaluation conducted in a PCE-based WDM network scenario shows that DPP + BR + PR and DPP + PR can significantly lower the blocking probability value compared to DPP (1:2) without compromising too much in terms of connection availability.

Two approaches based on backup reprovisioning and path restoration are proposed for dynamic failure recovery in survivable, PCE-based, WDM networks. Results show that proposed schemes can achieve high connection availability in double link failure scenarios.

These days there is a clear trend toward extending the reach of passive optical networks to cover large geographical areas, which enables reduction of the number of central offices and hence has the potential of cost saving in network operation. On the other hand, this reach extension necessitates the design of efficient dynamic bandwidth allocation schemes in order to tackle performance degradation caused by the increased propagation delay in long reach PONs. Among many existing approaches, the multi-thread-based DBA scheme where several bandwidth allocation processes are performed in parallel is considered one of the most effective options to improve network performance in LRPONs. We have found that without proper intercommunication between the overlapped threads, multi-thread DBA may lose efficiency and even perform worse than the conventional single-thread algorithm. With this in mind, this article reviews different inter-thread scheduling schemes for LR-PONs, and proposes a novel approach of integrating the key ideas of the existing ones. Extensive simulation results confirm that our proposed scheme can significantly improve DBA performance for LR-PONs under a variety of scenarios with consideration of different values of network load and reach.

The  Path  Computation  Element  (PCE)  concept  is  considered  to  be  beneficial  in  the network  connection  setup  operations,  especially  in  optical  networks  based  on wavelength  division  multiplex  (WDM)  transport  technology.  In  the  PCE  paradigm, communication  between  a  node  and  the  PCE  is  specified  by  the  Path  Computation Element  Communication  Protocol  (PCEP).  PCEP  allows  the  PCC  (Path  Computation Client) to send to the PCE more than one LSP (path computation) request at a time, i.e., multiple LSP requests can be bundled together before being sent to the PCE. Enabling bundling, and consequently the concurrent optimization of a large set of LSP requests at the PCE, may result in significant improvements in terms of network optimization and reduced  control  plane  overhead.  However,  these  advantages  come  at  a  cost  of increased  connection  setup-delay.  This  paper  explores  pros  and  cons  of  enabling bundling of LSP requests in terms of both control plane overhead reduction and benefits of  sequential  vs.  concurrent  path  computation  operations.  A  variety  of  scenarios  are analyzed,  including  a  WDM  mesh  network  providing  LSPs  with  both  dedicated  and shared  path  protection.  Results  demonstrate  significant  gains  in  terms  of  reduced control  overhead  using  LSP  bundling,  and  reduction  in  blocking  probability  using concurrent processing of bundled LSP requests at the PCE.

Concurrent RWA algorithm for differentiated services to process multiple LSP bundles at PCE is proposed. Significant blocking probability reduction has been observed at the expense of slightly increased LSP setup-time compared to a sequential approach.

LSP requests may be bundled to improve the network optimization process at the expense of an increased connection setup delay. A detailed study is conducted to evaluate the pros and cons of the bundling approach.

In Wavelength Division Multiplexed (WDM) networks with dynamic lightpath provisioning, connection requests are served without any prior knowledge of their arrival and departure times. As time passes, network resources may become fragmented because of the network dynamism. Under these circumstances it is highly beneficial to re-optimize (i.e., de-fragment) the existing lightpath configuration at some specific time instances to improve the network resource utilization and reduce the risk that future connection requests will be blocked. Assuming that this de-fragmentation process occurs during a re-optimization phase, this paper presents a set of strategies which govern the time instances when this re-optimization phase should be triggered as well as a set of strategies to decide which of the currently active lightpaths should be optimized at any given re-optimization phase. These strategies are referred to as when-to-re-optimize (when-t-r) and what-to-re-optimize (what-t-r) strategies, respectively. During the evaluation process particular attention is devoted to study the impact that when-t-r and what-t-r strategies have on the traffic disruption metrics (i.e., number of total disrupted connections, disruption time, reconfiguration time) inherent with the re-optimization process. Based on the evaluation results, it can be concluded that the choice of an optimal "when" and an optimal "what" to re-optimize strategy is dependent upon the performance objective (e.g. lower blocking probability or network disruption) in a given network scenario.

Elastic optical networking is considered a promising candidate to improve the spectral efficiency of optical networks. One of the most important planning challenges of elastic optical networks is the NP-hard routing and spectrum assignment (RSA) problem. In this paper, we investigate offline RSA in elastic filterless optical networks, which use a passive broadcast-and-select architecture to offer network agility. Here elastic optical network is referred to as the optical network that can adapt the channel bandwidth, data rate, and transmission format for each traffic demand in order to offer maximum throughput. In elastic filterless networks, the presence of unfiltered signals resulting from the drop-and-continue node architecture must be considered as an additional constraint in the RSA problem. In this paper, first the RSA problem in elastic filterless networks is formulated by using integer linear program (ILP) to obtain optimal solutions for small networks. Due to the problem complexity, two efficient RSA heuristics are also proposed to achieve suboptimal solutions for larger networks in reasonable time. Simulation results show that significant bandwidth savings in elastic filterless networks can be achieved compared to the fixed-grid filterless solutions. The proposed approach is further tested in multi-period traffic scenarios and combined with periodical spectrum defragmentation, leading to additional improvement in spectrum utilization of elastic filterless optical networks.

This papers reports the advances in optical core networks research coordinated in the framework of the e- photon/ONe and e-photon/ONe+ networks of excellence.

This paper reports the view of the e-Photon/ONe community regarding the research directions in optical transport networks. Results stemming from joint international research activities in the framework of the project are summarized. This summary presents, necessarily, little details, but wants to draw the attention of the reader towards the most advanced research activities ongoing on the broad topics of optical core networks, from wavelength routed network, to optical burst switching and optical packet switching.

In recent years, energy-efficient design of optical WDM networks has become increasingly important as efforts to reduce the operational expenditure (Opex) and the carbon footprint of the internet are prioritized. In this paper we focus on energy- efficient survivable network design where backup resources are shared for efficient capacity consumption. However there is a trade-off between energy-efficiency and survivability. Survivable network design strategies lead to lightly loaded links in order to minimize the risk in case of a failure and to increase the shareability of backup resources. On the contrary, energy-efficient network design strategies tend to increase the load in a set of links as a consequence of concentrating the traffic in order to be able to switch off as much network resources as possible. In this study, we present a novel method to simultaneously minimize Capex and Opex while providing an energy-efficient, shared backup protected network, under the assumption of backup capacity in sleep mode. For the first time we propose an ILP formulation for the energy-efficient shared backup protection problem. By exploiting the sleep mode for the backup resources, we observe that the ILP solution of our mathematical model brings up to 40% gain in energy efficiency in comparison to energy-unaware shared backup protection approach.

Energy consumption of communication networks is growing very fast due to the rapidly increasing traffic demand. Consequently, design of green communication networks gained a lot of attention. In this paper we focus on optical Wavelength Division Multiplexing (WDM) networks, able to support this growing traffic demand. Several energy-aware routing and wavelength assignment (EA-RWA) techniques have been proposed for WDM networks in order to minimize their operational cost. These techniques aim at minimizing the number of active links by packing the traffic as much as possible, thus avoiding the use of lightly loaded links. As a result, EA-RWA techniques may lead to longer routes and to a high utilization on some specific links. This has a detrimental effect on the signal quality of the optical connections, i.e., lightpaths. In this study we quantify the impact of power consumption minimization on the optical signal quality. and address this problem by proposing a combined impairment and energy-aware RWA (IEA-RWA) approach. Towards this goal we developed a complete mathematical model that incorporates both linear and non-linear physical impairments together with an energy efficiency objective. The IEA-RWA problem is formulized as a Mixed Integer Linear Programming (MILP) model where both energy efficiency and signal quality considerations are jointly optimized. By comparing the proposed IEA-RWA approach with existing RWA (IA-RWA and EA-RWA) schemes, we demonstrate that our solution allows for a reduction of energy consumption close to the one obtained by EA-RWA approaches, while still guaranteeing a sufficient level of the optical signal quality.

We investigate the benefit of utilizing electric price difference between time intervals of the day for networks covering multiple time-zones. We show that by exploiting this time-and-price variation electric bill can be significantly reduced.

We consider an energy-aware WDM core network, where the electric prices direr according to time intervals of the day, referred to as time-of-use rates. Since a large backbone network may have nodes scattered over several time zones, we exploit the time-of-use rates to lower the electric bill of the network provider. By using an ILP formulation of the bill minimization problem, we investigate the gain of total electric bill over two approaches where capacity and energy are optimized under different network traffic load conditions.

In this paper, we propose and evaluate a novel joint scheduling algorithm for multiple services which can support differentiated services and bandwidth allocation with global priority of different traffic such as Triple-play: voice, video and Internet services. Simulation results show that for the higher priority traffic better delay and jitter performance can be achieved compared with the lower priority traffic.

Dynamic bandwidth allocation (DBA) is one of the key issues for the current (1G) and next-generation (10G) Ethernet-based passive optical network (EPON) systems. We present a novel bandwidth scheduling scheme that integrates specific scheduling implementations in the optical line terminal and optical network units. This scheduling enables multiservice access with scalable quality of service support for the triple-play (video, voice, and data) services and open access. Our simulation results show that the proposed scheduling algorithm performs very well in supporting service differentiation and fair allocation of bandwidth to different service providers. A performance comparison between 1G and 10G systems is also presented. To the best of our knowledge, no detailed study of DBA in a 10G EPON can be found in the literature so far.

In this paper we propose an enhanced IPACT with limited service for multi-thread DBA in long-reach EPON. We evaluate our scheme by simulations in single-thread and double-thread cases and show that the DBA performance in terms of average delay and jitter can be significantly improved by the proposed algorithm.

We present and evaluate two novel node architectures based on optical switch matrices and wavelength converters (WCs). Relatively small and cheap switches are required while WCs efficiently improve blocking probability

The significance of broadband and multimedia telecommunications is still increasing and the use of fibre-optic technology in the access network is growing very fast in order to meet customers demand. Along with the higher bandwidth demand, increasing number of subscribers, and advances in the Wavelength Division Multiplexing (WDM) device technology, the WDM Passive Optical Network (PON) and hybrid WDM/TDM (Time Division Multiplexed) PON has been considered as a next generation solution for the broadband access. Meanwhile, in order to meet Service Level Agreement (SLA) and guarantee the appropriate level of connection availability, fault management within any type of the PONs becomes more significant for the reliable service delivery and business continuance. Connection availability is an issue of deep concern to network operators since failure of any access network component, and thus interruption of their services, could result in significant losses of revenue. This chapter reviews some protection schemes in PONs and provides reliability performance evaluation for the considered architectures. A customer typically expects the end-to-end service availability at least at the same level as that provided by the traditional copper based systems. Thus, before investing in a new technology network operators need to make sure that it will not degrade the service quality perceived by the customers. To avoid misunderstanding we provide the set of definitions used in this chapter. Next, we review the protection schemes in PONs and follow with the reliability analysis and performance evaluation. Finally, we draw some conclusions.

Advances in optical coherent transmission and electrical compensation technologies (such as coherent receiver and forward error correction FEC) have stimulated ideas for novel optical network architectures. Recently proposed passive wide area network solution, referred to as filterless optical network [1 #x2013;2] eliminates or minimizes the usage of active photonic reconfigurable network elements. In this approach, only the passive splitters and combiners for interconnecting the fiber links are utilized, which makes this network architecture more cost- and energy-effective as well as more reliable compared with networks based on active optical switching. However, the filterless optical network architecture implies some constraints on fiber interconnection design, maximum fiber-tree length and wavelength reuse due to its broadcast nature. Consequently, filterless solution always requires more resources (i.e. number of wavelengths) compared with the active switched optical networks which are allowed to utilize reconfigurable and coloured components. In order to improve the wavelength utilization while maintaining flexibility of resource allocation, this work extends the idea of filterless optical network by introducing some passive coloured components (e.g., fiber Bragg grating FBG, red/blue filters, etc) to drop local signals at some determined nodes. This approach is referred to as semi-filterless optical network. Furthermore, the semi-filterless solution maintains the passive feature, enabling high reliability and efficiency of cost and energy. Meanwhile, its non-broadcast property at some determined nodes has potential to decrease the transmission impairments and hence relax the constraints on fiber interconnection design and the maximal transparent length, which are strict in the filterless optical network. Our preliminary results confirm the advantages of semi-filterless solution.

Reliability performance of next generation optical access (NGOA) networks offering high bandwidth and large service coverage is evaluated. A particular attention is paid to the reduction of the impact of failures in NGOA.

We propose a fast fault supervision technique compatible with an efficient resilience scheme for WDM PON. Several drop fibres can be monitored simultaneously so that a significant reduction of operational cost can be achieved.

Access networks based on optical fiber can easily fulfill high bandwidth demand and cover large service areas. Fiber access networks are also scalable to meet the future capacity request. Among several alternatives, passive optical network (PON)is considered as the most promising one, because its passive point-to-multipoint architecture is characterized by relatively low deployment and operational cost. On the other hand, the passive point-to-multipoint feature in PONs also creates a number of challenges, such as efficient resource allocation and cost-effective protection. This paper provides an overview of the advances related to these two issues and points out the research topics that are still open and need to be investigated.

We propose what we believe to be a novel protection scheme compatible with smooth migration from a time-division multiplexing (TDM) passive optical network (PON) to a WDM/TDM-PON. We show that our scheme is very cost effective while keeping connection availability, recovery time, and power budget at an acceptable level. We focus on the protection schemes, overview the existing methods, and introduce a link protection scheme compatible with smooth migration from TDM-PON to hybrid WDM/ TDM-PON. Furthermore, we analyze the cost, connection availability, recovery time, and optical link budget for different protection schemes in order to find the cost-effective solution both for the TDM-PON and hybrid WDM/ TDM- PON. (c) 2007 Optical Society of America.

Due to the tremendous growth of traffic volume caused by both exponential increase of number of Internet users and continual emergence of new bandwidth demanding applications, high capacity networks are required in order to satisfactorily handle the extremely large amount of traffic. Hence, optical fiber communication is the key technology for the network infrastructure. This book addresses design and analysis of access and core networks targeting important research problems, which need to be tackled for the effective realization of next-generation optical networks. We hope that this book will give the researchers and engineers who work on optical networks an overview of the current trends and contribute to a deeper understanding of problems related to next-generation optical access and wide-area networks, which have tremendous potential to increase network capacity and support the rapidly growing bandwidth demand.

An optical communications network ( 10) comprising an optical line termination ( 12), and a first optical multiplexer/demultiplexer (MUX/DEMUX) ( 14), comprising N input ports and N output ports. The network (10) further comprises a plurality of wavelength division multiplexed passive optical network (WDM PON) distribution networks (22) coupled to a respective one of the MUX/DEMUX output ports and comprising optical network units (20). Feeder fibres ( 16) are coupled between the optical line termination and a respective one of the MUX/DEMUX input ports. Interconnection fibres ( 18) are coupled between respective pairs of the optical network units.

We propose a novel protection scheme compatible with smooth migration from TDM-PON to WDM/TDM-PON. We show that our scheme is very cost-effective while keeping connection availability, recovery time and power budget at the acceptable level.

A novel shared protection scheme using WDM in overlaid PONs is presented and evaluated. A large number of users can be supported and connection availability higher than 99.999% can be offered at a minimum cost.

A scalable and reliable architecture for both a wavelength division multiplexing passive optical network and a hybrid wavelength and time division multiplexing passive optical network with self-healing capability is presented and evaluated. Our protection scheme is compatible with a cascaded arrayed waveguide grating that can accommodate an ultra-large number of end users. A simple interconnection pattern between two adjacent optical network units (ONUs) is applied in order to provide protection for distributed fibers between a remote node and the ONUs. Therefore, the investment cost on a per-user basis can be significantly reduced. Meanwhile, the performance evaluation shows that our approach can achieve high connection availability while maintaining the support of long reach and high splitting ratio.

This paper proposes a novel protection scheme based on the cyclic property of an array waveguide grating (AWG) and neighboring connection pattern between two adjacent optical network units (ONUs) for the hybrid WDM/TDM passive optical networks (PONs). Our scheme uses 50% fewer wavelengths while offering one order of magnitude better connection availability than the existing scheme.

A novel protection architecture for passive optical networks (PONs) is presented and evaluated. It is based on the cyclic property of arrayed waveguide gratings (AWGs) and the interconnection between two adjacent optical network units. The proposed scheme is compatible with both wavelength-division-multiplexing (WDM) PONs and hybrid WDM/time-division-multiplexing PONs. It is compared with two existing schemes and shown to have several advantages: 1) 50% less wavelengths is needed; 2) the fiber interconnections are simplified; 3) the connection availability is improved by one order of magnitude.

We propose two asynchronous optical packet switch architectures, with efficient contention resolution based on controllable optical buffers and tunable wavelength converters TWCs. Providing a few shared optical buffers significantly boosts the performance obtained by TWCs.

We present the evolution of PON protection and compare reliability performance related to investment and management cost for some representative cases. Our results can indicate the most cost efficient architectures.

Fiber to the home is the future-proof technology for broadband access networks. Several fiber access network architectures have been developed (e.g., point-to-point, active optical network, and passive optical network). PON is considered the most promising solution due to the relatively low deployment cost and high resource efficiency. Meanwhile, because of the growing demand for reliable service delivery, fault management is becoming more significant in all parts of communications networks. However, there is a trade-off between the cost of protection and the level of service reliability. Since economical aspects are most critical in the access part of networks, improving reliability performance by duplication of network resources (and capital expenditures) could be too expensive. Therefore, recent work has focused on PON protection schemes with reduced CAPEX. The future trend will probably migrate toward minimizing operational expenditures during the access network lifetime. The main contributions of this article include providing a general method for CAPEX and OPEX analysis that can be applied to any type of fiber access network with consideration of changed component cost in time and variable take rates, and comparing the total cost (i.e., sum of CAPEX and OPEX) for the selected representative architectures with and without protection for business and residential users in relation to reliability performance. The aim is to give a guideline for the design of the most cost-effective protection schemes, while maintaining acceptable service reliability.

This Study investigates the use of information Summary (IS) applied to the advertised available resources to reduce the amount of link state advertisement (LSA) overhead, which is necessary to achieve distributed dynamic routing in WDM networks,. As illustrated in the Study, if carefully designed, the resulting IS-LSA protocol can significantly contain the size of the advertised data set without excessively affecting the network performance, i.e., the blocking probability caused by routing decisions based on incomplete link state information.

This paper shows that internal transmission capacity limitations in disaggregated data centers cannot be ignored. Insufficient capacity may reduce the inherent benefits of resource disaggregation in terms of resource utilization compared to the integrated solutions.

Several optical interconnect architectures inside data centers (DCs) have been proposed to efficiently handle the rapidly growing traffic demand. However, not many works have tackled the interconnects at top-of-rack (ToR), which have a large impact on the performance of the data center networks (DCNs) and can introduce serious scalability limitations due to their high cost and power consumption. In this paper, we propose a passive optical ToR interconnect architecture (POTORI) to replace the conventional electronic packet switch (EPS) in the access tier of DCNs. In the data plane, POTORI relies on a passive optical coupler to interconnect the servers within the rack and interfaces toward the aggregation/core tiers. The POTORI control plane is based on a centralized rack controller responsible for managing the communications among the servers in the rack. We propose a cycle-based medium access control (MAC) protocol to efficiently manage the exchange of control messages and the data transmission inside the rack. We also introduce and evaluate a dynamic bandwidth allocation algorithm for POTORI, namely largest first (LF). Extensive simulation results show that, with the use of fast tunable optical transceivers, POTORI and the proposed LF strategy are able to achieve an average packet delay below 10 μs under realistic DC traffic scenarios, outperforming conventional EPSs. On the other hand, with slower tunable optical transceivers, a careful configuration of the network parameters (e.g., maximum cycle time of the MAC protocol) is necessary to obtain a good network performance in terms of the average packet delay.

To efficiently handle the fast growing traffic inside data centers, several optical interconnect architectures have been recently proposed. However, most of them are targeting the aggregation and core tiers of the data center network, while relying on conventional electronic top-of-rack (ToR) switches to connect the servers inside the rack. The electronic ToR switches pose serious limitations on the data center network in terms of high cost and power consumption. To address this problem, we recently proposed a passive optical top-of-rack interconnect architecture, where we focused on the data plane design utilizing simple passive optical components to interconnect the servers within the rack. However, an appropriate control plane tailored for this architecture is needed to be able to analyze the network performance, e.g., packet delay, drop rate, etc., and also obtain a holistic network design for our passive optical top-of-rack interconnect, which we refer to as POTORI. To fill in this gap, this paper proposes the POTORI control plane design which relies on a centralized rack controller to manage the communications inside the rack. To achieve high network performance in POTORI, we also propose a centralized medium access control (MAC) protocol and two dynamic bandwidth allocation (DBA) algorithms, namely Largest First (LF) and Largest First with Void Filling (LFVF). Simulation results show that POTORI achieves packet delays in the order of microseconds and negligible packet loss probability under realistic data center traffic scenarios.

The growing popularity of cloud based applications is drastically increasing the traffic volume that datacenters have to handle. This brings the need for scalable, reliable, and energy-efficient interconnection networks inside the datacenters. Optical communication has been considered as a promising technology for datacenter applications due to its high energy- and cost-efficiency at ultra-high capacity. A typical datacenter interconnection network includes several tiers. Figure 1 shows an example with three tiers: edge, aggregation and core. Due to the large number of active devices on top of the rack (ToR) energy consumption at the edge tier is dominating the overall power consumed by all the switches within the datacenter [1-2]. Several passive optical interconnect approaches [2-4] have been proposed showing that replacing active optical devices by passive ones is possible to significantly reduce both the hardware cost and energy consumption, achieve lower maintenance complexity and offer a sufficient level of scalability. [GRAPHICS] Furthermore, several topologies, e.g., fat-tree [5], Quartz [6], are investigated in order to improve the resiliency and scalability, particularly for large-scale datacenters. However, it should be noted that the redundancy for these proposed topologies is often added in the aggregation and core tiers rather than the edge tier, due to the cost issue. Although passive optical ToR solution by nature could provide better reliability performance than its active counterpart, the intra-rack communication may still need survivability strategies to meet very high connection availability requirement. For instance, the required availability of fault-tolerant datacenter infrastructure (including electrical power supply, storage and distribution facilities) should be higher than 99.995% [7]. Then the expected availability for any connection established within the datacenter needs to be even higher, since the communication system is only a part of the site infrastructure. In this regard, we analyse reliability performance of optical interconnects and identify the key part to be protected. Based on it, we propose some reliable passive optical interconnects for the edge tier of the datacenter interconnection networks. They can achieve ultra-high connection availability for intra-rack communications and adapt to any topology, e.g., fat-tree and Quartz, designed to increase scalability and reliability performance for the overall datacenter network.

To address the sustainability, scalability, and reliability problems that data centers are currently facing, we propose three passive optical interconnect (POI) architectures on top of the rack. The evaluation results show that all three architectures offer high reliability performance (connection availability for intra-rack interconnections higher than 99.999%) in a cost-efficient way.