pith. sign in

arxiv: 1907.04784 · v1 · pith:VL3DBLJ6new · submitted 2019-07-10 · 💻 cs.NI · eess.SP

AWG-based Nonblocking Shuffle-Exchange Networks

Tong Ye , Jingjie Ding , Tony Tong Lee , Guido Maier This is my paper

Pith reviewed 2026-05-24 23:30 UTC · model grok-4.3

classification 💻 cs.NI eess.SP
keywords AWGshuffle-exchange networkWDMnonblocking routingself-routingoptical networkswavelength converterinterconnection networks
0
0 comments X

The pith

An AWG is functionally equivalent to a classical shuffle network due to its wavelength routing property.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper establishes that arrayed waveguide gratings can serve as the core of optical shuffle-exchange networks. By matching the wavelength routing of AWGs to the shuffle permutation, the design uses small AWGs and tunable wavelength converters to create large WDM networks. The approach preserves self-routing and nonblocking properties from classical designs. A sympathetic reader would care because it enables modular, scalable optical interconnects that fully utilize wavelength channels without needing large custom components.

Core claim

According to the wavelength routing property of AWGs, an AWG is functionally equivalent to a classical shuffle network by nature. Based on this, a systematic method designs large-scale WDM shuffle networks using sets of small-size AWGs with the same wavelength set. Combining these with TWCs of small conversion range yields an AWG-based WDM SEN that is scalable and achieves 100% utilization when all input wavelength channels are busy. The self-routing property and nonblocking routing conditions of classical SENs are preserved.

What carries the argument

The wavelength routing property of AWGs, which produces the exact shuffle permutation for valid wavelength assignments.

If this is right

  • Large-scale WDM SENs can be constructed modularly from small AWGs.
  • The network achieves full 100% channel utilization.
  • Self-routing is maintained without additional complexity.
  • Nonblocking conditions from classical SENs apply directly.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This equivalence might allow easier integration of optical networks in existing WDM systems.
  • Future work could test the design with real AWG hardware for routing delays.
  • Connections to other permutation networks in optical computing could be explored.
  • Scalability in size and wavelength count could be quantified in prototypes.

Load-bearing premise

That the AWG's wavelength routing produces an exact match to the shuffle permutation without extra blocking from specific WDM mappings or TWC conversion limits.

What would settle it

Finding a wavelength assignment or input pattern where the AWG-based SEN blocks a valid connection that a classical SEN would route.

Figures

Figures reproduced from arXiv: 1907.04784 by Guido Maier, Jingjie Ding, Tong Ye, Tony Tong Lee.

Figure 1
Figure 1. Figure 1: An 8 × 8 shuffle exchange network. can provide large switching capacity [4], [9]–[11], especially when it employs wavelength division multiplexing (WDM) technology [12]. Second, the structure of SENs is very regular and thus easy to deploy [4], [13]. Third, the SEN has a less network diameter than other networks, and thus less implementation cost [3]. Fourth, the routing algorithm of the SEN is simple sinc… view at source ↗
Figure 2
Figure 2. Figure 2: An example of 18 × 18 generalized shuffle N (3, 6). approach to construct a modular AWG-based WDM shuffle network in Section III. Based on the result in Section III, we design AWG-based WDM SENs in Section IV. At last, Section V shows that the self-routing property and the nonblocking conditions of classical SENs are also preserved by such WDM SENs. Finally, Section VI concludes this paper. II. GENERALIZED… view at source ↗
Figure 3
Figure 3. Figure 3: Equivalence between a single AWG and a shuffle: (a) AWG [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Modular AWG-based shuffle network (a) W(3, 6) and (b) its equivalent shuffle network N (3, 6). AWG a = 1. In other words, this connection goes through the following path: input a = 1 of input group p = 1 → input p = 1 of AWG a = 1 → input q 0 = 1 of AWG a = 1. According to wavelength routing property (1), input p = 1 of AWG 1 links to output q 0 = 1 of AWG 1 via wavelength λi , where i = [p + q 0 ]m = [1 +… view at source ↗
Figure 5
Figure 5. Figure 5: Illustration of TWC-modules: (a) a 3 × 3 TWC-module and (b) its space representation. A. mn × mn AWG-based Shuffle Network The mn × mn shuffle network W(m, mn−1 ) is a shuffle network W(m, rm) with r = mn−2 . Since there are m input groups, mn−2 inputs in each group, and m wavelength channels at each input in W(m, mn−1 ), for an input wave￾length channel X = paq 0 , the address X can be relabeled by an m-a… view at source ↗
Figure 6
Figure 6. Figure 6: Illustration of AWG-based WDM SENs: (a) An AWG-based WDM SEN [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The kth shuffle stage Wk. m × m TWC-modules in S(m, n), and each column contains mn−1 TWC-modules. An example is the 3 3 ×3 3 AWG-based WDM SEN S(3, 3) in [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Route of multi-requests without contention in [PITH_FULL_IMAGE:figures/full_fig_p011_8.png] view at source ↗
read the original abstract

Optical shuffle-exchange networks (SENs) have wide application in different kinds of interconnection networks. This paper proposes an approach to construct modular optical SENs, using a set of arrayed waveguide gratings (AWGs) and tunable wavelength converters (TWCs). According to the wavelength routing property of AWGs, we demonstrate for the first time that an AWG is functionally equivalent to a classical shuffle network by nature. Based on this result, we devise a systematic method to design a large-scale wavelength-division-multiplexing (WDM) shuffle network using a set of small-size AWGs associated with the same wavelength set. Combining the AWG-based WDM shuffle networks and the TWCs with small conversion range, we finally obtain an AWG-based WDM SEN, which not only is scalable in several ways, but also can achieve 100% utilization when the input wavelength channels are all busy. We also study the routing and wavelength assignment (RWA) problem of the AWG-based WDM SEN, and prove that the self-routing property and the nonblocking routing conditions of classical SENs are preserved in such AWG-based WDM SEN.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes constructing modular optical shuffle-exchange networks (SENs) using arrays of arrayed waveguide gratings (AWGs) combined with tunable wavelength converters (TWCs). It claims to demonstrate for the first time that an AWG is functionally equivalent to a classical shuffle network due to its wavelength routing property, enabling a systematic design of large-scale wavelength-division-multiplexing (WDM) shuffle networks from small AWGs sharing the same wavelength set. The resulting AWG-based WDM SEN is asserted to be scalable, achieve 100% utilization when all input channels are busy, and preserve the self-routing property and nonblocking routing conditions of classical SENs, with the routing and wavelength assignment (RWA) problem inheriting the classical nonblocking conditions.

Significance. If the claimed functional equivalence holds and the nonblocking properties are rigorously preserved without additional constraints from WDM channel mappings or TWC limits, the work would provide a concrete, modular construction for scalable optical interconnection networks. This could enable practical implementations of large SENs in optical computing and high-performance networking, with the 100% utilization and self-routing preservation offering clear engineering advantages over prior optical SEN designs.

major comments (2)
  1. [Abstract / equivalence derivation section] Abstract and the section deriving the AWG-shuffle equivalence: the central claim that the AWG's fixed cyclic wavelength routing produces an exact functional match to the classical shuffle permutation for any valid wavelength assignment is load-bearing for all subsequent results on scalability and nonblocking. The manuscript must exhibit the explicit permutation mapping (input port + wavelength → output port) and prove it coincides with the shuffle-exchange permutation independently of the chosen WDM channel set; without this, the assertion that nonblocking conditions carry over cannot be assessed.
  2. [WDM SEN construction and RWA section] The section combining AWG-based WDM shuffle stages with TWCs: the claim that the construction achieves 100% utilization and preserves classical nonblocking conditions assumes that the finite conversion range of the TWCs does not introduce new blocking states. The manuscript should verify that every wavelength conversion required by the inherited nonblocking routing conditions lies within the stated small conversion range; otherwise the 100% utilization guarantee fails for some traffic patterns.
minor comments (2)
  1. Notation for wavelength sets and port indexing should be defined once at the beginning and used consistently; several passages reuse symbols without redefinition.
  2. Figure captions for the AWG routing diagrams should explicitly label the input-port/wavelength pairs and resulting output ports to allow direct comparison with the claimed shuffle permutation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful review and constructive comments. We respond to each major comment below. The referee correctly identifies areas where explicit mappings and verifications would strengthen the manuscript, and we will revise accordingly.

read point-by-point responses

  1. Referee: [Abstract / equivalence derivation section] Abstract and the section deriving the AWG-shuffle equivalence: the central claim that the AWG's fixed cyclic wavelength routing produces an exact functional match to the classical shuffle permutation for any valid wavelength assignment is load-bearing for all subsequent results on scalability and nonblocking. The manuscript must exhibit the explicit permutation mapping (input port + wavelength → output port) and prove it coincides with the shuffle-exchange permutation independently of the chosen WDM channel set; without this, the assertion that nonblocking conditions carry over cannot be assessed.

    Authors: The manuscript derives the equivalence from the AWG wavelength routing property implementing a cyclic shift. We agree an explicit mapping would make the independence from specific WDM channel sets clearer. In revision we will add the permutation formula (input port i and wavelength index j maps to output port (i + j) mod N) together with a short proof that the mapping is identical to the classical shuffle for any equally spaced wavelength set, since AWG routing depends only on relative indices. revision: yes

  2. Referee: [WDM SEN construction and RWA section] The section combining AWG-based WDM shuffle stages with TWCs: the claim that the construction achieves 100% utilization and preserves classical nonblocking conditions assumes that the finite conversion range of the TWCs does not introduce new blocking states. The manuscript should verify that every wavelength conversion required by the inherited nonblocking routing conditions lies within the stated small conversion range; otherwise the 100% utilization guarantee fails for some traffic patterns.

    Authors: The manuscript asserts that the RWA inherits classical nonblocking conditions while using TWCs of small range. We agree that an explicit verification is needed to confirm no new blocking arises from the limited range. In the revision we will add a short analysis showing that all conversions required by the inherited SEN nonblocking conditions fall inside the stated small conversion range, thereby preserving the 100% utilization claim. revision: yes

Circularity Check

0 steps flagged

No significant circularity; equivalence derived from independent physical property

full rationale

The paper's core claim of functional equivalence between AWG and classical shuffle network is explicitly grounded in the external wavelength routing property of AWGs (an established physical fact), not in any self-definition, fitted parameter, or self-citation chain. The subsequent preservation of self-routing and nonblocking conditions follows from combining this equivalence with TWC conversion ranges, without the derivation reducing to its own inputs by construction. No equations, ansatzes, or load-bearing self-citations appear in the provided text that would trigger the enumerated circularity patterns. This is the common case of a modeling step justified by independent domain knowledge.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The claims rest on the domain assumption that AWG wavelength routing produces an exact shuffle permutation and that TWCs with small range suffice to complete the SEN without introducing new blocking; no free parameters or new entities are introduced in the abstract.

axioms (1)
  • domain assumption Wavelength routing property of AWGs produces a functional equivalent to a classical shuffle network
    Invoked to establish the core equivalence result stated in the abstract.

pith-pipeline@v0.9.0 · 5730 in / 1330 out tokens · 26981 ms · 2026-05-24T23:30:58.338973+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages

  1. [1]

    A cost-effective architecture for optical multistage interconnection network,

    M. Mehrnaz and M. Othman, “A cost-effective architecture for optical multistage interconnection network,” Journal of Networks , vol. 8, pp. 345–350, Feb. 2013

    work page 2013

  2. [2]

    A novel resource scheduling approach to improve the reliability of shuffle-exchange networks,

    F. Mashhadi, A. Asaduzzaman, and M. F. Mridha, “A novel resource scheduling approach to improve the reliability of shuffle-exchange networks,” in Proc. IEEE icIVPR , Feb. 2017, pp. 1–6

    work page 2017

  3. [3]

    The shuffle- exchange mesh topology for 3D NoCs,

    A. Sharifi, R. Sabbaghi-Nadooshan, and H. Sarbazi-Azad, “The shuffle- exchange mesh topology for 3D NoCs,” in Proc. International Sympo- sium on Parallel Architectures, Algorithms, and Networks (PAAP) , May 2008, pp. 275–280

    work page 2008

  4. [4]

    A recon- figurable, regular-topology cluster/datacenter network using commodity optical switches,

    D. Lugones, K. Katrinis, G. Theodoropoulos, and M. Collier, “A recon- figurable, regular-topology cluster/datacenter network using commodity optical switches,” Future Generation Computer Systems, pp. 78–89, Apr. 2014

    work page 2014

  5. [5]

    Novel intra- and inter-datacenter converged network exploiting space- and wavelength- dimensional switches,

    K. Ueda, Y . Mori, H. Hasegawa, and K. Sato, “Novel intra- and inter-datacenter converged network exploiting space- and wavelength- dimensional switches,” in Proc. OFC, Mar. 2017, pp. 1–3

    work page 2017

  6. [6]

    Study on big data center traffic management based on the separation of large-scale data stream,

    H. W. Park, I. Y . Yeo, J. R. Lee, and H. Jang, “Study on big data center traffic management based on the separation of large-scale data stream,” in Proc. Conference on Innovative Mobile and Internet Services in Ubiquitous Computing , Jul. 2017, pp. 591–594

    work page 2017

  7. [7]

    Towards petabit/s all-optical flat data center networks based on WDM optical cross-connect switches with flow control,

    W. Miao, F. Yan, and N. Calabretta, “Towards petabit/s all-optical flat data center networks based on WDM optical cross-connect switches with flow control,” J. Lightw. Technol., vol. 34, pp. 4066–4075, Sep. 2016

    work page 2016

  8. [8]

    Constructing transition ma- trices for routing BPC permutations on shuffle-exchange recirculating networks,

    G. Veselovsky and S. Ackarachalanonth, “Constructing transition ma- trices for routing BPC permutations on shuffle-exchange recirculating networks,” in Proc. International Conference on Digital Information and Communication Technology and it’s Applications (DICTAP), May 2012, pp. 378–382

    work page 2012

  9. [9]

    The development of crosstalk- free scheduling algorithms for routing in optical multistage interconnec- tion networks,

    M. Othman and T. D. S. R. M. Auzar, “The development of crosstalk- free scheduling algorithms for routing in optical multistage interconnec- tion networks,” in Trends in Telecommunications Technologies . Inte- chOpen, 2010, ch. 29

    work page 2010

  10. [10]

    Integrating RLP and fast zero algorithm to improve routing performance in optical multistage interconnection networks,

    T. D. Shahida, M. Othman, and M. Khazani, “Integrating RLP and fast zero algorithm to improve routing performance in optical multistage interconnection networks,” in Proc. International Symposium on High Capacity Optical Networks and Enabling Technologies , Nov. 2008, pp. 34–38

    work page 2008

  11. [11]

    Link and switch conflicts-free routing in optical multistage interconnection networks,

    S. Arya and Nipur, “Link and switch conflicts-free routing in optical multistage interconnection networks,” in Proc. International Conference on Next Generation Computing Technologies (NGCT) , Sep. 2015, pp. 312–315

    work page 2015

  12. [12]

    Experimental demonstration of multidimensional switching nodes for all-optical data center networks,

    V . Kamchevska, A. K. Medhin, F. D. Ros, F. Ye, R. Asif, A. M. Fagertun, S. Ruepp, M. Berger, L. Dittmann, T. Morioka, L. K. Oxenlwe, and M. Galili, “Experimental demonstration of multidimensional switching nodes for all-optical data center networks,” J. Lightw. Technol., vol. 34, pp. 1837–1843, Apr. 2016

    work page 2016

  13. [13]

    Optical omega network: a compact implementation technique,

    K. W. Wong and L. M. Cheng, “Optical omega network: a compact implementation technique,” Appl. Opt., vol. 34, no. 29, pp. 6648–6651, Oct. 1995

    work page 1995

  14. [14]

    On shortest path routing in single stage shuffle-exchange networks,

    K. Sunil and V . V . Alexander, “On shortest path routing in single stage shuffle-exchange networks,” in Proc. Annual ACM Symposium on Parallel Algorithms and Architectures, May 1995, pp. 298–307

    work page 1995

  15. [15]

    T. T. Lee and S. C. Liew, Principle of broadband switching and networking. Wiley-Interscience, 2010

    work page 2010

  16. [16]

    An improved simulated annealing algorithm to avoid crosstalk in optical omega network,

    M. Abdullah, M. Othman, and R. Johari, “An improved simulated annealing algorithm to avoid crosstalk in optical omega network,” in Proc. International Conference on Computing Informatics , Jun. 2006, pp. 1–7

    work page 2006

  17. [17]

    AWG-based shuffle-exchange optical-interconnection architecture,

    G. Maier, P. Valzasina, and A. Pattavina, “AWG-based shuffle-exchange optical-interconnection architecture,” in Proc. International Conference on Computer Communications and Networks (ICCCN) , Jul. 2011, pp. 1–6

    work page 2011

  18. [18]

    Error-free 320-gb/s all-optical wavelength conversion using a single semiconductor optical amplifier,

    Y . Liu, E. Tangdiongga, Z. Li, H. de Waardt, A. M. J. Koonen, G. D. Khoe, X. Shu, I. Bennion, and H. J. S. Dorren, “Error-free 320-gb/s all-optical wavelength conversion using a single semiconductor optical amplifier,” J. Lightw. Technol., vol. 25, pp. 103–108, Jan. 2007

    work page 2007

  19. [19]

    Simple optical fabrics for scalable terabit packet switches,

    R. Gaudino, G. A. G. Castillo, F. Neri, and J. M. Finochietto, “Simple optical fabrics for scalable terabit packet switches,” in Proc. IEEE ICC, 2008, pp. 5331–5337

    work page 2008

  20. [20]

    Non-blocking wdm switches based on arrayed waveguide grating and shared wavelength conversion,

    A. Pattavina and R. Zanzottera, “Non-blocking wdm switches based on arrayed waveguide grating and shared wavelength conversion,” in Proc. IEEE INFOCOM, Apr. 2006, pp. 1–11

    work page 2006

  21. [21]

    Design and analysis of optical flow-switched networks,

    G. Weichenberg, V . W. S. Chan, and M. Mdard, “Design and analysis of optical flow-switched networks,” IEEE J. Opt. Commun. Netw. , vol. 1, pp. 1943–0620, Aug. 2009

    work page 1943

  22. [22]

    Low-cost, degree-expandable and contention-free ROADM architecture based on M ×N WSS,

    L. Zong, H. Zhao, Z. Feng, and Y . Yan, “Low-cost, degree-expandable and contention-free ROADM architecture based on M ×N WSS,” in Proc. OFC, Mar. 2016, pp. 1–3

    work page 2016

  23. [23]

    Optical implementations of the perfect shuffle interconnection,

    K.-H. Brenner and A. Huang, “Optical implementations of the perfect shuffle interconnection,” Appl. Opt. , vol. 27, no. 1, pp. 135–137, Jan. 1988

    work page 1988

  24. [24]

    Folded perfect shuffle optical processor,

    C. W. Stirk, R. A. Athale, and M. W. Haney, “Folded perfect shuffle optical processor,” Appl. Opt., vol. 27, no. 2, pp. 202–203, Jan. 1988

    work page 1988

  25. [25]

    Design of micro-optics array to realize two dimensional perfect shuffle transform,

    J. Yang, Z. Shao, K. Zhou, J. Xu, and P. Xu, “Design of micro-optics array to realize two dimensional perfect shuffle transform,” Optical Switching and Networking , vol. 12, no. 2, pp. 68–79, 2014

    work page 2014

  26. [26]

    Passive optical interconnection network employing a shuffle–exchange topology,

    C. Waterson and B. K. Jenkins, “Passive optical interconnection network employing a shuffle–exchange topology,” Appl. Opt., vol. 33, no. 8, pp. 1575–1586, Mar. 1994

    work page 1994

  27. [27]

    Optical perfect-shuffle-exchange interconnection network using a liquid-crystal spatial light switch,

    M. Cao, F. Luo, H. Li, and S. Wang, “Optical perfect-shuffle-exchange interconnection network using a liquid-crystal spatial light switch,” Appl. Opt., vol. 31, no. 32, pp. 6817–6819, Nov. 1992

    work page 1992

  28. [28]

    A survey of free space optical networks,

    I. K. Son and S. Mao, “A survey of free space optical networks,” Digital Communications and Networks , vol. 3, no. 2, pp. 67 – 77, 2017

    work page 2017

  29. [29]

    Optimal all-to-all personalized exchange in a class of optical multistage networks,

    Y . Yang and J. Wang, “Optimal all-to-all personalized exchange in a class of optical multistage networks,” IEEE Trans. Parallel Distrib. Syst., vol. 12, no. 6, pp. 567 – 582, 2001

    work page 2001

  30. [30]

    Optical interconnection networks in data centers: recent trends and future challenges,

    C. Kachris, K. Kanonakis, and I. Tomkos, “Optical interconnection networks in data centers: recent trends and future challenges,” IEEE Commun. Mag., vol. 51, no. 9, pp. 39–45, Sep. 2013

    work page 2013

  31. [31]

    AWG-based non-blocking Clos networks,

    T. Ye, T. T. Lee, and W. Hu, “AWG-based non-blocking Clos networks,” IEEE/ACM Trans. Netw., vol. 23, no. 2, pp. 491–504, Apr. 2015

    work page 2015

  32. [32]

    Multicast routing and wavelength assignment in AWG-based Clos networks,

    M. Ge, T. Ye, T. T. Lee, and W. Hu, “Multicast routing and wavelength assignment in AWG-based Clos networks,” IEEE/ACM Trans. Netw. , vol. 25, no. 3, pp. 1892–1909, Jun. 2017

    work page 1909

  33. [33]

    AWG-based architecture for optical interconnection in asynchronous systems,

    D. Lucerna, G. Maier, and A. Pattavina, “AWG-based architecture for optical interconnection in asynchronous systems,” in Proc. IEEE HPSR, 2011, pp. 250–255

    work page 2011

  34. [34]

    Testa and L

    F. Testa and L. Pavesi, Optical Switching in Next Generation Data Centers. Springer, Cham, 2018, ch. 2

    work page 2018

  35. [35]

    A large-scale wavelength routing optical switch for data center networks,

    K. Sato, H. Hasegawa, T. Niwa, and T. Watanabe, “A large-scale wavelength routing optical switch for data center networks,” IEEE Commun. Mag., vol. 51, no. 9, pp. 46–52, Sep. 2013

    work page 2013

  36. [36]

    N× N cyclic-frequency router with improved performance based on arrayed- waveguide grating,

    S. Kamei, M. Ishii, A. Kaneko, T. Shibata, and M. Itoh, “ N× N cyclic-frequency router with improved performance based on arrayed- waveguide grating,” J. Lightw. Technol., vol. 27, pp. 4097–4104, Sep. 2009

    work page 2009

  37. [37]

    Experimental demonstration of flex-LIONS for reconfigurable all-to-all optical inter- connects,

    X. Xiao, R. Proietti, K. Zhang, and S. J. B. Yoo, “Experimental demonstration of flex-LIONS for reconfigurable all-to-all optical inter- connects,” in Proc. ECOC, 2018, pp. 1–3

    work page 2018