Jiandong Meng

Reducing electricity cost through virtual machine placement in high performance computing clouds

In this paper, we first study the impact of load placement policies on cooling and maximum data center temperatures in cloud service providers that operate multiple geographically distributed data centers. Based on this study, we then propose dynamic load distribution policies that consider all electricity-related costs as well as transient cooling effects. Our evaluation studies the ability of different cooling strategies to handle load spikes, compares the behaviors of our dynamic cost-aware policies to cost-unaware and static policies, and explores the effects of many parameter settings. Among other interesting results, we demonstrate that (1) our policies can provide large cost savings, (2) load migration enables savings in many scenarios, and (3) all electricity-related costs must be considered at the same time for higher and consistent cost savings.

Transient Behavior of Thin-Film Deposition: Coupling Micro- and Macro-Scale Transport

The transient behavior of the thin-film deposition process in a chemical vapor deposition reactor is numerically investigated, considering gallium nitride as the deposited material. The numerical modeling involves coupling of microscale transport mechanisms and chemical reactions at the surface with the macroscale convective heat and mass transfer. The results obtained provide a detailed understanding of the entire process, including the microscale phenomena at the surface. The study also presents information that can be used to control the system operating time and the consumption of input power and precursors, which are generally quite expensive.

Fabrication of Gallium Nitride Films in a Chemical Vapor Deposition Reactor

A numerical study has been carried out on the metalorganic chemical vapor deposition (MOCVD) process for the fabrication of gallium nitride (GaN) thin films, which range from a few nanometers to micrometers in thickness. The numerical study is also coupled with an experimental study on the flow and thermal transport processes in the system. Of particular interest in this study is the dependence of the growth rate of GaN and of the uniformity of the film on the flow, resulting from the choice of various design and operating parameters involved in the MOCVD process. Based on an impingement type rotating-disk reactor, three-dimensional simulations have been preformed to indicate the deposition rate increases with reactor pressure, inlet velocity, and wafer rotating speed, while decreases with the precursor concentration ratio. Additionally, a better film uniformity is caused by reducing the reactor pressure, inlet velocity and wafer rotating speed, and increasing precursor concentration ratio. With the impact of wafer temperature included in this study as well, these results are expected to provide a quantitative basis for the prediction, design, and optimization of the process for the fabrication of GaN devices. The flow and the associated transport processes are discussed in detail on the basis of the results obtained to suggest approaches to improve the uniformity of thin film, minimize fluid loss, and reduce flow recirculation that could affect growth rate and uniformity.

Fabrication of GaN Films in a Chemical Vapor Deposition Reactor

A three-dimensional numerical study has been carried out on the rotating disk GaN MOCVD process, and it is also coupled with an experimental study on the flow and thermal transport processes in the system. An impingement type reactor, with a rotating base, is considered. The dependence of the thin film growth rate and uniformity on operating conditions such as inflow velocity, rotational speed, and susceptor temperature are investigated in detail. Similarly, the effect of the geometry and configuration of the reactor are studied. The study also considers the effect of thermal and solutal buoyancy on the resulting flow. The flow and the associated transport processes are discussed in detail on the basis of the results obtained to suggest approaches to improve the uniformity of the film, minimize fluid loss and reduce flow recirculation that could affect growth rate and uniformity.Copyright

Thermal Transport in the Gallium Nitride Chemical Vapor Deposition Process

A numerical study has been carried out to characterize the metalorganic chemical vapor deposition (MOCVD) growth of Gallium Nitride (GaN) in a rotating-disk reactor. The major objective of this work is to examine the dependence of the growth rate and thin film uniformity on the primary parameters. First of all, for a rotating-disk system, the governing equations involved are obtained. Then, with the effect of thermal buoyancy included and based on the detailed mathematical model and chemical reaction mechanisms, the 3D simulation study is conducted for a rotating reactor. A comparison between the predicted growth rate and experimental data is presented. In addition, the effect of various primary operating and design parameters on the growth rate of GaN and thin-film uniformity is also examined. This provides further insight into the reactor performance and the characteristics of the entire process. The results obtained can also form the basis for the future design and optimization of this system.© 2013 ASME

Transient Behavior of the Gallium Nitride Chemical Vapor Deposition Process

The transient behavior of the Gallium Nitride deposition process in a CVD reactor is numerically investigated. A two-dimensional impinging reactor is considered to examine the time-dependent transport in the MOCVD process, including the steady-state deposition process, and the system start-up and shut-down. The study involves the consideration of complicated transport phenomena, including fluid flow, heat and mass transfer, and chemical reactions between the reactants and the intermediate species. The temperature field and the deposition rate are studied as functions of time, as well as the precursor mass fraction at certain times. The results obtained provide an in-depth understanding about the entire MOCVD process, and the possibility to control the system operating time and the consumption of input power and precursors, which are generally quite expensive. It also provided inputs on the effects of changing operating conditions and the duration of starting and shut down effects.Copyright

Numerical Simulation of GaN Growth in a MOCVD Process

This paper describes a model for the growth of gallium nitride in a vertical impinging metalorganic chemical vapor deposition (MOCVD) reactor. With trimethylgallium (TMGa) and ammonia (NH3 ) carried by hydrogen (H2 ) as precursors, the flow, temperature and concentration profiles are predicted by numerical modeling, which is performed using a commercial CFD software package CFD-ACE+. The growth rate is predicted based on detailed reaction mechanisms given in the literature, and related studies are carried out to verify the reliability and adaptability of the chosen chemical kinetics. A detailed mathematical model is developed first, and the complete chemical mechanisms are introduced. Then, the dependence of the growth rate and uniformity of the deposited layers on operating conditions, such as reactor operating pressure, susceptor temperature, inlet velocity and concentration of the precursors, is investigated to gain greater insight into the reactor performance and characteristics. Based on the simulation results, discussion is presented in this paper to offer the possibility of better control of the GaN film growth process, and to ultimately lead to an optimization of the process, with respect to production rate and film quality.Copyright

Invited) Numerical Simulation of the GaN Growth Process in a MOCVD Process

This paper describes a mathematical model for the growth of gallium nitride in a vertical impinging metalorganic chemical vapor deposition (MOCVD) reactor. With trimethylgallium and ammonia carried by hydrogen as precursors, the flow, temperature and concentration profiles are predicted by numerical modeling, which was carried out using Computational Fluid Dynamics simulation methods. The growth rate was predicted based on comprehensive reaction mechanisms available in the literature, and related studies were carried out to verify the reliability and adaptability of the chosen chemical kinetics.