Your new reading list.

Sidekick: In-Network Assistance for Secure End-to-End Transport Protocols
Gina Yuan, Matthew Sotoudeh, David K. Zhang,
Michael Welzl, David Mazières, and Keith Winstein
NSDI 2024, pages 1813–1830
https://www.usenix.org/conference/nsdi24/presentation/yuan
NSDI 2024 Outstanding Paper Award and Community Award Winner!
Published online in April 2024.

Abstract: In response to concerns about protocol ossification and privacy, post-TCP transport protocols such as QUIC and Web-RTC include end-to-end encryption and authentication at the transport layer. This makes their packets opaque to middleboxes, freeing the transport protocol to evolve but preventing some in-network innovations and performance improvements. This paper describes sidekick protocols: an approach to in-network assistance for opaque transport protocols where in-network intermediaries help endpoints by sending information adjacent to the underlying connection, which remains opaque and unmodified on the wire.

A key technical challenge is how the sidekick connection can efficiently refer to ranges of packets of the underlying connection without the ability to observe cleartext sequence numbers. We present a mathematical tool called a quACK that concisely represents a selective acknowledgment of opaque packets, without access to cleartext sequence numbers.

In real-world and emulation-based evaluations, the sidekick improved performance in several scenarios: early retransmission over lossy Wi-Fi paths, proxy acknowledgments to save energy, and a path-aware congestion-control mechanism we call PACUBIC that emulates a “split” connection.

An explicit 16-stage Runge–Kutta method of order 10 discovered by numerical search
David K. Zhang
Numerical Algorithms, volume 96, pages 1243–1267
DOI: 10.1007/s11075-024-01783-2
Published online on March 1, 2024.

Abstract: This article presents the discovery of an explicit 16-stage Runge–-Kutta method that numerically satisfies the Runge-–Kutta order conditions (in Butcher form) to 10th order, conjecturally improving the best-known number of stages in an explicit 10th-order Runge-–Kutta method from 17 to 16. Unlike the vast majority of published Runge-–Kutta methods, the method presented in this paper was not constructed via symbolic analysis or the use of simplifying assumptions, but instead by directly applying numerical optimization and root-finding algorithms to the order conditions. A naïve implementation of these algorithms would be made computationally infeasible by the considerable size and complex structure of the Butcher equations. However, we present a collection of software optimizations that greatly accelerate the evaluation of the Butcher equations and their derivatives while mitigating the effects of destructive cancellation, pushing these techniques within the realm of feasibility. While we do not have a formal proof of order, we present the results of numerical experiments to demonstrate that our method satisfies the Butcher equations to an accuracy of over 3000 decimal digits.

Sidecar: in-network performance enhancements in the age of paranoid transport protocols
Gina Yuan, David K. Zhang, Matthew Sotoudeh, Michael Welzl, and Keith Winstein
HotNets 2022, pages 221–227
DOI: 10.1145/3563766.3564113
Published online on November 14, 2022.

Abstract: In response to ossification and privacy concerns, post-TCP transport protocols such as QUIC are designed to be “paranoid”—opaque to meddling middleboxes by encrypting and authenticating the header and payload—making it impossible for Performance-Enhancing Proxies (PEPs) to provide the same assistance as before. We propose a research agenda towards an alternate approach to PEPs, creating a sidecar protocol that is loosely-coupled to the unchanged and opaque, underlying transport protocol. The key technical challenge to sidecar protocols is how to usefully refer to the packets of the underlying connection without ossification. We have made progress on this problem by creating a tool we call a quACK (quick ACK), a concise representation of a multiset of numbers that can be used to efficiently decode the randomly-encrypted packet contents a sidecar has received. We implement the quACK and discuss how to achieve several applications with this approach: alternate congestion control, ACK reduction, and PEP-to-PEP retransmission across a lossy subpath.

Manipulating Elections by Selecting Issues
Jasper Lu, David Kai Zhang, Zinovi Rabinovich, Svetlana Obraztsova, and Yevgeniy Vorobeychik
AAMAS 2019, pages 529–537
DOI: 10.5555/3306127.3331736
Published online on May 8, 2019.

Abstract: Constructive election control considers the problem of an adversary who seeks to sway the outcome of an electoral process in order to ensure that their favored candidate wins. We consider the computational problem of constructive election control via issue selection. In this problem, a part decides which political issues to focus on to ensure victory for the favored candidate. We also consider a variation in which the goal is to maximize the number of voters supporting the favored candidate. We present strong negative results, showing, for example, that the latter problem is inapproximable for any constant factor. On the positive side, we show that when issues are binary, the problem becomes tractable in several cases, and admits a 2-approximation in the two-candidate case. Finally, we develop integer programming and heuristic methods for these problems.

Discovering New Runge–Kutta Methods Using Unstructured Numerical Search
David K. Zhang
Vanderbilt University, Undergraduate Honors Thesis
https://arxiv.org/abs/1911.00318
Submitted on April 16, 2019.

Abstract: Runge–Kutta methods are a popular class of numerical methods for solving ordinary differential equations. Every Runge–Kutta method is characterized by two basic parameters: its order, which measures the accuracy of the solution it produces, and its number of stages, which measures the amount of computational work it requires. The primary goal in constructing Runge–Kutta methods is to maximize order using a minimum number of stages. However, high-order Runge–Kutta methods are difficult to construct because their parameters must satisfy an exponentially large system of polynomial equations. This paper presents the first known 10^th-order Runge–Kutta method with only 16 stages, breaking a 40-year standing record for the number of stages required to achieve 10^th-order accuracy. It also discusses the tools and techniques that enabled the discovery of this method using a straightforward numerical search.

Binding energies and structures of two-dimensional excitonic complexes in transition metal dichalcogenides
Daniel W. Kidd, David K. Zhang. and Kálmán Varga
Physical Review B, volume 93, issue 12, page 125423 (1–10)
DOI: 10.1103/PhysRevB.93.125423
Published online on March 18, 2016.

Abstract: The stochastic variational method is applied to excitonic formations within semiconducting transition metal dichalcogenides using a correlated Gaussian basis. The energy and structure of two- to six-particle systems are investigated along with their dependence on the effective screening length of the two-dimensional Keldysh potential and the electron-hole effective mass ratio. Excited state biexcitons are shown to be bound, with binding energies of the $L=0$ state showing good agreement with experimental measurements of biexciton binding energies. Ground and newly discussed excited state exciton-trions are predicted to be bound and their structures are investigated.

Excited Biexcitons in Transition Metal Dichalcogenides
David K. Zhang, Daniel W. Kidd, and Kálmán Varga
Nano Letters, volume 15, issue 10, pages 7002–7005
DOI: 10.1021/acs.nanolett.5b03009
Published online on September 30, 2015.

Abstract: The Stochastic Variational Method (SVM) is used to show that the effective mass model correctly estimates the binding energies of excitons and trions but fails to predict the experimental binding energy of the biexciton. Using high-accuracy variational calculations, it is demonstrated that the biexciton binding energy in transition metal dichalcogenides is smaller than the trion binding energy, contradicting experimental findings. It is also shown that the biexciton has bound excited states and that the binding energy of the $L=0$ excited state is in very good agreement with experimental data. This excited state corresponds to a hole attached to a negative trion and may be a possible resolution of the discrepancy between theory and experiment.

Excited Biexcitons in Transition Metal Dichalcogenides (contributed conference talk)
David K. Zhang
MAR16 Meeting of the American Physical Society
Talk delivered March 15, 2016, 4:30 PM–4:42 PM
See conference program listing and published abstract.

Abstract: Recently, experimental measurements and theoretical modeling have been in a disagreement concerning the binding energy of biexctions in transition metal dichalcogenides. While theory predicts a smaller binding energy (∼20 meV) that is, as logically expected, lower than that of the trion, experiment finds values much larger (∼60 meV), actually exceeding those for the trion. In this work, we show that there exists an excited state of the biexciton which yields binding energies that match well with experimental findings and thus gives a plausible explanation for the apparent discrepancy. Furthermore, it is shown that the electron-hole correlation functions of the ground state biexciton and trion are remarkably similar, possibly explaining why a distinct signature of ground state biexcitons would not have been noticed experimentally.

A General Algorithm for the Efficient Derivation of Linear Multistep Methods (contributed conference talk)
David K. Zhang and Samuel N. Jator
AMS Southeastern Spring Sectional Meeting #1097 (UT Knoxville)
Talk delivered March 22, 2014, 3:15 p.m.
See conference program listing and published abstract.

Abstract: Traditionally, linear multistep methods (LMMs) for the numerical solution of initial value problems, such as Adams methods and backward differentiation formulas, have been derived through the use of polynomial interpolation and collocation through continuous schemes. While these methods can be implemented in modern computer algebra systems, they require the use of highly expensive operations such as symbolic matrix inversion. This imposes a severe limit on the complexity of LMMs that can be derived. In this presentation, we present a generalized algorithm for deriving LMMs based upon Taylor series expansion. By our approach, we show that the derivation of a LMM containing $k + 1$ terms is reducible to the numerical solution of a $k \times k$ linear system, allowing for the efficient derivation of methods including hundreds or thousands of terms. Furthermore, we show that this algorithm is trivially generalizable to methods including arbitrarily many off-grid points, and that it can be generalized to create LMMs for directly solving initial value problems of arbitrarily high order, with the inclusion of all intermediate derivative terms. Specific methods are stated and tested numerically on well-known problems given in the literature.