pith. sign in

arxiv: 1907.06302 · v1 · pith:WIL52GRTnew · submitted 2019-07-15 · 💻 cs.NI · cs.SY· eess.SY

Compound TCP with Random Early Detection (RED): stability, bifurcation and performance analyses

Sreelakshmi Manjunath , Gaurav Raina This is my paper

Pith reviewed 2026-05-24 21:34 UTC · model grok-4.3

classification 💻 cs.NI cs.SYeess.SY
keywords Compound TCPRandom Early DetectionActive Queue ManagementStability AnalysisHopf BifurcationQueueing DelayTCP Congestion Control
0
0 comments X

The pith

Compound TCP with RED undergoes Hopf bifurcation as RTT or parameters vary, and a threshold policy outperforms it on delay and loss.

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

The paper derives a nonlinear time-delayed fluid model for Compound TCP interacting with RED to study local stability under changes in round-trip time, the queue averaging parameter, and packet-dropping thresholds. It obtains a sufficient stability condition, shows that the equilibrium loses stability through a Hopf bifurcation when any of those parameters is varied, and derives the necessary-and-sufficient stability condition when queue averaging is removed. Packet-level simulations confirm limit-cycle oscillations in queue size, and direct comparison indicates that averaging is not beneficial for stability; a simple threshold-based policy then yields lower queueing delay, shorter flow completion times, and less packet loss.

Core claim

We derive a non-linear time-delayed model for Compound TCP-RED and obtain a sufficient condition for local stability. The system undergoes a Hopf bifurcation as round-trip time, queue averaging parameter or packet-dropping thresholds are varied. In the regime without queue size averaging, we derive the necessary and sufficient condition for local stability. A comparison reveals that averaging may not be beneficial to system stability. The threshold-based queue policy outperforms RED in queueing delay, flow completion time and packet loss.

What carries the argument

The nonlinear time-delayed fluid model coupling Compound TCP window dynamics to RED's exponentially weighted moving average of queue size, from which local stability conditions and Hopf bifurcation boundaries are derived.

If this is right

  • Varying RTT or RED parameters induces limit-cycle oscillations in queue size that can synchronize TCP flows and reduce link utilization.
  • Removing queue-size averaging changes the stability condition from sufficient to necessary and sufficient.
  • The threshold-based policy produces stable low-latency operation with measurable gains in queueing delay, flow completion time and packet loss over RED.

Where Pith is reading between the lines

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

  • AQM designs that rely on exponential averaging may need re-examination if the stability penalty observed here appears in other TCP variants.
  • The threshold policy could be tested under heterogeneous RTTs and mixed traffic to check robustness beyond the single-link fluid model.
  • The Hopf analysis supplies explicit parameter boundaries that could be used to set RED thresholds in practice to avoid the oscillatory regime.

Load-bearing premise

The non-linear time-delayed fluid model accurately represents the packet-level dynamics of Compound TCP with RED.

What would settle it

Packet-level simulations or measurements in which queue size remains stable after parameters cross the predicted Hopf boundary, or in which the threshold policy shows no reduction in delay, flow completion time or loss relative to RED.

Figures

Figures reproduced from arXiv: 1907.06302 by Gaurav Raina, Sreelakshmi Manjunath.

Figure 1
Figure 1. Figure 1: Schematic diagram explaining the network scenario c [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Local stability charts, for Compound TCP with RED, sh [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Phase portraits for Compound TCP-RED: (a) convergen [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Local stability charts, for Compound TCP-RED in the a [PITH_FULL_IMAGE:figures/full_fig_p019_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Single bottleneck dumbbell topology, showing many e [PITH_FULL_IMAGE:figures/full_fig_p021_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: For RTT = 10 ms, RED maintains the average queue size between the thresholds, as intended. The average window size also appears to be randomly varying around 10 packets, this implies that the TCP flows are desyn￾chronised and the system is stable. The link utilisation is seen to be 100 %. When the RTT is increased to 200 ms, we observe a qualitative change in the queue size dynamics. The queue size and the… view at source ↗
Figure 6
Figure 6. Figure 6: Impact of RTT variation (Packet-level simulations. [PITH_FULL_IMAGE:figures/full_fig_p022_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Impact of RTT variation (Packet-level simulations. [PITH_FULL_IMAGE:figures/full_fig_p023_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Impact of averaging (Packet-level simulations. Hom [PITH_FULL_IMAGE:figures/full_fig_p024_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Impact of packet-dropping threshold (Packet-level [PITH_FULL_IMAGE:figures/full_fig_p025_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Impact of RTT variation (Packet-level simulations [PITH_FULL_IMAGE:figures/full_fig_p026_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Impact of RTT variation (Packet-level simulations [PITH_FULL_IMAGE:figures/full_fig_p027_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Impact of averaging (Packet-level simulations. He [PITH_FULL_IMAGE:figures/full_fig_p028_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Impact of packet-dropping threshold (Packet-leve [PITH_FULL_IMAGE:figures/full_fig_p028_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Local stability and Hopf bifurcation in a system of C [PITH_FULL_IMAGE:figures/full_fig_p029_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Comparison of RED and threshold-based queue policy [PITH_FULL_IMAGE:figures/full_fig_p031_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Parking-lot topology for the performance evaluati [PITH_FULL_IMAGE:figures/full_fig_p032_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Comparison of RED and threshold-based queue policy [PITH_FULL_IMAGE:figures/full_fig_p033_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Performance evaluation of RED and threshold-based [PITH_FULL_IMAGE:figures/full_fig_p034_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Performance evaluation of RED and threshold-based [PITH_FULL_IMAGE:figures/full_fig_p035_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Plots of µ2 and β2 versus the non-dimensional bifurcation parameter κ. Observe that µ2 > 0 and β2 < 0. This implies that, for the chosen parameter values, the Hopf bifurcation is super-critical, and the emergent limit cycles are asymptotically orbitally stable. REFERENCES [1] C. Villamizar and C. Song, “High performance TCP in ANSNET,” ACM SIGCOMM Computer Communication Review, vol. 24, no. 5, pp. 45–60, … view at source ↗
read the original abstract

The problem of increased queueing delays in the Internet motivates the study of currently implemented transport protocols and active queue management (AQM) policies. We study Compound TCP (default protocol in Windows) with Random Early Detection (RED). RED uses an exponentially weighted moving average of the queue size to make packet-dropping decisions, aiming to control the queue size. One must study RED with current protocols in order to explore its viability in the context of increased queueing delays. We derive a non-linear time-delayed model for Compound TCP-RED. We derive a sufficient condition for local stability of this model, and examine the impact of (i) round-trip time (RTT) of the TCP flows, (ii) queue averaging parameter and (iii) packet-dropping thresholds. Further, we establish that the system undergoes a Hopf bifurcation as any of the above parameters is varied. This suggests the emergence of limit cycles in the queue size, which may lead to synchronisation of TCP flows and loss of link utilisation. Next, we study a regime where queue size averaging is not performed, and packet-dropping decisions are based on instantaneous queue size. In this regime, we derive the necessary and sufficient condition for local stability. A comparison of the stability results for Compound TCP-RED in the two regimes--with and without queue size averaging--reveals that averaging may not be beneficial to system stability. Packet-level simulations show that the queue size indeed exhibits limit cycle oscillations as system parameters are varied. We then outline a simple threshold-based queue policy, that could ensure stable low-latency operation. We show that the threshold policy outperforms RED in terms of queueing delay, flow completion time and packet loss. We highlight that the threshold-based policy could mitigate the issue of increased queueing delays in the Internet.

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

3 major / 2 minor

Summary. The paper derives a nonlinear time-delayed fluid model for Compound TCP interacting with RED (using exponentially weighted moving average of queue size), obtains a sufficient condition for local stability of this model, shows that the system undergoes a Hopf bifurcation as RTT, averaging weight, or drop thresholds vary, derives necessary and sufficient stability conditions for the instantaneous-queue variant of RED, compares the two regimes to conclude that queue averaging may not benefit stability, validates the predicted limit-cycle oscillations via packet-level simulations, and outlines a simple threshold-based AQM that outperforms RED on queueing delay, flow completion time, and packet loss.

Significance. If the fluid model accurately captures the packet-level dynamics and the stability boundaries can be compared on equal footing, the results would provide concrete guidance on AQM design for Compound TCP, including a cautionary note on the use of queue averaging and evidence that a threshold policy can achieve stable low-latency operation. The explicit derivation of necessary-and-sufficient conditions in one regime and the packet-level validation are strengths that would make the work a useful reference for TCP-AQM analysis.

major comments (3)
  1. [Abstract; stability-comparison paragraph] Abstract and the stability-comparison section: the claim that 'averaging may not be beneficial to system stability' is obtained by juxtaposing a sufficient condition (averaged-queue RED) against necessary-and-sufficient conditions (instantaneous-queue RED). Because a sufficient condition certifies stability only inside a (possibly strict) subset of the true region, the comparison does not establish that the averaged regime has a smaller stability region unless the sufficient condition is shown to be tight or the actual boundary is computed for both cases.
  2. [Section deriving sufficient condition for averaged-queue model] Model derivation and local-stability section for averaged RED: the sufficient stability condition is stated without an accompanying tightness argument or numerical verification that the boundary is close to the true stability limit; this directly affects the load-bearing comparison with the instantaneous case.
  3. [Hopf-bifurcation subsection] Hopf-bifurcation analysis: the transversality condition and the sign of the first Lyapunov coefficient are asserted for the averaged model, but the paper does not report the explicit algebraic expressions or numerical checks confirming that the bifurcation is supercritical (or subcritical) for the parameter ranges used in the simulations.
minor comments (2)
  1. [Model-equation section] Notation for the averaging weight w and the drop thresholds min_th, max_th should be introduced once with consistent symbols across the model equations and the stability conditions.
  2. [Simulation section] The packet-level simulation figures would benefit from explicit reporting of the number of flows, link capacity, and the exact parameter values at which oscillations appear, to allow direct comparison with the analytically predicted boundaries.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments on our analysis of Compound TCP with RED. We address each major point below and will incorporate revisions to strengthen the stability comparison and bifurcation details.

read point-by-point responses

  1. Referee: [Abstract; stability-comparison paragraph] Abstract and the stability-comparison section: the claim that 'averaging may not be beneficial to system stability' is obtained by juxtaposing a sufficient condition (averaged-queue RED) against necessary-and-sufficient conditions (instantaneous-queue RED). Because a sufficient condition certifies stability only inside a (possibly strict) subset of the true region, the comparison does not establish that the averaged regime has a smaller stability region unless the sufficient condition is shown to be tight or the actual boundary is computed for both cases.

    Authors: We acknowledge that juxtaposing a sufficient condition against a necessary-and-sufficient one provides only a conservative indication rather than a definitive comparison of the full stability regions. The sufficient condition was derived via the same linearization and Routh-Hurwitz approach used for the instantaneous case, and packet-level simulations corroborate the predicted instability onset. To address the concern rigorously, we will add numerical boundary tracing (via continuation or grid search on the characteristic equation) for the averaged-queue model in the revision, allowing a direct comparison of the actual stability boundaries. revision: partial

  2. Referee: [Section deriving sufficient condition for averaged-queue model] Model derivation and local-stability section for averaged RED: the sufficient stability condition is stated without an accompanying tightness argument or numerical verification that the boundary is close to the true stability limit; this directly affects the load-bearing comparison with the instantaneous case.

    Authors: The sufficient condition arises from applying the Routh-Hurwitz criterion to the linearized delay-differential system; we did not include a separate tightness proof because the derivation already yields explicit parameter inequalities. We agree that explicit verification strengthens the claim. In the revised manuscript we will include a short subsection with numerical checks (solving the quasi-polynomial for marginal stability) showing that the analytic boundary lies close to the numerically computed one for representative parameter values. revision: yes

  3. Referee: [Hopf-bifurcation subsection] Hopf-bifurcation analysis: the transversality condition and the sign of the first Lyapunov coefficient are asserted for the averaged model, but the paper does not report the explicit algebraic expressions or numerical checks confirming that the bifurcation is supercritical (or subcritical) for the parameter ranges used in the simulations.

    Authors: The transversality condition (non-zero derivative of the real part of the critical eigenvalue with respect to the bifurcation parameter) and the first Lyapunov coefficient were computed symbolically during the analysis, but the lengthy intermediate expressions were omitted for brevity. We will add the explicit formulas for both quantities together with numerical evaluations over the RTT, averaging weight, and threshold ranges used in the simulations, confirming the bifurcation type. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper derives a non-linear time-delayed fluid model for Compound TCP-RED from standard modeling premises, then obtains a sufficient stability condition for the averaged-queue regime and necessary-and-sufficient conditions for the instantaneous-queue regime via direct mathematical analysis of that model. The interpretive comparison of these conditions (leading to the claim that averaging may not benefit stability) does not reduce any result to a fitted parameter, self-citation, or definitional equivalence; the derivations remain independent of the target claims. No load-bearing self-citations, ansatzes smuggled via prior work, or renamings of known results are present. The chain is therefore self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based solely on the abstract; no explicit free parameters, axioms, or invented entities are identifiable. The model itself is treated as given rather than derived from first principles in the provided text.

pith-pipeline@v0.9.0 · 5872 in / 1323 out tokens · 39871 ms · 2026-05-24T21:34:08.760526+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

60 extracted references · 60 canonical work pages

  1. [1]

    High performance TCP in ANSNE T,

    C. Villamizar and C. Song, “High performance TCP in ANSNE T,” ACM SIGCOMM Computer Communication Review , vol. 24, no. 5, pp. 45–60, 1994

    work page 1994

  2. [2]

    On packet switches with infinite storage,

    J. Nagle, “On packet switches with infinite storage,” IEEE Transactions on Communications , vol. 35, no. 4, pp. 435–438, 1987

    work page 1987

  3. [3]

    Recommendations on queue management and congestion avoidance in the Interne t,

    B. Braden, D. Clark, J. Crowcroft, B. Davie, S. Deering, D . Estrin, S. Floyd, V . Jacobson, G. Minshall, C. Partridge et al., “Recommendations on queue management and congestion avoidance in the Interne t,” Tech. Rep., 1998

    work page 1998

  4. [4]

    Random early detection gatewa ys for congestion avoidance,

    S. Floyd and V . Jacobson, “Random early detection gatewa ys for congestion avoidance,” IEEE/ACM Transactions on Networking , vol. 1, no. 4, pp. 397–413, 1993

    work page 1993

  5. [5]

    Rem: Active queue management,

    S. Athuraliya, S. H. Low, V . H. Li, and Q. Yin, “Rem: Active queue management,” IEEE Network , vol. 15, no. 3, pp. 48–53, 2001

    work page 2001

  6. [6]

    Analys is and design of controllers for AQM routers supporting TCP fl ows,

    C. V . Hollot, V . Misra, D. Towsley, and W. B. Gong, “Analys is and design of controllers for AQM routers supporting TCP fl ows,” IEEE Transactions on Automatic Control , vol. 47, no. 6, pp. 945–959, 2002

    work page 2002

  7. [7]

    Exponential-RED: a sta bilizing AQM scheme for low- and high-speed TCP protocols,

    S. Liu, T. Basar, and R. Srikant, “Exponential-RED: a sta bilizing AQM scheme for low- and high-speed TCP protocols,” IEEE/ACM Transactions on Networking , vol. 13, no. 5, pp. 1068–1081, 2005

    work page 2005

  8. [8]

    PIE: A lightweight contro l scheme to address the bufferbloat problem,

    R. Pan, P . Natarajan, C. Piglione, M. S. Prabhu, V . Subram anian, F. Baker, and B. V erSteeg, “PIE: A lightweight contro l scheme to address the bufferbloat problem,” in High Performance Switching and Routing (HPSR), 2013 IEEE 14 th International Conference on . IEEE, 2013, pp. 148–155

    work page 2013

  9. [9]

    Bufferbloat: dark buffers in t he internet,

    J. Gettys and K. Nichols, “Bufferbloat: dark buffers in t he internet,” Communications of the ACM , vol. 55, no. 1, pp. 57–65, 2012

    work page 2012

  10. [10]

    Bufferbloat and other Internet challenges ,

    V . G. Cerf, “Bufferbloat and other Internet challenges ,” IEEE Internet Computing , vol. 18, no. 5, pp. 80–80, 2014

    work page 2014

  11. [11]

    Controlling queue delay,

    K. Nichols and V . Jacobson, “Controlling queue delay,” Communications of the ACM , vol. 55, no. 7, pp. 42–50, 2012

    work page 2012

  12. [12]

    Analytic evaluatio n of RED performance,

    T. Bonald, M. May, and J.-C. Bolot, “Analytic evaluatio n of RED performance,” in Proceedings of IEEE INFOCOM , vol. 3. IEEE, 2000, pp. 1415–1424

    work page 2000

  13. [13]

    Tun ing RED for web traffic,

    M. Christiansen, K. Jeffay, D. Ott, and F. D. Smith, “Tun ing RED for web traffic,” in ACM SIGCOMM Computer Communication Review , vol. 30, no. 4. ACM, 2000, pp. 139–150

    work page 2000

  14. [14]

    Reasons not to de ploy RED,

    M. May, J. Bolot, C. Diot, and B. Lyles, “Reasons not to de ploy RED,” in Quality of Service, 1999. IWQoS’99. 1999 Seventh Internati onal W orkshop on. IEEE, 1999, pp. 260–262

    work page 1999

  15. [15]

    Weighted RED (WTRED) strategy for TCP congestion control,

    N. Hamadneh, D. Murray, M. Dixon, and P . Cole, “Weighted RED (WTRED) strategy for TCP congestion control,” Informatics Engineering and Information Science , pp. 421–434, 2011

    work page 2011

  16. [16]

    RRED: robust RED al gorithm to counter low-rate denial-of-service attacks,

    C. Zhang, J. Yin, Z. Cai, and W. Chen, “RRED: robust RED al gorithm to counter low-rate denial-of-service attacks,” IEEE Communications Letters, vol. 14, no. 5, 2010

    work page 2010

  17. [17]

    Subsidized red: an active queue management mechanism for short-lived flows,

    B. Wang, B. Kasthurirangan, and J. Xu, “Subsidized red: an active queue management mechanism for short-lived flows, ” Computer communications, vol. 28, no. 5, pp. 540–549, 2005. 44

    work page 2005

  18. [18]

    Implementation of network monitoring and packets captu ring using random early detection (red) method,

    B. Siregar, M. Manik, R. Rahmat, U. Andayani, and F. Fahm i, “Implementation of network monitoring and packets captu ring using random early detection (red) method,” in 2017 IEEE International Conference on Communication, Netw orks and Satellite (Comnetsat) . IEEE, 2017, pp. 42–47

    work page 2017

  19. [19]

    Efr ed: Enhancement of fair random early detection algorithm,

    M. Abdulkareem, K. Akil, A. Kalakech, and S. Kadry, “Efr ed: Enhancement of fair random early detection algorithm,” International Journal of Communications, Network and System Sciences , vol. 8, no. 07, p. 282, 2015

    work page 2015

  20. [20]

    Wide-area Int ernet traffic patterns and characteristics,

    K. Thompson, G. J. Miller, and R. Wilder, “Wide-area Int ernet traffic patterns and characteristics,” IEEE Network , vol. 11, no. 6, pp. 10–23, 1997

    work page 1997

  21. [21]

    Internet traffic measurement,

    C. Williamson, “Internet traffic measurement,” IEEE Internet Computing , vol. 5, no. 6, pp. 70–74, 2001

    work page 2001

  22. [22]

    Towards understanding residential internet traffic: From packets t o services,

    Z. Aouini, A. Kortebi, and Y . Ghamri-Doudane, “Towards understanding residential internet traffic: From packets t o services,” in Network of the Future (NOF), 2016 7th International Conference on th e. IEEE, 2016, pp. 1–7

    work page 2016

  23. [23]

    A compound T CP approach for high-speed and long distance networks,

    K. Tan, J. Song, Q. Zhang, and M. Sridharan, “A compound T CP approach for high-speed and long distance networks,” in Proceedings-IEEE INFOCOM, 2006

    work page 2006

  24. [24]

    CUBIC: a new TCP-friendly high -speed TCP variant,

    S. Ha, I. Rhee, and L. Xu, “CUBIC: a new TCP-friendly high -speed TCP variant,” ACM SIGOPS Operating Systems Review , vol. 42, no. 5, pp. 64–74, 2008

    work page 2008

  25. [25]

    TCP c ongestion avoidance algorithm identification,

    P . Y ang, J. Shao, W. Luo, L. Xu, J. Deogun, and Y . Lu, “TCP c ongestion avoidance algorithm identification,” IEEE/ACM Transactions on Networking, vol. 22, no. 4, pp. 1311–1324, 2014

    work page 2014

  26. [26]

    Modeling compound tcp over wifi for iot,

    S. R. Pokhrel, C. Williamson, C. Williamson, and S. R. Po khrel, “Modeling compound tcp over wifi for iot,” IEEE/ACM Transactions on Networking, vol. 26, no. 2, pp. 864–878, 2018

    work page 2018

  27. [27]

    Delay and loss-based transport pr otocols: Buffer-sizing and stability,

    P . Raja and G. Raina, “Delay and loss-based transport pr otocols: Buffer-sizing and stability,” in Communication Systems and Networks (COMSNETS), 2012 F ourth International Conference on . IEEE, 2012, pp. 1–10

    work page 2012

  28. [28]

    G. F. Franklin, J. D. Powell, and A. Emami-Naeini, Feedback control of dynamic systems . Addison-Wesley Reading, MA, 1994, vol. 3

    work page 1994

  29. [29]

    Srikant, The mathematics of Internet congestion control

    R. Srikant, The mathematics of Internet congestion control . Springer Science & Business Media, 2012

    work page 2012

  30. [30]

    Fluid-based anal ysis of a network of AQM routers supporting TCP flows with an ap plication to RED,

    V . Misra, W.-B. Gong, and D. Towsley, “Fluid-based anal ysis of a network of AQM routers supporting TCP flows with an ap plication to RED,” in ACM SIGCOMM Computer Communication Review , vol. 30, no. 4. ACM, 2000, pp. 151–160

    work page 2000

  31. [31]

    A con trol theoretic analysis of RED,

    C. V . Hollot, V . Misra, D. Towsley, and W.-B. Gong, “A con trol theoretic analysis of RED,” in INFOCOM 2001. Twentieth Annual Joint Conference of the IEEE Computer and Communications Societi es. Proceedings. IEEE , vol. 3. IEEE, 2001, pp. 1510–1519

    work page 2001

  32. [32]

    Dynamics of TCP/RED and a scalable control,

    S. H. Low, F. Paganini, J. Wang, S. Adlakha, and J. C. Doyl e, “Dynamics of TCP/RED and a scalable control,” in INFOCOM 2002. Twenty-First Annual Joint Conference of the IEEE Computer a nd Communications Societies. Proceedings. IEEE , vol. 1. IEEE, 2002, pp. 239–248

    work page 2002

  33. [33]

    A contro l-theoretic analysis of low-priority congestion control r eprioritization under aqm,

    L. D. Cicco, Y . Gong, D. Rossi, and E. Leonardi, “A contro l-theoretic analysis of low-priority congestion control r eprioritization under aqm,” ACM Transactions on Modeling and Performance Evaluation of Computing Systems , vol. 1, no. 4, p. 17, 2016

    work page 2016

  34. [34]

    Guckenheimer and P

    J. Guckenheimer and P . J. Holmes, Nonlinear oscillations, dynamical systems, and bifurcati ons of vector fields . Springer Science & Business Media, 2013, vol. 42

    work page 2013

  35. [35]

    Part II: Control t heory for buffer sizing,

    G. Raina, D. Towsley, and D. Wischik, “Part II: Control t heory for buffer sizing,” ACM SIGCOMM Computer Communication Review , vol. 35, no. 3, pp. 79–82, 2005

    work page 2005

  36. [36]

    Oscillating behavior of network traffic: A case study simulation,

    L. Zhang and D. Clark, “Oscillating behavior of network traffic: A case study simulation,” Internetworking: research and experience , vol. 1, no. 2, pp. 101–112, 1990

    work page 1990

  37. [37]

    ns-Network simulator,

    S. McCanne, “ns-Network simulator,” http://www. isi. edu/nsnam/ns , 1995

    work page 1995

  38. [38]

    Congestion avoidance and control,

    V . Jacobson, “Congestion avoidance and control,” in ACM SIGCOMM computer communication review , vol. 18, no. 4. ACM, 1988, pp. 314–329

    work page 1988

  39. [39]

    Modified TCP congestion avoidance algorithm,

    ——, “Modified TCP congestion avoidance algorithm,” end2end-interest mailing list , 1990

    work page 1990

  40. [40]

    TCP V egas: End to end con gestion avoidance on a global Internet,

    L. S. Brakmo and L. L. Peterson, “TCP V egas: End to end con gestion avoidance on a global Internet,” IEEE Journal on selected Areas in communications , vol. 13, no. 8, pp. 1465–1480, 1995

    work page 1995

  41. [41]

    FAST TCP: motiv ation, architecture, algorithms, performance,

    D. X. Wei, C. Jin, S. H. Low, and S. Hegde, “FAST TCP: motiv ation, architecture, algorithms, performance,” IEEE/ACM Transactions on Networking, vol. 14, no. 6, pp. 1246–1259, 2006

    work page 2006

  42. [42]

    Tcp-illinois: A loss -and delay-based congestion control algorithm for high-sp eed networks,

    S. Liu, T. Bas ¸ar, and R. Srikant, “Tcp-illinois: A loss -and delay-based congestion control algorithm for high-sp eed networks,” Performance Evaluation, vol. 65, no. 6-7, pp. 417–440, 2008

    work page 2008

  43. [43]

    Tcp-africa: An adap tive and fair rapid increase rule for scalable tcp,

    R. King, R. Baraniuk, and R. Riedi, “Tcp-africa: An adap tive and fair rapid increase rule for scalable tcp,” in INFOCOM 2005. 24th Annual Joint Conference of the IEEE Computer and Communicat ions Societies. Proceedings IEEE , vol. 3. IEEE, 2005, pp. 1838–1848. 45

    work page 2005

  44. [44]

    Buffer sizes for large multipl exers: TCP queueing theory and instability analysis,

    G. Raina and D. Wischik, “Buffer sizes for large multipl exers: TCP queueing theory and instability analysis,” in Next Generation Internet Networks, 2005 . IEEE, 2005, pp. 173–180

    work page 2005

  45. [45]

    An asymptotic approximation for tcp compound,

    S. Poojary and V . Sharma, “An asymptotic approximation for tcp compound,” Queueing Systems , vol. 85, no. 3-4, pp. 211–247, 2017

    work page 2017

  46. [46]

    The macrosc opic behavior of the tcp congestion avoidance algorithm,

    M. Mathis, J. Semke, J. Mahdavi, and T. Ott, “The macrosc opic behavior of the tcp congestion avoidance algorithm,” ACM SIGCOMM Computer Communication Review , vol. 27, no. 3, pp. 67–82, 1997

    work page 1997

  47. [47]

    A mean-fiel d model for multiple tcp connections through a buffer implem enting red,

    F. Baccelli, D. R. McDonald, and J. Reynier, “A mean-fiel d model for multiple tcp connections through a buffer implem enting red,” Performance Evaluation, vol. 49, no. 1-4, pp. 77–97, 2002

    work page 2002

  48. [48]

    Analysis and design of an ad aptive virtual queue (avq) algorithm for active queue manag ement,

    S. Kunniyur and R. Srikant, “Analysis and design of an ad aptive virtual queue (avq) algorithm for active queue manag ement,” in ACM SIGCOMM Computer Communication Review , vol. 31, no. 4. ACM, 2001, pp. 123–134

    work page 2001

  49. [49]

    On de signing improved controllers for aqm routers supporting tc p flows,

    C. V . Hollot, V . Misra, D. Towsley, and W.-B. Gong, “On de signing improved controllers for aqm routers supporting tc p flows,” in Proceedings IEEE INFOCOM 2001. Conference on Computer Comm unications. Twentieth Annual Joint Conference of the IEEE C omputer and Communications Society (Cat. No. 01CH37213) , vol. 3. IEEE, 2001, pp. 1726–1734

    work page 2001

  50. [50]

    B. D. Hassard, N. D. Kazarinoff, and Y .-H. Wan, Theory and applications of Hopf bifurcation . CUP Archive, 1981, vol. 41

    work page 1981

  51. [51]

    Y . A. Kuznetsov, Elements of applied bifurcation theory . Springer Science & Business Media, 2013, vol. 112

    work page 2013

  52. [52]

    Numerical bifurcation analysis of delay differential equations usin g DDE-BIFTOOL,

    K. Engelborghs, T. Luzyanina, and D. Roose, “Numerical bifurcation analysis of delay differential equations usin g DDE-BIFTOOL,” ACM Transactions on Mathematical Software (TOMS) , vol. 28, no. 1, pp. 1–21, 2002

    work page 2002

  53. [53]

    DDE-BIFT OOL v. 2.00: a Matlab package for bifurcation analysis of del ay differential equations,

    K. Engelborghs, T. Luzyanina, and G. Samaey, “DDE-BIFT OOL v. 2.00: a Matlab package for bifurcation analysis of del ay differential equations,” 2001

    work page 2001

  54. [54]

    Stochastic models for generating synthetic HT TP source traffic,

    J. Cao, W. S. Cleveland, Y . Gao, K. Jeffay, F. D. Smith, an d M. Weigle, “Stochastic models for generating synthetic HT TP source traffic,” in INFOCOM 2004. Twenty-third AnnualJoint Conference of the I EEE Computer and Communications Societies , vol. 3. IEEE, 2004, pp. 1546–1557

    work page 2004

  55. [55]

    Local bifurcation analysis of some dual cong estion control algorithms,

    G. Raina, “Local bifurcation analysis of some dual cong estion control algorithms,” IEEE Transactions on Automatic Control , vol. 50, no. 8, pp. 1135–1146, 2005

    work page 2005

  56. [56]

    Desi gn of robust active queue management controllers for a class of TCP communication networks,

    C.-K. Chen, Y .-C. Hung, T.-L. Liao, and J.-J. Y an, “Desi gn of robust active queue management controllers for a class of TCP communication networks,” Information Sciences , vol. 177, no. 19, pp. 4059–4071, 2007

    work page 2007

  57. [57]

    Congestion contro l for high bandwidth-delay product networks,

    D. Katabi, M. Handley, and C. Rohrs, “Congestion contro l for high bandwidth-delay product networks,” ACM SIGCOMM computer communication review, vol. 32, no. 4, pp. 89–102, 2002

    work page 2002

  58. [58]

    Analysi s and design of the virtual rate control algorithm for stabil izing queues in TCP networks,

    E.-C. Park, H. Lim, K.-J. Park, and C.-H. Choi, “Analysi s and design of the virtual rate control algorithm for stabil izing queues in TCP networks,” Computer Networks , vol. 44, no. 1, pp. 17–41, 2004

    work page 2004

  59. [59]

    Why flow-completion time i s the right metric for congestion control,

    N. Dukkipati and N. McKeown, “Why flow-completion time i s the right metric for congestion control,” ACM SIGCOMM Computer Communication Review, vol. 36, no. 1, pp. 59–62, 2006

    work page 2006

  60. [60]

    S. H. Strogatz, Nonlinear dynamics and chaos: with applications to physics , biology, chemistry, and engineering . CRC Press, 2018

    work page 2018