Performance Of Future Systems Research Articles

Understanding and Improving the Latency of DRAM-Based Memory Systems

Over the past two decades, the storage capacity and access bandwidth of main memory have improved tremendously, by 128x and 20x, respectively. These improvements are mainly due to the continuous technology scaling of DRAM (dynamic random-access memory), which has been used as the physical substrate for main memory. In stark contrast with capacity and bandwidth, DRAM latency has remained almost constant, reducing by only 1.3x in the same time frame. Therefore, long DRAM latency continues to be a critical performance bottleneck in modern systems. Increasing core counts, and the emergence of increasingly more data-intensive and latency-critical applications further stress the importance of providing low-latency memory accesses. In this dissertation, we identify three main problems that contribute significantly to long latency of DRAM accesses. To address these problems, we present a series of new techniques. Our new techniques significantly improve both system performance and energy efficiency. We also examine the critical relationship between supply voltage and latency in modern DRAM chips and develop new mechanisms that exploit this voltage-latency trade-o to improve energy efficiency. First, while bulk data movement is a key operation in many applications and operating systems, contemporary systems perform this movement inefficiently, by transferring data from DRAM to the processor, and then back to DRAM, across a narrow o -chip channel. The use of this narrow channel for bulk data movement results in high latency and high energy consumption. This dissertation introduces a new DRAM design, Low-cost Inter-linked SubArrays (LISA), which provides fast and energy-efficient bulk data movement across sub- arrays in a DRAM chip. We show that the LISA substrate is very powerful and versatile by demonstrating that it efficiently enables several new architectural mechanisms, including low-latency data copying, reduced DRAM access latency for frequently-accessed data, and reduced preparation latency for subsequent accesses to a DRAM bank. Second, DRAM needs to be periodically refreshed to prevent data loss due to leakage. Unfortunately, while DRAM is being refreshed, a part of it becomes unavailable to serve memory requests, which degrades system performance. To address this refresh interference problem, we propose two access-refresh parallelization techniques that enable more overlap- ping of accesses with refreshes inside DRAM, at the cost of very modest changes to the memory controllers and DRAM chips. These two techniques together achieve performance close to an idealized system that does not require refresh. Third, we find, for the first time, that there is significant latency variation in accessing different cells of a single DRAM chip due to the irregularity in the DRAM manufacturing process. As a result, some DRAM cells are inherently faster to access, while others are inherently slower. Unfortunately, existing systems do not exploit this variation and use a fixed latency value based on the slowest cell across all DRAM chips. To exploit latency variation within the DRAM chip, we experimentally characterize and understand the behavior of the variation that exists in real commodity DRAM chips. Based on our characterization, we propose Flexible-LatencY DRAM (FLY-DRAM), a mechanism to reduce DRAM latency by categorizing the DRAM cells into fast and slow regions, and accessing the fast regions with a reduced latency, thereby improving system performance significantly. Our extensive experimental characterization and analysis of latency variation in DRAM chips can also enable development of other new techniques to improve performance or reliability. Fourth, this dissertation, for the first time, develops an understanding of the latency behavior due to another important factor { supply voltage, which significantly impacts DRAM performance, energy consumption, and reliability. We take an experimental approach to understanding and exploiting the behavior of modern DRAM chips under different supply voltage values. Our detailed characterization of real commodity DRAM chips demonstrates that memory access latency reduces with increasing supply voltage. Based on our characterization, we propose Voltron, a new mechanism that improves system energy efficiency by dynamically adjusting the DRAM supply voltage based on a performance model. Our extensive experimental data on the relationship between DRAM supply voltage, latency, and reliability can further enable developments of other new mechanisms that improve latency, energy efficiency, or reliability. The key conclusion of this dissertation is that augmenting DRAM architecture with simple and low-cost features, and developing a better understanding of manufactured DRAM chips together leads to significant memory latency reduction as well as energy efficiency improvement. We hope and believe that the proposed architectural techniques and detailed experimental data on real commodity DRAM chips presented in this dissertation will enable developments of other new mechanisms to improve the performance, energy efficiency, or reliability of future memory systems.

Evaluating real-life performance of the state-of-the-art in facial expression recognition using a novel YouTube-based datasets

Facial expression recognition (FER) is one of the most active areas of research in computer science, due to its importance in a large number of application domains. Over the years, a great number of FER systems have been implemented, each surpassing the other in terms of classification accuracy. However, one major weakness found in the previous studies is that they have all used standard datasets for their evaluations and comparisons. Though this serves well given the needs of a fair comparison with existing systems, it is argued that this does not go in hand with the fact that these systems are built with a hope of eventually being used in the real-world. It is because these datasets assume a predefined camera setup, consist of mostly posed expressions collected in a controlled setting, using fixed background and static ambient settings, and having low variations in the face size and camera angles, which is not the case in a dynamic real-world. The contributions of this work are two-fold: firstly, using numerous online resources and also our own setup, we have collected a rich FER dataset keeping in mind the above mentioned problems. Secondly, we have chosen eleven state-of-the-art FER systems, implemented them and performed a rigorous evaluation of these systems using our dataset. The results confirm our hypothesis that even the most accurate existing FER systems are not ready to face the challenges of a dynamic real-world. We hope that our dataset would become a benchmark to assess the real-life performance of future FER systems.

Vulnerability to climate change impacts of present renewable energy systems designed for achieving net-zero energy buildings

There are high hopes and expectations in the potential of Net-Zero Energy Buildings (NZEB) to contribute to combat the adverse environmental effects from intensive human activity, including energy use. An important related question that has received insufficient attention and is therefore addressed in this study, is how NZEB equipped with renewable energy (RE) systems would perform in future climates that will be caused by the ongoing environmental impacts. In this research, downscaled future hourly weather data from the Global Climate Models (GCM) are used to predict future performance of RE systems for low energy using residential buildings in 10 different climate zones in the U.S. Renewable energy systems with different configurations of Photovoltaic (PV) solar energy and wind system power generation are modeled and coupled with hourly building energy loads. The research results show that buildings with the present configurations of RE will be losing their capability to meet the zero-energy goal in half of the considered climate zones. It was found that the RE systems for a future NZEB should be resized and reconfigured to accommodate climate change impacts. RE systems prioritizing PV systems shows good stability and performance in power generation under expected future climate conditions. A criterion was developed to assist a proposed grid search method for finding the RE system configurations for future NZEB that would identify vulnerabilities of present NZEB's performance under climate change.

Impact of Next Generation Electrode Materials on Abuse Response

The application space for lithium ion cells is rapidly expanding to include transportation, grid storage, space system, and military power needs. Increasingly demanding mission requirements in these applications has necessitated higher energy and higher power energy storage systems. As the size and capacity of these power systems increases, so do the consequences of an off normal battery event (thermal runaway, short circuit, mechanical crush, etc.). Significant scientific efforts have been made to understand the consequences and requirements for runaway, understand policy decisions pertaining to battery safety, understand safety response in packs, and model thermal abuse of cells. Despite great progress, the overall technology still poses safety concerns for the public which, along with other considerations, have hindered the widespread adoption of many lithium based storage technologies. Several high-profile safety occurrences in recent history have brought the subject of lithium ion battery safety into the forefront of discussion within the community. While many solutions have been proposed, there has yet to be a robust and cost effective solution for the issue of battery safety, particularly in high capacity storage systems. Additionally, next generation materials such as silicon graphite/graphene anodes and high voltage cathodes are posed to increase the electrochemical performance of future systems, but little is known with regards to the abuse response of these materials. Our research efforts have focused on developing understanding and strategies to develop intrinsically safe lithium ion based systems. This material science based strategy to ensure battery safety offers benefits over engineered solutions to battery safety. As next generation materials mature, it becomes increasingly important to apply similar methodologies to understand and optimize the abuse response and safety envelope for these batteries. Intrinsically safe battery research topics have focused on materials development and coating efforts for anode and cathode safety, electrolyte flammability and reactivity reduction efforts, separator instability considerations, and solutions targeted at the mitigating the energy released in the case of a cell abuse. The understanding of next generation silicon carbon composites will be discussed to compare with current graphite based systems. Industrially relevant 18650 cell were built using silicon carbon composite anodes versus NMC 523 cathodes to understand both electrochemical and abuse performance, see Figure 1. Initial experiments show similar performance for specific capacity and rate capability for these cells. However, these next generation anode materials appear to have some potential differences in the degradation mechanisms and overall response during abuse conditions. These results may help to inform future development efforts, particularly as industry pushes towards higher silicon content in anode materials or towards significantly higher energy dense materials. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. Figure 1

(Invited) Diverse Accessible Heterogeneous Integration (DAHI) Foundry at Northrop Grumman Aerospace Systems (NGAS)

The performance, size, weight and power requirements for future systems are increasingly demanding. These can be met by intimate integration of lower power and higher performance scaled CMOS and compound semiconductor technologies into smaller areas and volumes. The more adaptable and more intimate is this integration between two or more technologies, the more flexibility is given to the designer for the selection of the technology for a specific function, or even better, for an optimum combination of different transistor technologies in the same function or cell in the design. Northrop Grumman Aerospace Systems (NGAS) under the Diverse Accessible Heterojunction Integration (DAHI) DARPA program is developing integration processes, design kits and thermal simulation tools to integrate submicron CMOS, InP HBT, GaN HEMT and high-Q passive technologies for advanced DoD systems. We have demonstrated integration of NGAS’ InP HBT and GaN HEMT technologies on 65nm and 45nm CMOS wafers.

An Experimental Investigation of Acoustic Black Hole Dynamics at Low, Mid, and High Frequencies

The acoustic black hole (ABH) has been developed in recent years as an effective, passive, and lightweight method for attenuating bending wave vibrations in beams and plates and reducing the sound radiation and structural-acoustic response of structures. The ABH effect utilizes a local change in the plate or beam thickness to reduce the bending wave speed and increase the transverse vibration amplitude. Attaching a viscoelastic damping layer to the ABH results in effective energy dissipation and vibration reduction. Surface-averaged mobility and radiated sound power measurements were performed on an aluminum plate containing an array of 20 two-dimensional ABHs with damping layers and compared to a similar uniform plate. Detailed laser vibrometer scans of an ABH cell (including the ABH and surrounding homogeneous plate) were also performed to analyze the vibratory characteristics of individual ABH cells and compared with mode shapes calculated using finite elements. The results showed that the surface-averaged mobility was reduced by up to 14 dB for the fully damped ABH plate compared to a uniform reference plate while also reducing the mass of the plate. The results demonstrated that the dynamics of plates with embedded ABHs can be characterized by low, mid, and high frequency ranges, with low-order local ABH modes contributing significantly to low frequency ABH performance. The effects of damping layer thickness and diameter were also investigated to assess ABH performance optimization. It was shown that the damping layer can have the added benefit of mass loading the ABH and enhancing low frequency performance. The results will be useful for designing the low frequency performance of future ABH systems and describing ABH performance in terms of design parameters.

Asymmetry-aware work-stealing runtimes

Amdahl's law provides architects a compelling reason to introduce system asymmetry to optimize for both serial and parallel regions of execution. Asymmetry in a multicore processor can arise statically (e.g., from core microarchitecture) or dynamically (e.g., applying dynamic voltage/frequency scaling). Work stealing is an increasingly popular approach to task distribution that elegantly balances task-based parallelism across multiple worker threads. In this paper, we propose asymmetry-aware work-stealing (AAWS) runtimes, which are carefully designed to exploit both the static and dynamic asymmetry in modern systems. AAWS runtimes use three key hardware/software techniques: work-pacing, work-sprinting, and work-mugging. Work-pacing and work-sprinting are novel techniques that combine a marginal-utility-based approach with integrated voltage regulators to improve performance and energy efficiency in high- and low-parallel regions. Work-mugging is a previously proposed technique that enables a waiting big core to preemptively migrate work from a busy little core. We propose a simple implementation of work-mugging based on lightweight user-level interrupts. We use a vertically integrated research methodology spanning software, architecture, and VLSI to make the case that holistically combining static asymmetry, dynamic asymmetry, and work-stealing runtimes can improve both performance and energy efficiency in future multicore systems.

UltraDense Networks: The New Wireless Frontier for Enabling 5G Access

The extreme traffic load that future wireless networks are expected to accommodate requires a rethinking of the system design. The initial estimations indicate that, unlike the evolutionary path of previous cellular generations that was based on spectral efficiency improvements, the most substantial amount of future system performance gains will be obtained by means of network infrastructure densification. By increasing the density of operator-deployed infrastructure elements along with incorporation of user-deployed access nodes (ANs) and mobile user devices acting as infrastructure prosumers, having one or more ANs exclusively dedicated to each user is expected to become feasible, introducing the ultradense network (UDN) paradigm. Although it is clear that UDNs are able to take advantage of the significant benefits provided by proximal transmissions and increased spatial reuse of system resources, large node density and irregular deployment introduce new challenges, mainly due to the interference environment characteristics that are vastly different from previous cellular deployments. This article attempts to provide insights on fundamental issues related to UDN deployment, such as determining the infrastructure density required to support the given traffic load requirements and the benefits of network-wise coordination, demonstrating the potential of UDNs for fifth-generation (5G) wireless networks.

Back to the Future: Current-Mode Processor in the Era of Deeply Scaled CMOS

This paper explores the use of MOS current-mode logic (MCML) as a fast and low noise alternative to static CMOS circuits in microprocessors, thereby improving the performance, energy efficiency, and signal integrity of future computer systems. The power and ground noise generated by an MCML circuit is typically 10– $100 \times $ smaller than the noise generated by a static CMOS circuit. Unlike static CMOS, whose dominant dynamic power is proportional to the frequency, MCML circuits dissipate a constant power independent of clock frequency. Although these traits make MCML highly energy efficient when operating at high speeds, the constant static power of MCML poses a challenge for a microarchitecture that operates at the modest clock rate and with a low activity factor. To address this challenge, a single-core microarchitecture for MCML is explored that exploits the $C$ -slow retiming technique, and operates at a high frequency with low complexity to save energy. This design principle contrasts with the contemporary multicore design paradigm for static CMOS that relies on a large number of gates operating in parallel at the modest speeds. The proposed architecture generates 10– $40\times $ lower power and ground noise, and operates within 13% of the performance (i.e., 1/ExecutionTime) of a conventional, eight-core static CMOS processor while exhibiting $1.6\times $ lower energy and 9% less area. Moreover, the operation of an MCML processor is robust under both systematic and random variations in transistor threshold voltage and effective channel length.

Comparison of an analog direct detection and a micropulse aerosol lidar at 1.5 - μ m wavelength for wind field observations—with first results over the ocean

The performance of two direct-detection atmospheric lidar systems with very different methods of generating and detecting laser radiation is compared as the result of a field experiment held in March 2015, in Chico, California. During the noncontinuous, 11-day test period, in which the systems operated side by side, the micropulse lidar was operated at its maximum pulse repetition frequency (15 kHz) and integrated elastic backscatter over the interpulse period of the analog direct-detection lidar (0.1 s). Operation at the high pulse repetition frequency resulted in second-trip echoes that contaminated portions of the data. The performance of the micropulse lidar varied with background brightness—as expected with a photon-counting receiver—yet showed equal or larger backscatter intensity signal-to-noise ratio throughout the experiment. Examples of wind fields and time series of wind vectors from both systems during the Chico experiment are presented. In addition, scans over the ocean that were collected by the micropulse lidar during a subsequent deployment on the northern California coast are presented. We conclude by reviewing the advantages and disadvantages of each system and make some suggestions to improve the design and performance of future systems.

The human factor in production planning and control: considering human needs in computer aided decision-support systems

The unawareness of production planners about the interaction of key performance indicators (KPIs) in manufacturing systems and misinterpretations of system states often leads to problems when they try to improve them. Exemplary, the lead time syndrome (LTS) represents one of these counterproductive actions. While planners aim to improve due date reliability by planned lead time adjustments, the result is often an aggravation of due date reliability. This is due to human behaviour in the decision-making process, which can be biased. Despite its potential to improve the logistic performance, human needs were out of scope of recent investigations in the field of logistics. The aim of this paper is to close this research gap by means of a case study and research on human behaviour in the field of psychology. The result is the determination of guidelines to improve the logistic performance in future decision-support systems by considering human needs.

ADAMANT: Tools to Capture, Analyze, and Manage Data Movement

In the converging world of High Performance Computing and Big Data, moving data is becoming a critical aspect of performance and energy efficiency. In this paper we present the Advanced DAta Movement Analysis Toolkit (ADAMANT), a set of tools to capture and analyze data movement within an application, and to aid in understanding performance and energy efficiency in current and future systems. ADAMANT identifies all the data objects allocated by an application and uses instrumentation modules to monitor relevant events (e.g. cache misses). Finally, ADAMANT produces a per-object performance profile.In this paper we demonstrate the use of ADAMANT in analyzing three applications, BT, BFS, and Velvet, and evaluate the impact of different memory technology. With the information produced by ADAMANT we were able to model and compare different memory configurations and object placement solutions. In BFS we devised a placement which outperforms caching, while in the other two cases we were able to point out which data objects may be problematic for the configurations explored, and would require refactoring to improve performance.

Microwave photonic delay line signal processing.

This paper provides a path for the design of state-of-the-art fiber-optic delay lines for signal processing. The theoretical forms for various radio-frequency system performance metrics are derived for four modulation types: X- and Z-cut Mach-Zehnder modulators, a phase modulator with asymmetric Mach-Zehnder interferometer, and a polarization modulator with control waveplate and polarizing beam splitter. Each modulation type is considered to cover the current and future needs for ideal system designs. System gain, compression point, and third-order output intercept point are derived from the transfer matrices for each modulation type. A discussion of optical amplifier placement and fiber-effect mitigation is offered. The paper concludes by detailing two high-performance delay lines, built for unique applications, that exhibit performance levels an order of magnitude better than commercial delay lines. This paper should serve as a guide to maximizing the performance of future systems and offer a look into current and future research being done to further improve photonics technologies.

Smart Antennas for Communications in Line of Sight

Throughout the world, including the United States, there is significant research and development on smart antennas for wireless systems. This is because smart antennas have tremendous potential to enhance the performance of future generation wireless systems as evidenced by the antennas? recent deployment in many systems. It covers smart antenna technology, including software and system aspects. First the two basic types of smart antennas, adaptive and phased arrays are described and then their current use and proposed use in future wireless systems is discussed. A smart antenna is therefore a phased or adaptive array that adjusts to the environment. That is, for the adaptive array, the beam pattern changes as the desired user and the interference move; and for the phased array the beam is steered or different beams are selected as the desired user moves. We describe and analyze adaptive antenna array technology to improve naval communications systems to beyond line-of-sight at microwave frequencies by using the ducting layer as a leaky waveguide and the adaptive array to resolve and coherently combine multipath in this layer.

What Matters Most in Transportation Demand Model Specifications: A Comparison of Outputs in a Mid-size Network

This paper examines the impact of travel demand modeling (TDM) disaggregation techniques in the context of medium-sized communities. Specific TDM improvement strategies are evaluated for predictive power and flexibility with case studies based on the Tyler, Texas, network. Results suggest that adding time-of-day disaggregation, particularly in conjunction with multi-class assignment, to a basic TDM framework has the most significant impacts on outputs. Other strategies shown to impact outputs include adding a logit mode choice model and incorporating a congestion feedback loop. For resource-constrained communities, these results show how model output and flexibility vary for different settings and scenarios.BACKGROUND Transportation directly provides for the mobility of people and goods, while influencing land use patterns and economic activity, which in turn affect air quality, social equity, and investment decisions. Driven by the need to forecast future transportation demand and system performance, Manheim (1979) and Florian et al. (1988) introduced a transportation analysis framework for traffic forecasting using aggregated data that provide the basis for what is known as the four-step model: a process involving trip generation, then trip distribution and mode choice, followed by route choice. Aggregating demographic data at the zone level, the four-step model generates trip productions based on socioeconomic data (e.g., household counts by income and size) and trip attractions primarily based on jobs counts. The model then proportionally distributes trips between each origin and destination (OD) zone pair based on competing travel attractions and impedances, under the assumption that OD pairings with higher travel costs draw fewer trips. Trips between each OD pair are split among a variety of transportation modes, allocating trips to private vehicle, transit, or other

Control and Monitoring for Grid-Friendly Wind Turbines: Research Overview and Suggested Approach

Recent years have witnessed a continuing increase in the wind turbine (WT) size and penetration level into the grid, revealing critical challenges to sustainable market development and urgent demands of low voltage ride-through, load reduction, and optimal operation. This paper presents an overview on grid-friendly WTs and relevant technologies for control and monitoring. The development trends are discussed as well to explore potential approaches for optimizing WT systems. The main body contains three parts: the “grid-friendly” criterion and a prototype, the overview and suggestions for WT controller design concerning load reduction and power optimization, and suggested solutions for wind power monitoring systems. This paper is expected to be a reference for extending the knowledge of grid-friendly WTs and optimizing the performance of future large-scale wind power systems.

Perturbation Methods for Real-Time In Situ Evaluation of Hot-Side Thermal Resistances in Thermoelectric Energy Recovery Systems

Thermoelectric (TE) power systems in high-temperature industrial, transportation, and military energy systems require high-performance hot-side and cold-side heat transfer to provide the critical temperature differential and transfer the required thermal energy to create the power output. Hot- and cold-side heat transfer performance is typically characterized by the hot-side and cold-side thermal resistance, Rh,th and Rc,th, respectively. This heat transfer performance determines the hot-side temperature, Th, and cold-side temperature, Tc, conditions when operating in energy recovery environments with available temperature differentials characterized by an external driving temperature, Tsrc, and ambient temperature, Tamb. It is crucial to monitor and track the hot-side thermal performance at all times during TE energy recovery system operation, thereby allowing one to track the system “health,” predict future expected system performance, and anticipate/prevent system failures. This paper describes the use of a perturbation methodology and a direct coupling between the TE current, voltage, and hot-side energy flow to extract a real-time in situ evaluation of hot-side thermal resistances. External measurable TE parameters, either system current or Tsrc, can be perturbed during system operation, and the resulting TE system response can then be coupled mathematically to the hot-side thermal transfer performance (i.e., thermal resistance). This paper discusses the mathematical formalism of this technique, and TE module experimental data showing successful application of real-time current perturbation. This technique provides a pathway for developing faster, real-time system monitoring and diagnostics to alleviate system performance degradation, or prevent system damage from dramatic changes in hot-side thermal transfer conditions in industrial, transportation, and spacecraft TE power systems.

Application of Trip Generation Models for Urban Transport Planning In a Data Scarce Developing Country City: The Case of Dar es Salaam

With the increase in travel demand and traffic management problems in many developingcountries cities, travel demand forecasting models are being employed increasingly tomake informed decisions about the operational improvements to the existing transportationsystem and the design and performance of future transportation systems. The mainadvantage of using travel demand forecasting models for such purposes is that they arecapable of capturing the interactive effects of different components of the system understudy. However, for some time now there have been concerns about the application oftransport planning models in developing countries. The concerns have been mainly inrelation to the variables, coefficients and models borrowed from developed countries. Thispaper first discusses the characteristics of transport problems in developing cities andprovides a review of trip generation modelling approaches. Then, the discussion extendsfurther to cover available data for urban transport planning and trip generation modelsthat have found application in Dar es Salaam, Tanzania. The paper concludes byhighlighting how available data sources and trip generation modelling approach can beimproved to cope with the dynamic conditions in Dar es Salaam.

An SDMA-based MAC protocol for wireless ad hoc networks with smart antennas

A wireless ad hoc network consists of a set of wireless devices. The wireless devices are capable of communicating with each other without the assistance of base stations. Space Division Multiple Access (SDMA) is a new technology designed to optimize the performance of current and future mobile communication systems. In this paper, an SDMA-based MAC protocol (S-MAC) for wireless ad hoc networks with smart antennas is proposed. The proposed protocol exploits the SDMA system to allow reception of more than one packet from spatially separated transmitters. Using SDMA technology provides collision-free access to the communication medium based on the location of a node. The proposed protocol solves the hidden terminal problem, the exposed terminal problem, and the deafness problem. Simulation results demonstrate the effectiveness of the proposed S-MAC in improving throughput and increasing spatial channel reuse.

Derivation of Microwave Refractive Index Structure Constant <formula formulatype="inline"><tex Notation="TeX">$(C_{n}^{2})$</tex> </formula> of the Atmosphere From K-Band Interferometric Phase Measurements

Differential phase data has been collected using a site test interferometer (STI) at K-band for several years at various NASA ground station sites, including the NASA Deep Space Network (DSN) in Goldstone, CA, and the NASA Space Network (SN) sites in White Sands, NM, and Guam. An objective of this work is to characterize the site-dependent atmospheric phase stability to determine the viability of that site for arraying applications. A useful figure of merit to quantify this effect is the determination of the microwave refractive index structure constant, C n 2 , which is highly dependent on variations in the water vapor concentration of the troposphere. In this paper, a method is proposed to derive C n 2 from a two-element site test interferometer utilizing temporal phase structure function information. In this way, a path-averaged C n 2 at each ground station site is determined which can be used in the design and expected system performance of future Ka-band communication systems in an arrayed architecture. Statistical characterization of microwave C n 2 indicates that a mean value of 2.04 ×10 -14 , 2.08×10 -14 , and 9.8×10 -13 is expected for Goldstone, CA, White Sands, NM, and Guam, respectively.