Peter Gao

pHost: distributed near-optimal datacenter transport over commodity network fabric

The importance of minimizing flow completion times (FCT) in datacenters has led to a growing literature on new network transport designs. Of particular note is pFabric, a protocol that achieves near-optimal FCTs. However, pFabrics performance comes at the cost of generality, since pFabric requires specialized hardware that embeds a specific scheduling policy within the network fabric, making it hard to meet diverse policy goals. Aiming for generality, the recent Fastpass proposal returns to a design based on commodity network hardware and instead relies on a centralized scheduler. Fastpass achieves generality, but (as we show) loses many of pFabrics performance benefits. We present pHost, a new transport design aimed at achieving both: the near-optimal performance of pFabric and the commodity network design of Fastpass. Similar to Fastpass, pHost keeps the network simple by decoupling the network fabric from scheduling decisions. However, pHost introduces a new distributed protocol that allows end-hosts to directly make scheduling decisions, thus avoiding the overheads of Fastpasss centralized scheduler architecture. We show that pHost achieves performance on par with pFabric (within 4% for typical conditions) and significantly outperforms Fastpass (by a factor of 3.8×) while relying only on commodity network hardware.

Quartz: a new design element for low-latency DCNs

Most datacenter network (DCN) designs focus on maximizing bisection bandwidth rather than minimizing server-to-server latency. We explore architectural approaches to building low-latency DCNs and introduce Quartz, a design element consisting of a full mesh of switches. Quartz can be used to replace portions of either a hierarchical network or a random network. Our analysis shows that replacing high port-count core switches with Quartz can significantly reduce switching delays, and replacing groups of top-of-rack and aggregation switches with Quartz can significantly reduce congestion-related delays from cross-traffic. We overcome the complexity of wiring a complete mesh using low-cost optical multiplexers that enable us to efficiently implement a logical mesh as a physical ring. We evaluate our performance using both simulations and a small working prototype. Our evaluation results confirm our analysis, and demonstrate that it is possible to build low-latency DCNs using inexpensive commodity elements without significant concessions to cost, scalability, or wiring complexity.

Demonstration of a near-IR line-referenced electro-optical laser frequency comb for precision radial velocity measurements in astronomy

An important technique for discovering and characterizing planets beyond our solar system relies upon measurement of weak Doppler shifts in the spectra of host stars induced by the influence of orbiting planets. A recent advance has been the introduction of optical frequency combs as frequency references. Frequency combs produce a series of equally spaced reference frequencies and they offer extreme accuracy and spectral grasp that can potentially revolutionize exoplanet detection. Here we demonstrate a laser frequency comb using an alternate comb generation method based on electro-optical modulation, with the comb centre wavelength stabilized to a molecular or atomic reference. In contrast to mode-locked combs, the line spacing is readily resolvable using typical astronomical grating spectrographs. Built using commercial off-the-shelf components, the instrument is relatively simple and reliable. Proof of concept experiments operated at near-infrared wavelengths were carried out at the NASA Infrared Telescope Facility and the Keck-II telescope.We describe a successful effort to produce a laser comb around 1.55 μm in the astronomical H band (1.5-1.8 μm) using a method based on a line-referenced, electrooptical-modulation frequency comb (LR-EOFC). We discuss the experimental setup, laboratory results and proof of concept demonstrations at the NASA Infrared Telescope Facility (IRTF) and the Keck-II telescope. The laser comb has a demonstrated stability of < 200 kHz, corresponding to a Doppler precision of ∼ 0.3 m s−1. This technology, when coupled with a high spectral resolution spectrograph, offers the promise of ∼ 1 m s−1 radial velocity precision suitable for the detection of Earth-sized planets in the habitable zones of cool M-type stars. ar X iv :1 50 1. 02 50 9v 1 [ as tr oph .I M ] 1 1 Ja n 20 15

Bimodal distribution of sulfuric acid aerosols in the upper haze of Venus

Observations by the SPICAV/SOIR instruments aboard Venus Express have revealed that the upper haze (UH) of Venus, between 70 and 90 km, is variable on the order of days and that it is populated by two particle modes. We use a one-dimensional microphysics and vertical transport model based on the Community Aerosol and Radiation Model for Atmospheres to evaluate whether interaction of upwelled cloud particles and sulfuric acid particles nucleated in situ on meteoric dust are able to generate the two observed modes, and whether their observed variability are due in part to the action of vertical transient winds at the cloud tops. Nucleation of photochemically produced sulfuric acid onto polysulfur condensation nuclei generates mode 1 cloud droplets, which then diffuse upwards into the UH. Droplets generated in the UH from nucleation of sulfuric acid onto meteoric dust coagulate with the upwelled cloud particles and therefore cannot reproduce the observed bimodal size distribution. By comparison, the mass transport enabled by transient winds at the cloud tops, possibly caused by sustained subsolar cloud top convection, are able to generate a bimodal size distribution in a time scale consistent with Venus Express observations. Below the altitude where the cloud particles are generated, sedimentation and vigorous convection causes the formation of large mode 2 and mode 3 particles in the middle and lower clouds. Evaporation of the particles below the clouds causes a local sulfuric acid vapor maximum that results in upwelling of sulfuric acid back into the clouds. In the case where the polysulfur condensation nuclei are small and their production rate is high, coagulation of small droplets onto larger droplets in the middle cloud may set up an oscillation in the size modes of the particles such that precipitation of sulfuric acid “rain” may be possible immediately below the clouds once every few Earth months. Reduction of the polysulfur condensation nuclei production rate destroys this oscillation and reduces the mode 1 particle abundance in the middle cloud by two orders of magnitude. However, it better reproduces the sulfur-to-sulfuric-acid mass ratio in the cloud and haze droplets as constrained by fits to UV reflectivity data. In general we find satisfactory agreement between our nominal and transient wind results and observations from Pioneer Venus, Venus Express, and Magellan, though improvements could be made by incorporating sulfur microphysics.

Nonhydrostatic effects and the determination of icy satellites’ moment of inertia

We compare the moment of inertia (MOI) of a simple hydrostatic, two layer body as determined by the Radau–Darwin Approximation (RDA) to its exact hydrostatic MOI calculated to first order in the parameter q = Ω^2R^3/GM, where Ω, R, and M are the spin angular velocity, radius, and mass of the body, and G is the gravitational constant. We show that the RDA is in error by less than 1% for many configurations of core sizes and layer densities congruent with those of solid bodies in the Solar System. We then determine the error in the MOI of icy satellites calculated with the RDA due to nonhydrostatic effects by using a simple model in which the core and outer shell have slight degree 2 distortions away from their expected hydrostatic shapes. Since the hydrostatic shape has an associated stress of order ρΩ^2R^2 (where ρ is density) it follows that the importance of nonhydrostatic effects scales with the dimensionless number σ/ρΩ^2R^2, where σ is the nonhydrostatic stress. This highlights the likely importance of this error for slowly rotating bodies (e.g., Titan and Callisto) and small bodies (e.g., Saturn moons other than Titan). We apply this model to Titan, Callisto, and Enceladus and find that the RDA-derived MOI can be 10% greater than the actual MOI for nonhydrostatic stresses as small as ∼0.1 bars at the surface or ∼1 bar at the core–mantle boundary, but the actual nonhydrostatic stresses for a given shape change depends on the specifics of the interior model. When we apply this model to Ganymede we find that the stresses necessary to produce the same MOI errors as on Titan, Callisto, and Enceladus are an order of magnitude greater due to its faster rotation, so Ganymede may be the only instance where RDA is reliable. We argue that if satellites can reorient to the lowest energy state then RDA will always give an overestimate of the true MOI. Observations have shown that small nonhydrostatic gravity anomalies exist on Ganymede and Titan, at least at degree 3 and presumably higher. If these anomalies are indicative of the nonhydrostatic anomalies at degree 2 then these imply only a small correction to the MOI, even for Titan, but it is possible that the physical origin of nonhydrostatic degree 2 effects is different from the higher order terms. We conclude that nonhydrostatic effects could be present to an extent that allows Callisto and Titan to be fully differentiated.

Structure and composition of Pluto's atmosphere from the New Horizons solar ultraviolet occultation

The Alice instrument on NASAs New Horizons spacecraft observed an ultraviolet solar occultation by Plutos atmosphere on 2015 July 14. The transmission vs. altitude was sensitive to the presence of N_2, CH_4, C_2H_2, C_2H_4, C_2H_6, and haze. We derived line-of-sight abundances and local number densities for the 5 molecular species, and line-of-sight optical depth and extinction coefficients for the haze. We found the following major conclusions: (1) We confirmed temperatures in Plutos upper atmosphere that were colder than expected before the New Horizons flyby, with upper atmospheric temperatures near 65–68 K. The inferred enhanced Jeans escape rates were (3–7) × 10^(22) N_2 s^(−1) and (4–8) × 10^(25) CH_4 s^(−1) at the exobase (at a radius of ∼ 2900 km, or an altitude of ∼1710 km). (2) We measured CH_4 abundances from 80 to 1200 km above the surface. A joint analysis of the Alice CH_4 and Alice and REX N_2 measurements implied a very stable lower atmosphere with a small eddy diffusion coefficient, most likely between 550 and 4000 cm^2 s^(−1). Such a small eddy diffusion coefficient placed the homopause within 12 km of the surface, giving Pluto a small planetary boundary layer. The inferred CH_4 surface mixing ratio was ∼ 0.28–0.35%. (3) The abundance profiles of the “C_2H_x hydrocarbons” (C_2H_2, C_2H_4, C_2H_6) were not simply exponential with altitude. We detected local maxima in line-of-sight abundance near 410 km altitude for C_2H_4, near 320 km for C_2H_2, and an inflection point or the suggestion of a local maximum at 260 km for C_2H_6. We also detected local minima near 200 km altitude for C_2H_4, near 170 km for C_2H_2, and an inflection point or minimum near 170–200 km for C_2H_6. These compared favorably with models for hydrocarbon production near 300–400 km and haze condensation near 200 km, especially for C_2H_2 and C_2H_4 (Wong et al., 2017). (4) We found haze that had an extinction coefficient approximately proportional to N_2 density.

Design and Construction of Absorption Cells for Precision Radial Velocities in the K Band Using Methane Isotopologues

We present a method to optimize absorption cells for precise wavelength calibration in the near-infrared. We apply it to design and optimize methane isotopologue cells for precision radial velocity measurements in the K band. We also describe the construction and installation of two such cells for the CSHELL spectrograph at NASA’s IRTF. We have obtained their high-resolution laboratory spectra, which we can then use in precision radial velocity measurements and which can also have other applications. In terms of obtainable RV precision, methane should outperform other proposed cells, such as the ammonia cell (^(14)NH_3) recently demonstrated on CRIRES/VLT. The laboratory spectra of the ammonia and methane cells show strong absorption features in the H band that could also be exploited for precision Doppler measurements. We present spectra and preliminary radial velocity measurements obtained during our first-light run. These initial results show that a precision down to 20-30 m s^(-1)can be obtained using a wavelength interval of only 5 nm in the K band and S/N ∼ 150. This supports the prediction that a precision down to a few meters per second can be achieved on late-M dwarfs using the new generation of NIR spectrographs, thus enabling the detection of terrestrial planets in their habitable zones. Doppler measurements in the NIR can also be used to mitigate the radial velocity jitter due to stellar activity, enabling more efficient surveys on young active stars.

Constraints on the microphysics of Pluto's photochemical haze from New Horizons observations

The New Horizons flyby of Pluto confirmed the existence of hazes in its atmosphere. Observations of a large high- to low- phase brightness ratio, combined with the blue color of the haze (indicative of Rayleigh scattering), suggest that the haze particles are fractal aggregates, perhaps analogous to the photochemical hazes on Titan. Therefore, studying the Pluto hazes can shed light on the similarities and differences between the Pluto and Titan atmospheres. We model the haze distribution using the Community Aerosol and Radiation Model for Atmospheres assuming that the distribution is shaped by downward transport and coagulation of particles originating from photochemistry. Hazes composed of both purely spherical and purely fractal aggregate particles are considered. General agreement between model results and solar occultation observations is obtained with aggregate particles when the downward mass flux of photochemical products is equal to the column-integrated methane destruction rate ∼1.2 × 10^(−14) g cm^(−2) s^(−1), while for spherical particles the mass flux must be 2–3 times greater. This flux is nearly identical to the haze production flux of Titan previously obtained by comparing microphysical model results to Cassini observations. The aggregate particle radius is sensitive to particle charging effects, and a particle charge to radius ratio of 30 e − /µm is necessary to produce ∼0.1–0.2 µm aggregates near Plutos surface, in accordance with forward scattering measurements. Such a particle charge to radius ratio is 2–4 times higher than those previously obtained for Titan. Hazes composed of spheres with the same particle charge to radius ratio have particles that are 4 times smaller at Plutos surface. These results further suggest that the haze particles are fractal aggregates. We also consider the effect of condensation of HCN, C_2H_2, C_2H_4, and C_2H_6 on the haze particles, which may play an important role in shaping their altitude and size distributions.

Vertical Distribution of C_3-hydrocarbons in the Stratosphere of Titan

Motivated by the recent detection of propene (C_3H_6) in the atmosphere of Titan, we use a one-dimensional Titan photochemical model with an updated eddy diffusion profile to systematically study the vertical profiles of the stable species in the C_3-hydrocarbon family. We find that the stratospheric volume mixing ratio of propene (C_3H_6) peaks at 150 km with a value of 5 × 10^(−9), which is in good agreement with recent observations by the Composite Infrared Spectrometer on the Cassini spacecraft. Another important species that is currently missing from the hydrocarbon family in Titans stratosphere is allene (CH_2CCH_2), an isomer of methylacetylene (CH_3C_2H). We predict that its mixing ratio in the stratosphere is about 10^(−9), which is on the margin of the detection limit. CH_2CCH_2 and CH_3C_2H equilibrate at a constant ratio in the stratosphere by hydrogen-exchanging reactions. Thus, by precisely measuring the ratio of CH_2CCH_2 to CH_3C_2H, the abundance of atomic hydrogen in the atmosphere can be inferred. No direct yield for the production of cyclopropane (c-C_3H_6) is available. From the discharge experiments of Navarro-Gonzalez & Ramirez, the abundance of cyclopropane is estimated to be 100 times less than that of C_3H_6.

Aggregate particles in the plumes of Enceladus

Estimates of the total particulate mass of the plumes of Enceladus are important to constrain theories of particle formation and transport at the surface and interior of the satellite. We revisit the calculations of Ingersoll and Ewald (Ingersoll, A.P., Ewald, S.P. [2011]. Icarus 216(2), 492–506), who estimated the particulate mass of the Enceladus plumes from strongly forward scattered light in Cassini ISS images. We model the plume as a combination of spherical particles and irregular aggregates resulting from the coagulation of spherical monomers, the latter of which allows for plumes of lower particulate mass. Though a continuum of solutions are permitted by the model, the best fits to the ISS data consist either of low mass plumes composed entirely of small aggregates or high mass plumes composed of mostly spheres. The high particulate mass plumes have total particulate masses of (166 ± 42) × 10^3 kg, consistent with the results of Ingersoll and Ewald (Ingersoll, A.P., Ewald, S.P. [2011]. Icarus 216(2), 492–506). The low particulate mass plumes have masses of (25 ± 4) × 10^3 kg, leading to a solid to vapor mass ratio of 0.07 ± 0.01 for the plume. If indeed the plumes are made of such aggregates, then a vapor-based origin for the plume particles cannot be ruled out. Finally, we show that the residence time of the monomers inside the plume vents is sufficiently long for Brownian coagulation to form the aggregates before they are ejected to space.