Pith. sign in

REVIEW 3 major objections 6 minor 54 references

Unlocking Diversity of Fast-Switched Optical Data Center Networks with Unified Routing

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read URO claims that a single offline routing framework can make fast-switched optical data center networks work even when circuit time slices are shorter than a packet's own one-way delay, and it demonstrates this on commodity programmable…

desk verdict A real prototype and thorough simulations make URO worth reading, but the sub-OWD claim rests on an assumption that breaks in the very regime the paper motivates. read the letter →

arxiv 2412.00266 v1 pith:S6DXCS37 submitted 2024-11-29 cs.NI

Jialong Li , Federico De Marchi , Yiming Lei , Raj Joshi , Balakrishnan Chandrasekaran , Yiting Xia This is my paper · ORCID
classification cs.NI
keywords opticaldatacenternetworksfast-switchedcircuitstimeslicesroutingprogrammableswitchescalendarqueuesValiantloadbalancingflowcompletion
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

URO (Unified Routing for Optical networks) is a routing framework for fast-switched optical data center networks, where optical circuits are reconfigured at microsecond or even nanosecond scale and a time slice can be shorter than a packet's one-way delay. Because the optical schedule is known in advance, URO computes paths offline so that packets know where to wait, and programmable switches buffer packets in calendar queues that are paused and resumed on a per-slice beat. The authors argue that this single design covers a wide range of optical hardware, from very short sub-one-way-delay slices to long slices where it reduces to k-shortest-path routing. They demonstrate a prototype on Intel Tofino2 switches with a minimum circuit duration of 2 microseconds and loss-free end-to-end application traffic, and simulations against Opera, Sirius, and VBS show lower latency and shorter paths at microsecond-scale slices. The payoff is that network architects no longer need to co-design routing with each new optical hardware point; one routing layer can serve many.

What carries the argument

The carrying mechanism is an offline backtracking algorithm over a time-varying graph. The graph has top-of-rack switches as vertices and optical circuits as edges labeled with the time slices when they exist; path latency is defined as the number of elapsed time slices times the slice duration. ROUTING finds the earliest last-hop circuit to the destination, then SUBPATH recursively searches backward for the shortest feasible chain of intermediate ToRs whose circuit times are nondecreasing, so each hop's departure is no later than the next hop's circuit time. Per-hop lookup tables let each ToR act on only the next hop while preserving optimality, and calendar queues, physical queues per egress port cyclically assigned to time slices, implement the waiting; an on-chip packet generator triggers queue pause and resume each slice. When a packet misses its planned slice, a one-shot lookup consults a queue-state bit array to reroute onto the next optimal path without recirculation.

What would settle it

Load a URO testbed with a burst that makes a scheduled departure queue genuinely non-empty, then measure whether the rerouted packet's actual end-to-end latency ever exceeds the latency of the path that would have been chosen with accurate queue information. If such a case exists, the optimality claim under real queueing is refuted; if none exists, the zero-queue simplification holds under the tested regime.

Watch

Extended reading notes

Core claim

The paper claims that a single offline routing algorithm can make fast-switched optical data center networks work even when a circuit's time slice is shorter than a packet's one-way delay. Because the optical schedule is known ahead of time, URO computes the latency-minimizing path for each source-destination pair per time slice, allowing packets to stop and wait at intermediate top-of-rack switches until the next needed optical circuit appears. On Intel Tofino2 switches the authors implement this waiting as time-synchronized queue pausing and report a minimum circuit duration of 2 microseconds with loss-free application performance; in simulations with production DCN traffic, URO shortens paths and reduces flow completion times compared to the Opera architecture and to VLB-based designs at microsecond-scale slices.

Load-bearing premise

The whole path plan assumes packets arrive exactly at the start of a time slice and that top-of-rack switch queues are empty; if real queueing builds up, the precomputed fastest route may no longer be fastest, and the system's rerouting only re-optimizes under that same idealized model.

Editorial extensions

If this is right

  • Fast-switched architectures with sub-one-way-delay slices become deployable on commodity programmable switches rather than requiring custom silicon.
  • One routing layer can replace architecture-specific routing co-designs, so new optical hardware can be adopted without building a new network stack.
  • Mice flows get lower latency and shorter paths than continuous-path designs like Opera, while elephant flows can still be offloaded to Valiant load balancing for throughput.
  • At slice durations above one-way delay the algorithm collapses to ordinary shortest-path routing, so the same framework covers both fast-switched and slow-switched regimes.
  • If packet-granularity nanosecond slices ever become practical in ToR hardware, URO's schedule-based offline approach reaches lower-bound flow completion times comparable to dedicated designs like Sirius and VBS.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's zero-queue assumption suggests a testable design point: URO's advantages should shrink under sustained high load where queues are non-empty, and one could extend the offline path computation to use predicted queue occupancy per slice rather than zero.
  • Because URO treats any known cyclic schedule the same way, it opens the door to schedule design as a control knob, where future work could jointly optimize schedules with URO's latency metric.
  • The one-shot rerouting with multiple precomputed backup paths turns link failures into a missed-slice event, suggesting URO's robustness could be quantified analytically as a function of schedule redundancy rather than only by simulation.
  • Since URO separates routing from architecture, it could act as a compatibility layer that lets two different optical fabrics be bridged by a common ToR software stack, an interoperability benefit the authors invoke but do not demonstrate.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper presents Unified Routing for Optical networks (URO), a routing framework intended to work across fast-switched optical data center networks whose time slices may be shorter than a packet's one-way delay. The core idea is to exploit the fact that the cyclic optical schedule is known a priori: a backtracking algorithm (Algorithm 1) computes, for each source-destination pair and arrival slice, a path that may wait (buffer) at intermediate ToRs, minimizing a slice-counting latency metric (Eqn. 1). Three properties are proved: the chosen path has minimal latency (Property 1), per-hop lookup tables preserve the optimal path (Property 2), and rerouting after a missed slice yields the next optimal path (Property 3). A prototype on Intel Tofino2 implements calendar queues with queue pause/resume and reports a minimum 2 us slice; large-scale htsim simulations on Microsoft production traces compare URO with Opera and Sirius/VBS, reporting up to 12.8x lower FCTs at 2 us slices and resource usage within commodity limits.

Significance. If the claims hold, URO is a valuable step toward decoupling routing from the specifics of optical hardware, an area where most systems are tightly co-designed (Opera, Sirius, VBS). The paper's strengths are its explicitly stated modeling assumptions, a real prototype (2 us slices on Tofino2 is an impressive data point for commodity switches), the breadth of the evaluation (production traces, multiple baselines, failure scenarios, resource accounting), and the open-framework positioning (Lighthouse). The three properties are hand-checkable and appear correct within the stated discrete model. The main risk is that the model omits per-hop propagation and transmission delays, and the paper's own evaluation enters a regime (0.12 us slices with 500 ns per-hop propagation) where that omission is most consequential; the quantitative sub-microsecond comparisons should therefore be treated with caution until the model is refined. The needed corrections are within the manuscript's scope, so the contribution remains substantial.

major comments (3)
  1. [§III-A, §III-C, Algorithm 1] The model omits per-hop propagation and transmission delays, and this omission is load-bearing exactly in the regime the paper targets. The feasibility check at Algorithm 1 lines 17-18, t(src,r) <= t, together with the latency definition in Eqn. (1), treats a packet as arriving at the next ToR at the start of its departure slice. The footnote in §III-A justifies this by saying these delays are 'substantially smaller than lat(p)', but this is contradicted by the paper's own settings: §VI-A sets per-hop propagation to 500 ns, while §VI-C evaluates 0.12 us slices and Table I lists 20-40 ns slice architectures. For u = 120 ns, a packet sent over S->R in slice 0 arrives at R at 500 ns, i.e., during slice 4, so a planned R->D circuit in slice 1 is missed and the path declared feasible by the algorithm is physically infeasible. Consequently, Properties 1-3 prove optimality only with respect to the discrete slice model, and in the sub-OWD regime the path actually taken is the product of runtime rerouting, whose 'optimality' is relative to the same idealized model. The fix is within reach: introduce a per-hop constant delay delta (propagation + transmission + guardband), replace the check with t(src,r) + ceil(delta/u) <= t, and either re-derive the properties under this refined model or explicitly restrict the optimality claims to slices with u >= delta. In particular, the 0.12 us lower bound in Fig. 11c should be recomputed, since with delta = 500 ns and u = 120 ns a single hop already costs several slices.
  2. [§IV-C; abstract] The abstract's claim that the prototype 'ensur[es] end-to-end, loss-free application performance' is stronger than what the system actually provides. In §IV-C, the one-shot lookup drops a packet when no departure queue is feasible, and the paper reports only that no drops were observed in the evaluation. Because the offline plan is built under the empty-queue assumption (§III-A) and the number of alternative paths stored per lookup entry is fixed, there is no argument ruling out drops under heavy bursts or adversarial arrivals; loss-freedom is an empirical observation, not a guarantee. Please state the load conditions under which loss-freedom is expected to hold, or soften the abstract and §I claims accordingly.
  3. [§III-B] The complexity analysis of the offline algorithm is not rigorous as written. The claimed reduction to O(N^M / 2^(M^2+M)) rests on the unexplained assertion that 'half of the nodes are filtered out at each level', which is schedule-dependent and not established; moreover, the product in the text (1 × N/2 × N/4 × ... × N/2^M) evaluates to N^M / 2^(M(M+1)/2), not N^M / 2^(M^2+M), so the stated exponent needs correction. The 'polynomial time' wording is also imprecise: for a fixed hop limit M, the worst-case O(N^M) is polynomial in N, but the heuristic reduction is what makes the 1024-ToR projections in Table II plausible, and the pre-computation time at that scale is not reported. Please present the complexity bound formally and report the achieved pre-computation time for the larger configurations.
minor comments (6)
  1. [§IV-B] There is a typo in §IV-B: 'if the departure time slice is later than the arrival time slice of' should read '...arrival time slice of the packet'.
  2. [Fig. 4] The walk-through of Fig. 4 is hard to follow because the path numbering in Fig. 4a and the table entries in Fig. 4b are not visually aligned in monochrome; a small table listing (arrival slice, destination, egress port, departure slice) per path would clarify the example.
  3. [§V] Both the 2 us minimum-slice derivation and the 50 ns queuing-delay estimation accuracy are deferred to the companion Lighthouse paper [33]; since these numbers support headline claims, please include the key derivation steps here or confirm the stable public availability of [33].
  4. [Table I] The caption says architectures supported by URO are 'highlighted in blue', but the table is not color-coded in this version; please mark the supported rows with a symbol or footnote.
  5. [Fig. 2 and Fig. 6] Several axis labels and legends (e.g., Fig. 2, Fig. 6) show the unit as 's' where the micro sign is missing; the units should render as 'us'.
  6. [§VI-C] The footnote that h=2 outperforms h=3 for VBS in this scaled network is worth a sentence in the main text, since Fig. 11b otherwise suggests monotone improvement with h.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: URO's optimality proof is relative to an explicit latency model and performance claims are validated against external baselines.

full rationale

The paper's derivation chain is a standard algorithm-design argument. The URO algorithm minimizes the latency metric defined in Eq. 1 under the explicitly stated assumptions of slice-aligned arrivals and zero queueing; Property 1 proves by a search-order contradiction that the algorithm attains this objective, and Properties 2 and 3 reduce to Property 1. This is a correctness proof relative to a stated model, not a case where the prediction is equivalent to the input. The evaluation is anchored to external systems (Opera, Sirius, VBS), production traces, and a physical Tofino2 testbed, so the central performance claims do not reduce to fitted parameters or to the paper's own definitions. Self-citations to Lighthouse and HOHO supply engineering details (2 us timing limit, 50 ns queue-delay estimation, time synchronization), but the 2 us capability is also demonstrated by the testbed in this paper, and none of these citations is the load-bearing step in the routing derivation. The concern that the model omits per-hop propagation in the sub-OWD regime is a modeling-validity/correctness issue, not circularity: it challenges whether the defined latency corresponds to physical latency, but it does not make a derived result equal to its input. Under the stated rules, that concern belongs to correctness risk rather than a circularity score.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central routing algorithm leans on the idealizing assumptions (a) and (b) from Section III-A; the feasibility of the system implementation leans on hardware capabilities and on the companion Lighthouse paper for the 2 us limit and delay-estimation accuracy. The algorithm itself introduces no free parameters beyond the max-hop bound M and the alpha-based flow cutoff, both tunable design knobs rather than fits to the headline results.

free parameters (4)
  • slowdown metric alpha = 1.5
    Sets the flow-size cutoff for offloading mice vs elephant flows to VLB; chosen after sensitivity experiments showing insensitivity in 1.4-1.7 (Section IV-D, Section VI-D). Affects FCT results but not the core routing algorithm.
  • max hop count M = not stated explicitly; results show paths up to 6 hops
    Bounding parameter of the SUBPATH backtracking search; the algorithm's time complexity and completeness depend on it (Section III-B). Not reported as a concrete value.
  • guardband duration = 200 ns
    Fixed guardband in the testbed to account for circuit reconfiguration; reduces duty cycle at 2 us slices (Section V).
  • VLB cutoff flow size = 5 MB for 2 us slices, 13 MB for 5 us slices
    Derived from alpha and slice duration; determines which flows use VLB (Section IV-D). Not a fit to outcome data, but a tunable threshold.
assumptions (5)
  • domain assumption The optical schedule is known a priori and repeats cyclically, with each ToR pair assigned at least one circuit per cycle.
    Stated in Section II-A; required for the offline routing computation and for the path latency model.
  • ad hoc to paper Packets always arrive at the beginning of a time slice and there is no queuing delay at ToRs.
    Explicit assumptions in Section III-A under which the URO latency model and optimality proofs (Properties 1-3) are derived. The runtime system tries to mitigate real queuing, but the theoretical guarantees do not extend to non-empty queues.
  • domain assumption Transmission and propagation delays are negligible compared with time slice duration.
    Stated in Section III-A with numbers for 100 Gbps and 100 m fiber. Valid for microsecond-scale slices used in the testbed, but invalid for the hypothetical 0.12 us comparisons where the 500 ns propagation delay dominates.
  • domain assumption ToR clocks are synchronized to the optical schedule at nanosecond precision.
    The lookup table matches on arrival time slice, requiring synchronization; the paper cites Lighthouse [33], [36] for this capability rather than describing it.
  • domain assumption Commodity programmable switches expose queue pause/resume, packet generators, and register bit arrays as described.
    The system design (Section IV-B, IV-C) depends on Tofino2 features; the paper cites vendor talks and prior papers.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unlocking Diversity of Fast-Switched Optical Data Center Networks with Unified Routing." pith.science (2026). https://pith.science/paper/S6DXCS37

@misc{pith2026241200266,
  author       = {Pith},
  title        = {Pith review of: Unlocking Diversity of Fast-Switched Optical Data Center Networks with Unified Routing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S6DXCS37}},
  note         = {Machine review of arXiv:2412.00266}
}
read the original abstract

Optical data center networks (DCNs) are emerging as a promising solution for cloud infrastructure in the post-Moore's Law era, particularly with the advent of 'fast-switched' optical architectures capable of circuit reconfiguration at microsecond or even nanosecond scales. However, frequent reconfiguration of optical circuits introduces a unique challenge: in-flight packets risk loss during these transitions, hindering the deployment of many mature optical hardware designs due to the lack of suitable routing solutions. In this paper, we present Unified Routing for Optical networks (URO), a general routing framework designed to support fast-switched optical DCNs across various hardware architectures. URO combines theoretical modeling of this novel routing problem with practical implementation on programmable switches, enabling precise, time-based packet transmission. Our prototype on Intel Tofino2 switches achieves a minimum circuit duration of 2us, ensuring end-to-end, loss-free application performance. Large-scale simulations using production DCN traffic validate URO's generality across different hardware configurations, demonstrating its effectiveness and efficient system resource utilization.

Figures

Figures reproduced from arXiv: 2412.00266 by the authors.

Figure 1
Figure 1. Illustration of an optical DCN. to the existing network stack. Consequently, this line of work requires a complete redesign of the network from scratch, making end-to-end deployment a long-term endeavor. Decades of optics research has made available a wide variety of optical technologies spanning a full spectrum of reconfiguration delays. Numerous network architectures have been proposed to explore different time sl… view at source ↗
Figure 2
Figure 2. URO tail FCTs assuming empty queues vs. 80 packets [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the URO backtracking algorithm. The [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Illustration of the URO switch system (§IV). The current time slice is 2 and the corresponding active queue is 0. (a) URO paths (§III-B). (b) T oR1’s lookup table (§IV-A) corresponding to the paths, and queuing delay estimation for packet rerouting (§IV-C). (c) Calenda…
Figure 5
Figure 5. Figure 5: ToR-to-ToR delay with different packet sizes. 10 1 10 2 FCT ( s) 0.20 0.40 0.60 0.80 0.99 CDF URO 50 s URO 25 s URO 10 s URO 2 s VLB 50 s VLB 25 s VLB 10 s VLB 2 s Opera 50 s [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 7
Figure 7. Figure 7: (a) FCTs and (b) throughput comparison with Opera [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 10
Figure 10. Figure 10: (a) Path lengths compared to Opera. (b) Path lengths [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]
Figure 9
Figure 9. Figure 9: FCTs comparison with Opera under different time slices [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 11
Figure 11. Figure 11: Theoretical lower-bound FCTs achieved by (a) Sirius and (b) VBS. (c-e) Theoretical lower-bound FCT comparison [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]
Figure 12
Figure 12. Figure 12: FCT comparison with VLB under (a) web search and [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]
Figure 14
Figure 14. Figure 14: FCTs when setting slowdown metric α to (a) 1.4, (b) 1.5, (c) 1.6, and (d) 1.7. 0 10 20 Max #queues per port 2 s ( = 1.6) 2 s ( = 1.5) 2 s ( = 1.4) 5 s ( = 1.6) 5 s ( = 1.5) 5 s ( = 1.4) 10 s ( = 1.6) 10 s ( = 1.5) 10 s ( = 1.4) 50 s ( =1.4-1.6) (a) 0 100 200 300 400 Q…
Figure 15
Figure 15. Figure 15: (a) Maximum (blue) and 99th percentile (red) number of queues per port. (b) Queue occupancy per port. The results for 300 µs are similar to those for 50 µs and are not included in the figure. queues reached in simulation by each setting during stress tests with 30% tr…
Figure 18
Figure 18. Figure 18: Impact by slowdown metric α. 0.5 0.6 0.7 0.8 0.9 1.0 Edge-disjoint path ratio 0.0 0.2 0.4 0.6 0.8 1.0 CDF Opera full-round-robin Full-round-robin 1 Full-round-robin 2 [PITH_FULL_IMAGE:figures/full_fig_p012_18.png]
Figure 17
Figure 17. Figure 17: Connectivity loss under failed (a) ToRs and (b) OCSes. [PITH_FULL_IMAGE:figures/full_fig_p012_17.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 51 canonical work pages

  1. [1]

    Helios: a hybrid electrical/optical switch architecture for modular data centers,

    N. Farrington, G. Porter, S. Radhakrishnan, H. H. Bazzaz, V . Subramanya, Y . Fainman, G. Papen, and A. Vahdat, “Helios: a hybrid electrical/optical switch architecture for modular data centers,” in Proceedings of the ACM SIGCOMM 2010 Conference , 2010, pp. 339–350

    work page 2010

  2. [2]

    c-through: Part-time optics in data centers,

    G. Wang, D. G. Andersen, M. Kaminsky, K. Papagiannaki, T. E. Ng, M. Kozuch, and M. Ryan, “c-through: Part-time optics in data centers,” in Proceedings of the ACM SIGCOMM 2010 Conference , 2010, pp. 327–338

    work page 2010

  3. [3]

    OSA: An optical switching architecture for data center networks with unprecedented flexibility,

    K. Chen, A. Singla, A. Singh, K. Ramachandran, L. Xu, Y . Zhang, X. Wen, and Y . Chen, “OSA: An optical switching architecture for data center networks with unprecedented flexibility,” IEEE/ACM Transactions on Networking, vol. 22, no. 2, pp. 498–511, 2013

    work page 2013

  4. [4]

    Jupiter evolving: transforming google’s datacenter network via optical circuit switches and software-defined networking,

    L. Poutievski, O. Mashayekhi, J. Ong, A. Singh, M. Tariq, R. Wang, J. Zhang, V . Beauregard, P. Conner, S. Gribble et al., “Jupiter evolving: transforming google’s datacenter network via optical circuit switches and software-defined networking,” in Proceedings of the ACM SIGCOMM 2022 Conference, 2022, pp. 66–85

    work page 2022

  5. [5]

    A tale of two topologies: Exploring convertible data center network architectures with flat-tree,

    Y . Xia, X. S. Sun, S. Dzinamarira, D. Wu, X. S. Huang, and T. E. Ng, “A tale of two topologies: Exploring convertible data center network architectures with flat-tree,” in Proceedings of the Conference of the ACM Special Interest Group on Data Communication , 2017, pp. 295–308

    work page 2017

  6. [6]

    Wave- cube: A scalable, fault-tolerant, high-performance optical data center architecture,

    K. Chen, X. Wen, X. Ma, Y . Chen, Y . Xia, C. Hu, and Q. Dong, “Wave- cube: A scalable, fault-tolerant, high-performance optical data center architecture,” in 2015 IEEE Conference on Computer Communications (INFOCOM). IEEE, 2015, pp. 1903–1911

    work page 2015

  7. [7]

    En- abling wide-spread communications on optical fabric with megaswitch

    L. Chen, K. Chen, Z. Zhu, M. Yu, G. Porter, C. Qiao, and S. Zhong, “En- abling wide-spread communications on optical fabric with megaswitch.” in NSDI, vol. 17, 2017, pp. 577–593

    work page 2017

  8. [8]

    Firefly: A reconfigurable wireless data center fabric using free-space optics,

    N. Hamedazimi, Z. Qazi, H. Gupta, V . Sekar, S. R. Das, J. P. Longtin, H. Shah, and A. Tanwer, “Firefly: A reconfigurable wireless data center fabric using free-space optics,” in Proceedings of the 2014 ACM conference on SIGCOMM , 2014, pp. 319–330

    work page 2014

Show all 54 references
  1. [9]

    Enabling topological flexibility for data centers using omniswitch

    Y . Xia, M. Schlansker, T. E. Ng, and J. Tourrilhes, “Enabling topological flexibility for data centers using omniswitch.” in HotCloud, 2015

    2015

  2. [10]

    Integrating microsecond circuit switching into the data center,

    G. Porter, R. Strong, N. Farrington, A. Forencich, P. Chen-Sun, T. Rosing, Y . Fainman, G. Papen, and A. Vahdat, “Integrating microsecond circuit switching into the data center,” ACM SIGCOMM Computer Communica- tion Review, vol. 43, no. 4, pp. 447–458, 2013

    2013

  3. [11]

    Rotornet: A scalable, low-complexity, optical datacenter network,

    W. M. Mellette, R. McGuinness, A. Roy, A. Forencich, G. Papen, A. C. Snoeren, and G. Porter, “Rotornet: A scalable, low-complexity, optical datacenter network,” in Proceedings of the Conference of the ACM Special Interest Group on Data Communication , 2017, pp. 267–280

    2017

  4. [12]

    Expanding across time to deliver bandwidth efficiency and low latency,

    W. M. Mellette, R. Das, Y . Guo, R. McGuinness, A. C. Snoeren, and G. Porter, “Expanding across time to deliver bandwidth efficiency and low latency,” in 17th USENIX Symposium on Networked Systems Design and Implementation (NSDI 20) , 2020, pp. 1–18

    2020

  5. [13]

    Sirius: A flat datacenter network with nanosecond optical switching,

    H. Ballani, P. Costa, R. Behrendt, D. Cletheroe, I. Haller, K. Jozwik, F. Karinou, S. Lange, K. Shi, B. Thomsen et al., “Sirius: A flat datacenter network with nanosecond optical switching,” in Proceedings of the Annual conference of the ACM Special Interest Group on Data Comm...

    2020

  6. [14]

    Pulse: optical circuit switched data center architecture operating at nanosecond timescales,

    J. L. Benjamin, T. Gerard, D. Lavery, P. Bayvel, and G. Zervas, “Pulse: optical circuit switched data center architecture operating at nanosecond timescales,” Journal of Lightwave Technology, vol. 38, no. 18, pp. 4906– 4921, 2020

    2020

  7. [15]

    Modoru: Clos nanosecond optical switching for distributed deep training,

    C. Wang, N. Yoshikane, D. Elson, Y . Wakayama, D. Soma, S. Beppu, and T. Tsuritani, “Modoru: Clos nanosecond optical switching for distributed deep training,” Journal of Optical Communications and Networking , vol. 16, no. 1, pp. A40–A52, 2023

    2023

  8. [16]

    Ramp: a flat nanosecond optical network and mpi operations for distributed deep learning systems,

    A. Ottino, J. Benjamin, and G. Zervas, “Ramp: a flat nanosecond optical network and mpi operations for distributed deep learning systems,” Optical Switching and Networking , vol. 51, p. 100761, 2024

    2024

  9. [17]

    Hybrid wavelength switched-tdma high port count all-optical data centre switch,

    A. C. Funnell, K. Shi, P. Costa, P. Watts, H. Ballani, and B. C. Thomsen, “Hybrid wavelength switched-tdma high port count all-optical data centre switch,” Journal of Lightwave Technology, vol. 35, no. 20, pp. 4438–4444, 2017

    2017

  10. [18]

    Potori: a passive optical top-of-rack interconnect architecture for data centers,

    Y . Cheng, M. Fiorani, R. Lin, L. Wosinska, and J. Chen, “Potori: a passive optical top-of-rack interconnect architecture for data centers,” Journal of Optical Communications and Networking , vol. 9, no. 5, pp. 401–411, 2017

    2017

  11. [19]

    Silicon photonic flex-lions for reconfigurable multi-gpu systems,

    M. Fariborz, X. Xiao, P. Fotouhi, R. Proietti, and S. B. Yoo, “Silicon photonic flex-lions for reconfigurable multi-gpu systems,” Journal of Lightwave Technology, vol. 39, no. 4, pp. 1212–1220, 2021

    2021

  12. [20]

    Projector: Agile reconfigurable data center interconnect,

    M. Ghobadi, R. Mahajan, A. Phanishayee, N. Devanur, J. Kulkarni, G. Ranade, P.-A. Blanche, H. Rastegarfar, M. Glick, and D. Kilper, “Projector: Agile reconfigurable data center interconnect,” in Proceedings of the 2016 ACM SIGCOMM Conference , 2016, pp. 216–229

    2016

  13. [21]

    Optimal oblivious reconfigurable networks,

    D. Amir, T. Wilson, V . Shrivastav, H. Weatherspoon, R. Kleinberg, and R. Agarwal, “Optimal oblivious reconfigurable networks,” in Proceedings of the 54th Annual ACM SIGACT Symposium on Theory of Computing , 2022, pp. 1339–1352

    2022

  14. [22]

    Extending optimal oblivious reconfigurable networks to all n,

    T. Wilson, D. Amir, V . Shrivastav, H. Weatherspoon, and R. Kleinberg, “Extending optimal oblivious reconfigurable networks to all n,” in 2023 Symposium on Algorithmic Principles of Computer Systems (APOCS) . SIAM, 2023, pp. 1–16

    2023

  15. [23]

    Sqr: In-network packet loss recovery from link failures for highly reliable datacenter networks,

    T. Qu, R. Joshi, M. C. Chan, B. Leong, D. Guo, and Z. Liu, “Sqr: In-network packet loss recovery from link failures for highly reliable datacenter networks,” in Proceedings of ICNP , 2019

    2019

  16. [24]

    Parking packet payload with p4,

    S. Goswami, N. Kodirov, C. Mustard, I. Beschastnikh, and M. Seltzer, “Parking packet payload with p4,” in Proceedings of CoNEXT , 2020

    2020

  17. [25]

    Advanced congestion & flow control with programmable switches,

    J. Lee, “Advanced congestion & flow control with programmable switches,” in P4 Expert Roundtable Series , 2020. [Online]. Available: https://bit.ly/3J8x7fw

    2020

  18. [26]

    Backpressure flow control,

    P. Goyal, P. Shah, K. Zhao, G. Nikolaidis, M. Alizadeh, and T. E. Anderson, “Backpressure flow control,” in Proceedings of NSDI , 2022

    2022

  19. [27]

    Gemini: Practical reconfigurable datacenter networks with topology and traffic engineering,

    M. Zhang, J. Zhang, R. Wang, R. Govindan, J. C. Mogul, and A. Vahdat, “Gemini: Practical reconfigurable datacenter networks with topology and traffic engineering,” arXiv preprint arXiv:2110.08374 , 2021

    2021 arXiv

  20. [28]

    Shoal: A network architecture for disaggregated racks,

    V . Shrivastav, A. Valadarsky, H. Ballani, P. Costa, K. S. Lee, H. Wang, R. Agarwal, and H. Weatherspoon, “Shoal: A network architecture for disaggregated racks,” in 16th USENIX Symposium on Networked Systems Design and Implementation (NSDI 19) , 2019, pp. 255–270

    2019

  21. [29]

    Re-architecting datacenter networks and stacks for low latency and high performance,

    M. Handley, C. Raiciu, A. Agache, A. V oinescu, A. W. Moore, G. Antichi, and M. Wójcik, “Re-architecting datacenter networks and stacks for low latency and high performance,” in Proceedings of the Conference of the ACM Special Interest Group on Data Communication , 2017, pp. 29–42

    2017

  22. [30]

    Inside the social network’s (datacenter) network,

    A. Roy, H. Zeng, J. Bagga, G. Porter, and A. C. Snoeren, “Inside the social network’s (datacenter) network,” in Proceedings of the 2015 ACM Conference on Special Interest Group on Data Communication , 2015, pp. 123–137

    2015

  23. [31]

    High-resolution measurement of data center microbursts,

    Q. Zhang, V . Liu, H. Zeng, and A. Krishnamurthy, “High-resolution measurement of data center microbursts,” in Proceedings of the 2017 Internet Measurement Conference , 2017, pp. 78–85

    2017

  24. [32]

    VL2: A scalable and flexible data center network,

    A. Greenberg, J. R. Hamilton, N. Jain, S. Kandula, C. Kim, P. Lahiri, D. A. Maltz, P. Patel, and S. Sengupta, “VL2: A scalable and flexible data center network,” in Proceedings of the ACM SIGCOMM 2009 conference on Data communication , 2009, pp. 51–62

    2009

  25. [33]

    Lighthouse: An open research framework for optical data center networks,

    Y . Lei, F. De Marchi, J. Li, R. Joshi, B. Chandrasekaran, and Y . Xia, “Lighthouse: An open research framework for optical data center networks,” arXiv preprint arXiv:2411.18319 , 2024

    2024 arXiv

  26. [34]

    An open research framework for optical data center networks,

    Y . Lei, F. De Marchi, R. Joshi, J. Li, B. Chandrasekaran, and Y . Xia, “An open research framework for optical data center networks,” inProceedings TRANSACTIONS ON NETWORKING, VOL. XX, NO. XX, OCTOBER 2024 14 of the ACM SIGCOMM 2024 Conference: Posters and Demos , 2024, pp. 86–88

    2024

  27. [35]

    Corundum: An open-source 100-gbps nic,

    A. Forencich, A. C. Snoeren, G. Porter, and G. Papen, “Corundum: An open-source 100-gbps nic,” in 2020 IEEE 28th Annual International Sym- posium on Field-Programmable Custom Computing Machines (FCCM) . IEEE, 2020, pp. 38–46

    2020

  28. [36]

    Nanosecond precision time synchronization for optical data center networks,

    Y . Lei, J. Li, Z. Liu, R. Joshi, and Y . Xia, “Nanosecond precision time synchronization for optical data center networks,” arXiv preprint arXiv:2410.17012, 2024

    2024 arXiv

  29. [37]

    Programmable calendar queues for high-speed packet scheduling,

    N. K. Sharma, C. Zhao, M. Liu, P. G. Kannan, C. Kim, A. Krishnamurthy, and A. Sivaraman, “Programmable calendar queues for high-speed packet scheduling,” in Proceedings of NSDI , 2020

    2020

  30. [38]

    Timertasks: Towards time-driven execution in programmable dataplanes,

    R. Joshi, B. Leong, and M. C. Chan, “Timertasks: Towards time-driven execution in programmable dataplanes,” in Proceedings of SIGCOMM Posters and Demos, 2019

    2019

  31. [39]

    Homa: A receiver-driven low-latency transport protocol using network priorities,

    B. Montazeri, Y . Li, M. Alizadeh, and J. Ousterhout, “Homa: A receiver-driven low-latency transport protocol using network priorities,” in Proceedings of the 2018 Conference of the ACM Special Interest Group on Data Communication , 2018, pp. 221–235

    2018

  32. [40]

    Time-division tcp for reconfigurable data center networks,

    S. S. Chen, W. Wang, C. Canel, S. Seshan, A. C. Snoeren, and P. Steenkiste, “Time-division tcp for reconfigurable data center networks,” in Proceedings of the ACM SIGCOMM 2022 Conference , 2022, pp. 19–35

    2022

  33. [41]

    Adapting TCP for Reconfigurable Datacenter Networks,

    M. K. Mukerjee, C. Canel, W. Wang, D. Kim, S. Seshan, and A. C. Snoeren, “Adapting TCP for Reconfigurable Datacenter Networks,” in 17th USENIX Symposium on Networked Systems Design and Implementation (NSDI 20) . Santa Clara, CA: USENIX Association, Feb. 2020, pp. 651–666. [Onl...

    2020

  34. [42]

    Bolt: Sub-rtt congestion control for ultra-low latency,

    S. Arslan, Y . Li, G. Kumar, and N. Dukkipati, “Bolt: Sub-rtt congestion control for ultra-low latency,” in 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23) , 2023, pp. 219–236

    2023

  35. [43]

    Memchached,

    “Memchached,” https://memcached.org/, 2024

    2024

  36. [44]

    Memslap,

    “Memslap,” http://docs.libmemcached.org/bin/memslap.html, 2024

    2024

  37. [45]

    “iperf,” https://iperf.fr/, 2024

    2024

  38. [46]

    Improving datacenter performance and robustness with multipath tcp,

    C. Raiciu, S. Barre, C. Pluntke, A. Greenhalgh, D. Wischik, and M. Handley, “Improving datacenter performance and robustness with multipath tcp,” ACM SIGCOMM Computer Communication Review , vol. 41, no. 4, pp. 266–277, 2011

    2011

  39. [47]

    Jellyfish: Networking data centers randomly,

    A. Singla, C.-Y . Hong, L. Popa, and P. B. Godfrey, “Jellyfish: Networking data centers randomly,” in Presented as part of the 9th {USENIX} Symposium on Networked Systems Design and Implementation ( {NSDI} 12), 2012, pp. 225–238

    2012

  40. [48]

    Xpander: Towards optimal-performance datacenters,

    A. Valadarsky, G. Shahaf, M. Dinitz, and M. Schapira, “Xpander: Towards optimal-performance datacenters,” in Proceedings of the 12th International on Conference on emerging Networking EXperiments and Technologies, 2016, pp. 205–219

    2016

  41. [49]

    Data center tcp (dctcp),

    M. Alizadeh, A. Greenberg, D. A. Maltz, J. Padhye, P. Patel, B. Prabhakar, S. Sengupta, and M. Sridharan, “Data center tcp (dctcp),” in Proceedings of the ACM SIGCOMM 2010 Conference , 2010, pp. 63–74

    2010

  42. [50]

    Mars: Near-optimal throughput with shallow buffers in reconfigurable datacenter networks,

    V . Addanki, C. Avin, and S. Schmid, “Mars: Near-optimal throughput with shallow buffers in reconfigurable datacenter networks,” Proceedings of the ACM on Measurement and Analysis of Computing Systems , vol. 7, no. 1, pp. 1–43, 2023

    2023

  43. [51]

    Hop- on hop-off routing: A fast tour across the optical data center network for latency-sensitive flows,

    J. Li, Y . Lei, F. De Marchi, R. Joshi, B. Chandrasekaran, and Y . Xia, “Hop- on hop-off routing: A fast tour across the optical data center network for latency-sensitive flows,” in Proceedings of the 6th Asia-Pacific Workshop on Networking, 2022, pp. 63–69

    2022

  44. [52]

    Uniform-cost multi-path routing for reconfigurable data center networks,

    J. Li, H. Gong, F. De Marchi, A. Gong, Y . Lei, W. Bai, and Y . Xia, “Uniform-cost multi-path routing for reconfigurable data center networks,” in Proceedings of the ACM SIGCOMM 2024 Conference , 2024, pp. 433–448

    2024

  45. [53]

    Poster: Opportunistic credit- based transport for reconfigurable data center networks with tidal,

    F. De Marchi, J. Li, W. Bai, and Y . Xia, “Poster: Opportunistic credit- based transport for reconfigurable data center networks with tidal,” in Proceedings of the ACM SIGCOMM 2024 Conference: Posters and Demos, 2024, pp. 4–6

    2024

  46. [54]

    Rethinking transport protocols for reconfigurable data centers: An empirical study,

    F. De Marchi, W. Bai, J. Li, and Y . Xia, “Rethinking transport protocols for reconfigurable data centers: An empirical study,” in Proceedings of the 1st SIGCOMM Workshop on Hot Topics in Optical Technologies and Applications in Networking , 2024, pp. 7–13

    2024

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.