GRGeorge N. Rouskas

George N. Rouskas

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


George N. Rouskas has a total of 153 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Introduction to the Benchmarking in Optical Networks Special Issue

    by Alejandra Beghelli, George Rouskas, Paul Wright

    The first part of this special issue contains five papers addressing different aspects of good benchmarking practices in optical network research. We start with a brief introduction, followed by an overview of the papers and an invitation to contribute to the second part.


  • Symmetry-free routing and spectrum assignment: a universal algorithm based on first-fit

    by George N. Rouskas

    First-fit (FF) is a well-known and widely deployed algorithm for spectrum assignment (SA), but until our recent study [ J. Opt. Commun. Netw. 14,165(2022)JOCNBB 1943-0620 10.1364/JOCN.445492 ], investigations of the algorithm had been experimental in nature and no formal properties of the algorithm with respect to SA were known. In this work, we make two contributions. First, we show that FF is auniversalalgorithm for the SA problem in the sense that, for any variant, 1) it can be used to construct solutions equivalent to, or better than, any solution obtained by any other algorithm, and 2) it can construct an optimal solution. This universality property applies to both the min-max and min-frag objectives and to variants of the SA problem with or without guard band constraints. Consequently, the spectrum symmetry-free model of our recent study [ J. Opt. Commun. Netw. 14,165(2022)JOCNBB 1943-0620 10.1364/JOCN.445492 ] extends to all known SA variants, which therefore reduce to permutation problems. Second, we extend the spectrum symmetry-free model to the routing and spectrum assignment (RSA) problem in general topologies. This model allows for the design of more efficient algorithms as it eliminates from consideration an exponential number of equivalent symmetric solutions. By sidestepping symmetry, the RSA solution space is naturally and optimally decomposed into a routing space and a connection permutation space. Building upon this property, we introduce a two-parameter, symmetry-freeuniversalalgorithm that can be used to tackle any RSA variant in a uniform manner. The algorithm is amenable to multi-threaded execution to speed up the search process, and the value of the parameters can be adjusted to strike a balance between running time and solution quality. Our evaluation provides insight into the relative benefits of path diversity (which determines the size of the routing space) and connection diversity (which determines the size of the permutation space).


  • Fast and scalable all-optical network architecture for distributed deep learning

    by Wenzhe Li, Guojun Yuan, Zhan Wang, Guangming Tan, Peiheng Zhang, George N. Rouskas

    With the ever-increasing size of training models and datasets, network communication has emerged as a major bottleneck in distributed deep learning training. To address this challenge, we propose an optical distributed deep learning (ODDL) architecture. ODDL utilizes a fast yet scalable all-optical network architecture to accelerate distributed training. One of the key features of the architecture is its flow-based transmit scheduling with fast reconfiguration. This allows ODDL to allocate dedicated optical paths for each traffic stream dynamically, resulting in low network latency and high network utilization. Additionally, ODDL provides physically isolated and tailored network resources for training tasks by reconfiguring the optical switch using LCoS-WSS technology. The ODDL topology also uses tunable transceivers to adapt to time-varying traffic patterns. To achieve accurate and fine-grained scheduling of optical circuits, we propose an efficient distributed control scheme that incurs minimal delay overhead. Our evaluation on real-world traces showcases ODDL’s remarkable performance. When implemented with 1024 nodes and 100 Gbps bandwidth, ODDL accelerates VGG19 training by 1.6× and 1.7× compared to conventional fat-tree electrical networks and photonic SiP-Ring architectures, respectively. We further build a four-node testbed, and our experiments show that ODDL can achieve comparable training time compared to that of an ideal electrical switching network.


  • Symmetry-free algorithm for spectrum allocation: parallel implementations and evaluation

    by George N. Rouskas, Shubham Gupta, Priya Sharma

    Our symmetry-free model for spectrum allocation (SA) in networks of general topology leverages two properties: (1) SA is equivalent to a connection permutation problem, and (2) in assigning spectrum, it is sufficient to consider the allocation made by the first-fit (FF) algorithm. This model opens up algorithmic approaches that altogether sidestep spectrum symmetry, i.e., eliminate from consideration the exponential number of equivalent solutions resulting from spectrum slot permutations. Recursive FF (RFF) is such an algorithm; it applies FF recursively to search the connection permutation space and solve the SA problem optimally. Moreover, parallelism is inherent in the spectrum symmetry-free model, as the connection permutation space may be naturally decomposed into non-overlapping subsets that can be searched independently. Accordingly, RFF admits multi-threaded implementations that may be tailored to the computing environment at hand. In this work, we present two strategies for parallelizing the execution of RFF, and we evaluate them experimentally using a comprehensive set of metrics. Our experiments indicate that RFF explores a vast number of symmetry-free solutions, and for moderate-sized networks, it takes mere seconds to yield solutions that are either optimal or very close to the lower bound.

  • Open Access

    Qubit Recycling in Entanglement Distillation

    by Stuart Pelletier, Ruozhou Yu, George Rouskas, Jianqing Liu

    Quantum entanglement distillation is a process to extract a small number of high-fidelity entanglement from a large number of low-fidelity ones, which in essence is to trade yield (or survival rate) for fidelity. Among existing distillation approaches, Gisin's local filtering protocol is commonly adopted in photonic quantum systems for distilling entangled photons in polarization basis. Yet, the performance of Gisin's filter is cursed by the same fundamental trade-off between fidelity and yield. To address this challenge, in this work, we propose a protocol to recycle the disposed photons and improve their fidelity by a designed (and optimized) local operator. The key parameters of the proposed protocol are calculated by solving a constrained optimization problem. In so doing, we achieve significantly higher yield of high-fidelity entanglement pairs. We further evaluate the performance of our designed protocol under two common configurations of Gisin's filter, namely full filter and partial filter. Compared with existing distillation protocols, the results demonstrate that our design achieves as much as 31.2% gain in yield under the same fidelity, while only incurring moderate system complexity in terms of invested hardware and extra signaling for synchronization.

  • Recursive first fit: a highly parallel optimal solution to spectrum allocation

    by George N. Rouskas, Chaitanya Bandikatla

    We revisit the classical spectrum allocation (SA) problem, a fundamental subproblem in optical network design, and make three contributions. First, we show how some SA problem instances may be decomposed into smaller instances that may be solved independently without loss of optimality. Second, we prove an optimality property of the well-known first-fit (FF) heuristic. Finally, we leverage this property to develop a recursive and parallel algorithm that applies the FF heuristic to find an optimal solution efficiently. This recursive FF algorithm is highly scalable because of two unique properties: (1) it completely sidesteps the symmetry inherent in SA and hence drastically reduces the solution space compared to typical integer linear programming formulations, and (2) the solution space can be naturally decomposed in non-overlapping subtrees that may be explored in parallel almost independently of each other, resulting in faster than linear speedup.


  • Scalable Recursive First Fit: An Optimal Solution to the Spectrum Allocation Problem

    by George N. Rouskas, Chaitanya Bandikatla

    We revisit the classical spectrum allocation (SA) problem, a fundamental subproblem in optical network design, and make three contributions. First, we show how some SA problem instances may be decomposed into smaller instances that may be solved independently without loss of optimality. Second, we prove an optimality property of the well-known first-fit (FF) heuristic. Finally, we leverage this property to develop a recursive algorithm that applies the FF heuristic to find an optimal solution efficiently. This recursive first-fit (Rec-FF) algorithm complements our recent algorithm that recursively searches the routing space, and may be combined with it to solve large routing and spectrum allocation (RSA) problem instances to optimality.


  • Crosstalk-Aware Shared Backup Path Protection in Multi-Core Fiber Elastic Optical Networks

    by Fengxian Tang, Gangxiang Shen, George N. Rouskas

    Elastic optical networks employing multi-core fibers (MCF-EON) have the potential to expand significantly the transmission capacity of optical transport. However, wide deployment of such networks depends on addressing effectively two critical challenges: inter-core crosstalk, which may cause serious signal performance degradation in an MCF link, and survivability against network failures that may cause enormous data loss. In this article, we consider the design of MCF-EONs with shared-backup path protection (SBPP), one of the most efficient techniques for protecting network traffic. Specifically, we tackle the crosstalk-aware routing, core, and spectrum assignment (CA-RCSA) problem with the objective of jointly minimizing the network spectrum resources used and the total inter-core crosstalk. We formulate the problem as an integer linear programming (ILP) model subject to strict inter-core crosstalk limits for each provisioned lightpath, and we also propose an auxiliary graph (AG) based heuristic algorithm for lightpath provisioning. Simulation studies show that our algorithm is effective in terms of the objectives, and it is efficient to perform close to the ILP model in small networks, for which solving the ILP is feasible.


  • Congestion Minimization for Service Chain Routing Problems With Path Length Considerations

    by Lingnan Gao, George N. Rouskas

    Network function virtualization (NFV), with its perceived potential to accelerate service deployment and to introduce flexibility in service provisioning, has drawn a growing interest from industry and academia alike over the past few years. One of the key challenges in realizing NFV is the service chain routing problem, whereby traffic must be routed so as to traverse the various components of a network service that have been mapped onto the underlying network. In this work, we consider the online service chain routing problem. We route the service chain with the goal of jointly minimizing the maximum network congestion and the number of hops from the source to the destination. To this end, we present a simple yet effective online algorithm in which the routing decision is irrevocably made without prior knowledge of future requests. We prove that our algorithm is O(log m)-competitive in terms of congestion minimization, where m is the number of edges of the underlying network topology, and we show that this ratio is asymptotically optimal.