arXiv subjects

Yu Gan

Publications and source records attributed to Yu Gan.

54 records · Page 3Linked to original sources

arXiv2020

Beyond Fourier transform: super-resolving optical coherence tomography

Optical coherence tomography (OCT) is a volumetric imaging modality that empowers clinicians and scientists to noninvasively visualize the cross-sections of biological samples. As the latest generation of its kind, Fourier-domain OCT (FD-OCT) offers a micrometer-scale axial resolution by taking advantage of coherence gating. Based on the current theory, it is believed the only way to obtain a higher-axial-resolution OCT image is to physically extend the system's spectral bandwidth given a certain central wavelength. Here, we showed the belief is wrong. We proposed a novel reconstruction framework, which integrates prior knowledge and exploits the \emph{shift-variance}, to retrospectively super-resolve OCT images without altering the system configurations. Both numerical and experimental results confirmed the processed image manifested an axial resolution beyond the previous theoretical prediction. We believe this result not only opens new horizons for future research directions in OCT reconstruction but also promises an immediate upgrade to tens of thousands of legacy OCT units currently deployed.

eess.IV↗

arXiv2019

uqSim: Scalable and Validated Simulation of Cloud Microservices

Current cloud services are moving away from monolithic designs and towards graphs of many loosely-coupled, single-concerned microservices. Microservices have several advantages, including speeding up development and deployment, allowing specialization of the software infrastructure, and helping with debugging and error isolation. At the same time they introduce several hardware and software challenges. Given that most of the performance and efficiency implications of microservices happen at scales larger than what is available outside production deployments, studying such effects requires designing the right simulation infrastructures. We present uqSim, a scalable and validated queueing network simulator designed specifically for interactive microservices. uqSim provides detailed intra- and inter-microservice models that allow it to faithfully reproduce the behavior of complex, many-tier applications. uqSim is also modular, allowing reuse of individual models across microservices and end-to-end applications. We have validated uqSim both against simple and more complex microservices graphs, and have shown that it accurately captures performance in terms of throughput and tail latency. Finally, we use uqSim to model the tail at scale effects of request fanout, and the performance impact of power management in latency-sensitive microservices.

cs.DC↗

arXiv2019

Isotropic charge screening of the anisotropic black phosphorus revealed by potassium adatoms

Black phosphorus has attracted great research interest due to its numerous applications in electronic devices, optoelectronic devices, energy storages and so on. Compared with the majority of two-dimensional materials, black phosphorus possesses a unique property, i.e. the strong in-plane anisotropy. All the properties reported so far, including its effective mass, electron mobility, light absorption, thermal conductivity and so on, have shown great anisotropy in the basal plane. This property renders black phosphorus unique applications not achievable with other two-dimensional materials. In this work, however, we discover a remarkable isotropic behavior in the strongly anisotropic black phosphorus, i.e. its electrostatic screening of point charges. We use the tip-induced band bending of a scanning tunneling microscope to map out the Coulomb field of ionized potassium adatoms on black phosphorus, and reveal its isotropic charge screening. This discovery is important for understanding electron scattering and transport in black phosphorus.

cond-mat.mes-hall↗

arXiv2019

An Open-Source Benchmark Suite for Cloud and IoT Microservices

Cloud services have recently started undergoing a major shift from monolithic applications, to graphs of hundreds of loosely-coupled microservices. Microservices fundamentally change a lot of assumptions current cloud systems are designed with, and present both opportunities and challenges when optimizing for quality of service (QoS) and utilization. In this paper we explore the implications microservices have across the cloud system stack. We first present DeathStarBench, a novel, open-source benchmark suite built with microservices that is representative of large end-to-end services, modular and extensible. DeathStarBench includes a social network, a media service, an e-commerce site, a banking system, and IoT applications for coordination control of UAV swarms. We then use DeathStarBench to study the architectural characteristics of microservices, their implications in networking and operating systems, their challenges with respect to cluster management, and their trade-offs in terms of application design and programming frameworks. Finally, we explore the tail at scale effects of microservices in real deployments with hundreds of users, and highlight the increased pressure they put on performance predictability.

cs.DC↗

arXiv2019

Leveraging Deep Learning to Improve the Performance Predictability of Cloud Microservices

Performance unpredictability is a major roadblock towards cloud adoption, and has performance, cost, and revenue ramifications. Predictable performance is even more critical as cloud services transition from monolithic designs to microservices. Detecting QoS violations after they occur in systems with microservices results in long recovery times, as hotspots propagate and amplify across dependent services. We present Seer, an online cloud performance debugging system that leverages deep learning and the massive amount of tracing data cloud systems collect to learn spatial and temporal patterns that translate to QoS violations. Seer combines lightweight distributed RPC-level tracing, with detailed low-level hardware monitoring to signal an upcoming QoS violation, and diagnose the source of unpredictable performance. Once an imminent QoS violation is detected, Seer notifies the cluster manager to take action to avoid performance degradation altogether. We evaluate Seer both in local clusters, and in large-scale deployments of end-to-end applications built with microservices with hundreds of users. We show that Seer correctly anticipates QoS violations 91% of the time, and avoids the QoS violation to begin with in 84% of cases. Finally, we show that Seer can identify application-level design bugs, and provide insights on how to better architect microservices to achieve predictable performance.

cs.DC↗

arXiv2019

Chip-based frequency combs sources for optical coherence tomography

The Optical coherence tomography (OCT) is a powerful interferometric imaging technique widely used in medical fields such as ophthalmology, cardiology and dermatology, for which footprint and cost are becoming increasingly important. Here we present a platform for miniaturized sources for OCT based on chip-scale lithographically-defined microresonators. We show that the proposed platform is compatible with standard commercial spectral domain (SD) OCT systems and enable imaging of human tissue with an image quality comparable to the one achieved with tabletop commercial sources. This platform provides a path towards fully integrated OCT systems.

physics.optics↗

arXiv2018

The Architectural Implications of Microservices in the Cloud

Cloud services have recently undergone a shift from monolithic applications to microservices, with hundreds or thousands of loosely-coupled microservices comprising the end-to-end application. Microservices present both opportunities and challenges when optimizing for quality of service (QoS) and cloud utilization. In this paper we explore the implications cloud microservices have on system bottlenecks, and datacenter server design. We first present and characterize an end-to-end application built using tens of popular open-source microservices that implements a movie renting and streaming service, and is modular and extensible. We then use the end-to-end service to study the scalability and performance bottlenecks of microservices, and highlight implications they have on the design of datacenter hardware. Specifically, we revisit the long-standing debate of brawny versus wimpy cores in the context of microservices, we quantify the I-cache pressure they introduce, and measure the time spent in computation versus communication between microservices over RPCs. As more cloud applications switch to this new programming model, it is increasingly important to revisit the assumptions we have previously used to build and manage cloud systems.

cs.DC↗

arXiv2018

Seer: Leveraging Big Data to Navigate the Increasing Complexity of Cloud Debugging

Performance unpredictability in cloud services leads to poor user experience, degraded availability, and has revenue ramifications. Detecting performance degradation a posteriori helps the system take corrective action, but does not avoid the QoS violations. Detecting QoS violations after the fact is even more detrimental when a service consists of hundreds of thousands of loosely-coupled microservices, since performance hiccups can quickly propagate across the dependency graph of microservices. In this work we focus on anticipating QoS violations in cloud settings to mitigate performance unpredictability to begin with. We propose Seer, a cloud runtime that leverages the massive amount of tracing data cloud systems collect over time and a set of practical learning techniques to signal upcoming QoS violations, as well as identify the microservice(s) causing them. Once an imminent QoS violation is detected Seer uses machine-level hardware events to determine the cause of the QoS violation, and adjusts the resource allocations to prevent it. In local clusters with 10 40-core servers and 200-instance clusters on GCE running diverse cloud microservices, we show that Seer correctly anticipates QoS violations 91% of the time, and attributes the violation to the correct microservice in 89% of cases. Finally, Seer detects QoS violations early enough for a corrective action to almost always be applied successfully.

cs.DC↗

arXiv2017

The importance of sigma bonding electrons for the accurate description of electron correlation in graphene

Electron correlation in graphene is unique because of the interplay of the Dirac cone dispersion of $π$ electrons with long range Coulomb interaction. The random phase approximation predicts no metallic screening at long distance and low energy because of the zero density of states at Fermi level, so one might expect that graphene should be a poorly screened system. However, empirically graphene is a weakly interacting semimetal, which leads to the question of how electron correlations take place in graphene at different length scales. We address this question by computing the equal time and dynamic structure factor $S(\vec q)$ and $S(\vec q, ω)$ of freestanding graphene using {\it ab-initio} fixed-node diffusion Monte Carlo and the random phase approximation. We find that the $σ$ electrons contribute strongly to $S(\vec q,ω)$ for relevant experimental values of $ω$ even at distances up to around 80 Å. These findings illustrate how the emergent physics from underlying Coulomb interactions results in the observed weakly correlated semimetal.

cond-mat.mtrl-sci↗

arXiv2016

A reexamination of the effective fine structure constant of graphene, as measured in graphite

We present a refined and improved study of the influence of screening on the effective fine structure constant of graphene, $α^*$, as measured in graphite using inelastic x-ray scattering. This follow-up to our previous study [J. P. Reed, et al., Science 330, 805 (2010)] was carried out with two times better energy resolution, five times better momentum resolution, and improved experimental setup with lower background. We compare our results to RPA calculations and evaluate the relative importance of interlayer hopping, excitonic corrections, and screening from high energy excitations involving the $σ$ bands. We find that the static, limiting value of $α^*$ falls in the range 0.25 to 0.35, which is higher than our previous result of 0.14, but still below the value expected from RPA. We show the reduced value is not a consequence of interlayer hopping effects, which were ignored in our previous analysis, but of a combination of excitonic effects in the $π\rightarrow π^*$ particle-hole continuum, and background screening from the $σ$-bonded electrons. We find that $σ$-band screening is extremely strong at distances of the order of a few nm, and should be highly effective at screening out short-distance, Hubbard-like interactions in graphene, as well as other carbon allotropes.

cond-mat.str-el↗

arXiv2014

Temperature-resolution anomalies in the reconstruction of time dynamics from energy-loss experiments

Inelastic scattering techniques provide a powerful approach to studying electron and nuclear dynamics, via reconstruction of a propagator that quantifies the time evolution of a system. There is now growing interest in applying such methods to very low energy excitations, such as lattice vibrations, but in this limit the cross section is no longer proportional to a propagator. Significant deviations occur due to the finite temperature Bose statistics of the excitations. Here we consider this issue in the context of high-resolution electron energy loss experiments on the copper-oxide superconductor Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$. We find that simple division of a Bose factor yields an accurate propagator on energy scales greater than the resolution width. However, at low energy scales, the effects of resolution and finite temperature conspire to create anomalies in the dynamics at long times. We compare two practical ways for dealing with such anomalies, and discuss the range of validity of the technique in light of this comparison.

cond-mat.str-el↗

arXiv2013

First-principles method of propagation of tightly bound excitons: exciton band structure of LiF and verification with inelastic x-ray scattering

We propose a simple first-principles method to describe propagation of tightly bound excitons. By viewing the exciton as a composite object (an effective Frenkel exciton in Wannier orbitals), we define an exciton kinetic kernel to encapsulate the exciton propagation and decay for all binding energy. Applied to prototypical LiF, our approach produces three exciton bands, which we verified quantitatively via inelastic x-ray scattering. The proposed real-space picture is computationally inexpensive and thus enables study of the full exciton dynamics, even in the presence of surfaces and impurity scattering. It also provides intuitive understanding to facilitate practical exciton engineering in semiconductors, strongly correlated oxides, and their nanostructures.

cond-mat.str-el↗

arXiv2012

Optimal forwarding ratio on dynamical networks with heterogeneous mobility

As the discovery of non-Poissonian statistics of human mobility trajectories, more attention has been paid to understanding the role of these patterns in different dynamics. In this study, we first introduce the heterogeneous mobility of mobile agents into dynamical networks, and then investigate the forwarding strategy on the heterogeneous dynamical networks. We find that the faster speed and the higher proportion of high-speed agents can enhance the network throughput and reduce the mean traveling time in the case of random forwarding. A hierarchical structure in the dependence of high-speed is observed: the network throughput remains unchanged in small and large high-speed value. It is interesting to find that the slightly preferential forwarding to high-speed agents can maximize the network capacity. Through theoretical analysis and numerical simulations, we show that the optimal forwarding ratio stems from local structural heterogeneity of low-speed agents.

physics.soc-ph↗

arXiv2012

Crystallographic refinement of collective excitations using standing wave inelastic x-ray scattering

We propose a method for realizing true, real-space imaging of charge dynamics in a periodic system, with angstrom spatial resolution and attosecond time resolution. In this method, inelastic x-ray scattering (IXS) is carried out with a coherent, standing wave source, which provides the off-diagonal elements of the generalized dynamic structure factor, S(q_1,q_2,ω), allowing complete reconstruction of the inhomogeneous response function of the system, χ(x_1,x_2,t). The quantity χhas the physical meaning of a propagator for charge, so allows one to observe - in real time - the disturbance in the electron density created by a point source placed at a specified location, x_1 (on an atom vs. between atoms, for example). This method may be thought of as a generalization of x-ray crystallography that allows refinement of the excited states of a periodic system, rather than just its ground state.

cond-mat.str-el↗

arXiv2011

The electron many-body problem in graphene

We give a brief summary of the current status of the electron many-body problem in graphene. We claim that graphene has intrinsic dielectric properties which should dress the interactions among the quasiparticles, and may explain why the observation of electron-electron renormalization effects has been so elusive in the recent experiments. We argue that the strength of Coulomb interactions in graphene may be characterized by an effective fine structure constant given by $α^{\star}(\mathbf{k},ω)\equiv2.2/ε(\mathbf{k},ω)$, where $ε(\mathbf{k},ω)$ is the dynamical dielectric function. At long wavelengths, $α^{\star}(\mathbf{k},ω)$ appears to have its smallest value in the static regime, where $α^{\star}(\mathbf{k}\to0,0)\approx1/7$ according to recent inelastic x-ray measurements, and the largest value in the optical limit, where $α^{\star}(0,ω)\approx2.6$. We conclude that the strength of Coulomb interactions in graphene is not universal, but depends highly on the scale of the phenomenon of interest. We propose a prescription in order to reconcile different experiments.

cond-mat.str-el↗

arXiv2010

The effective fine structure constant of freestanding graphene measured in graphite

Electrons in graphene behave like Dirac fermions, permitting phenomena from high energy physics to be studied in a solid state setting. A key question is whether or not these Fermions are critically influenced by Coulomb correlations. We performed inelastic x-ray scattering experiments on crystals of graphite, and applied reconstruction algorithms to image the dynamical screening of charge in a freestanding, graphene sheet. We found that the polarizability of the Dirac fermions is amplified by excitonic effects, improving screening of interactions between quasiparticles. The strength of interactions is characterized by a scale-dependent, effective fine structure constant, α*(k,ω), whose value approaches α* ~ 1/7 at low energy and large distances. This value is substantially smaller than the nominal α= 2.2, suggesting that, on the whole, graphene is more weakly interacting than previously believed.

cond-mat.str-el↗

arXiv2009

Two-step stabilization of orbital order and the dynamical frustration of spin in the model charge-transfer insulator KCuF3

We report a combined experimental and theoretical study of KCuF3, which offers - because of this material's relatively simple lattice structure and valence configuration (d9, i.e., one hole in the d-shell) - a particularly clear view of the essential role of the orbital degree of freedom in governing the dynamical coupling between the spin and lattice degrees of freedom. We present Raman and x-ray scattering evidence that the phase behaviour of KCuF3 is dominated above the Neel temperature (T_N = 40 K) by coupled orbital/lattice fluctuations that are likely associated with rotations of the CuF6 octahedra, and we show that these orbital fluctuations are interrupted by a static structural distortion that occurs just above T_N. A detailed model of the orbital and magnetic phases of KCuF3 reveals that these orbital fluctuations - and the related frustration of in-plane spin-order-are associated with the presence of nearly degenerate low-energy spin-orbital states that are highly susceptible to thermal fluctuations over a wide range of temperatures. A striking implication of these results is that the ground state of KCuF3 at ambient pressure lies near a quantum critical point associated with an orbital/spin liquid phase that is obscured by emergent Neel ordering of the spins; this exotic liquid phase might be accessible via pressure studies.

cond-mat.str-el↗

arXiv2009

Implicit spatial averaging in inversion of inelastic x-ray scattering data

Inelastic x-ray scattering (IXS) now a widely used technique for studying the dynamics of electrons in condensed matter. We previously posed a solution to the phase problem for IXS [P. Abbamonte, et. al., Phys. Rev. Lett. {\bf 92}, 237401 (2004)], that allows explicit reconstruction of the density propagator of a system. The propagator represents, physically, the response of the system to an idealized, point perturbation, so provides direct, real-time images of electron motion with attosecond time resolution and $Å$-scale spatial resolution. Here we show that the images generated by our procedure, as it was originally posed, are spatial averages over all source locations. Within an idealized, atomic-like model, we show that in most cases a simple relationship to the complete, un-averaged response can still be determined. We illustrate this concept for recent IXS measurements of single crystal graphite.

cond-mat.str-el↗