System-level Performance Metrics Research Articles

Packing optimization of practical systems using a dynamic acceleration methodology

System design is a challenging and time-consuming task which often requires close collaboration between several multidisciplinary design teams to account for complex interactions between components and sub-systems. As such, there is a growing demand in industry to create better performing, efficient, and cost-effective development tools to assist in the system design process. Additionally, the ever-increasing complexity of systems today often necessitates a shift away from manual expertise and a movement towards computer-aided design tools. This work narrows the scope of the system design process by focusing on one critical design aspect: the packaging of system components. The algorithm presented in this paper was developed to optimize the packaging of system components with consideration of practical, system-level functionalities and constraints. Using a dynamic acceleration methodology, the algorithm packages components from an initial position to a final packed position inside of a constrained volume. The motion of components from initial to final positions is driven by several acceleration forces imposed on each component. These accelerations are based on physical interactions between components and their surrounding environment. Various system-level performance metrics such as center of mass alignment and rotational inertia reduction are also considered throughout optimization. Results of several numerical case studies are also presented to demonstrate the functionality and capability of the proposed packaging algorithm. These studies include packaging problems with known optimal solutions to verify the efficacy of the algorithm. Finally, the proposed algorithm was used in a more practical study for the packaging of an urban air mobility nacelle to demonstrate the algorithm’s prospective capabilities in solving real-world packaging problems.

(Invited) Multiscale Diffusion-Reaction Heat Transfer Analysis for Predicting Thermal Runaway in Li-Ion Cells and Battery Packs

Thermal runaway in Li-ion cells and battery packs remains a topic of much importance for ensuring the safety of electrochemical energy conversion and storage systems. Thermal runaway also indirectly affects performance, as conservative overdesign to reduce the risk of thermal runaway affects system-level performance metrics such as energy storage density. Given the considerable challenges in experimental investigation of thermal runaway, theoretical models to predict thermal runaway characteristics of a cell or battery pack are critically needed. Such models can help guide the design of experiments, and ensure that practical systems are reasonably safe without being needlessly overdesigned.This talk will summarize ongoing work on theoretical analysis of multilayer diffusion-reaction heat transfer problems for predicting thermal runaway in Li-ion cells and battery packs. Following a linearization of heat generation terms, the governing energy conservation equation is solved in a multilayer geometry representative of a Li-ion cell comprising multiple electrode and separator layers, or of a battery pack containing several cells. It is mathematically shown that under certain conditions, this problem may admit a finite number of imaginary eigenvalues, which are shown to cause thermal runaway. It is shown that the occurrence of imaginary eigenvalues is governed by a delicate balance between heat generation, thermal conduction within the multilayer body and heat removal from the boundaries, based on which, an explicit threshold condition for thermal runaway is derived. This threshold condition is interpreted in the context of parameter values relevant to Li-ion cells. For example, it is shown that boundary cooling is severely limited in its role to prevent thermal runaway, whereas improving thermal transport within the cell is likely to be a lot more effective. This theoretical model is also extended to a two-layer problem comprising a finite layer adjoined by a semi-infinite body that may model immersion cooling of a Li-ion cell or battery pack. It is shown that while this diffusion-reaction problem is, in principle, unconditionally unstable, the finite time period of any discharge process may allow the temperature field to still remain within a safe envelope, if designed appropriately. Based on this mathematical model, scenarios related to immersion cooling of pouch cells are investigated.The theoretical work described in this talk is broadly relevant to several diffusion-reaction systems. In the particular context of Li-ion cells, these results may form the foundation of techniques to proactively predict and prevent thermal runaway in Li-ion cells and battery packs.

Network Uncertainty Quantification for Analysis of Multi-Component Systems

Abstract In order to impact physical mechanical system design decisions and realize the full promise of high-fidelity computational tools, simulation results must be integrated at the earliest stages of the design process. This is particularly challenging when dealing with uncertainty and optimizing for system-level performance metrics, as full-system models (often notoriously expensive and time-consuming to develop) are generally required to propagate uncertainties to system-level quantities of interest. Methods for propagating parameter and boundary condition uncertainty in networks of interconnected components hold promise for enabling design under uncertainty in real-world applications. These methods avoid the need for time consuming mesh generation of full-system geometries when changes are made to components or subassemblies. Additionally, they explicitly tie full-system model predictions to component/subassembly validation data which is valuable for qualification. These methods work by leveraging the fact that many engineered systems are inherently modular, being comprised of a hierarchy of components and subassemblies that are individually modified or replaced to define new system designs. By doing so, these methods enable rapid model development and the incorporation of uncertainty quantification earlier in the design process. The resulting formulation of the uncertainty propagation problem is iterative. We express the system model as a network of interconnected component models, which exchange solution information at component boundaries. We present a pair of approaches for propagating uncertainty in this type of decomposed system and provide implementations in the form of an open-source software library. We demonstrate these tools on a variety of applications and demonstrate the impact of problem-specific details on the performance and accuracy of the resulting UQ analysis. This work represents the most comprehensive investigation of these network uncertainty propagation methods to date.

Time-Constrained Capacitated Vehicle Routing Problem in Urban E-Commerce Delivery

Electric vehicle routing problems can be particularly complex when recharging must be performed mid-route. In some applications, such as e-commerce parcel delivery truck routing, however, mid-route recharging may not be necessary because of constraints on vehicle capacities and the maximum allowed time for delivery. In this study, we develop a mixed-integer optimization model that exactly solves such a time-constrained capacitated vehicle routing problem, especially of interest for e-commerce parcel delivery vehicles. We compare our solution method with an existing metaheuristic and carry out exhaustive case studies considering four U.S. cities—Austin, TX; Bloomington, IL; Chicago, IL; and Detroit, MI—and two vehicle types: conventional vehicles and battery electric vehicles (BEVs). In these studies we examine the impact of vehicle capacity, maximum allowed travel time, service time (dwelling time to physically deliver the parcel), and BEV range on system-level performance metrics, including vehicle miles traveled (VMT). We find that the service time followed by the vehicle capacity plays a key role in the performance of our approach. We assume an 80-mi BEV range as a baseline without mid-route recharging. Our results show that the BEV range has a minimal impact on performance metrics because the VMT per vehicle averages around 72 mi. In a case study for shared-economy parcel deliveries, we observe that VMT could be reduced by 38.8% in Austin if service providers were to operate their distribution centers jointly.

A Construction Kit for Efficient Low Power Neural Network Accelerator Designs

Implementing embedded neural network processing at the edge requires efficient hardware acceleration that combines high computational throughput with low power consumption. Driven by the rapid evolution of network architectures and their algorithmic features, accelerator designs are constantly being adapted to support the improved functionalities. Hardware designers can refer to a myriad of accelerator implementations in the literature to evaluate and compare hardware design choices. However, the sheer number of publications and their diverse optimization directions hinder an effective assessment. Existing surveys provide an overview of these works but are often limited to system-level and benchmark-specific performance metrics, making it difficult to quantitatively compare the individual effects of each utilized optimization technique. This complicates the evaluation of optimizations for new accelerator designs, slowing-down the research progress. In contrast to previous surveys, this work provides a quantitative overview of neural network accelerator optimization approaches that have been used in recent works and reports their individual effects on edge processing performance. The list of optimizations and their quantitative effects are presented as a construction kit, allowing to assess the design choices for each building block individually. Reported optimizations range from up to 10,000× memory savings to 33× energy reductions, providing chip designers with an overview of design choices for implementing efficient low power neural network accelerators.

Deep Learning Based Intrusion Detection in Cloud Services for Resilience Management

In the global scenario one of the important goals for sustainable development in industrial field is innovate new technology, and invest in building infrastructure. All the developed and developing countries focus on building resilient infrastructure and promote sustainable developments by fostering innovation. At this juncture the cloud computing has become an important information and communication technologies model influencing sustainable development of the industries in the developing countries. As part of the innovations happening in the industrial sector, a new concept termed as ‘smart manufacturing’ has emerged, which employs the benefits of emerging technologies like internet of things and cloud computing. Cloud services deliver an on-demand access to computing, storage, and infrastructural platforms for the industrial users through Internet. In the recent era of information technology the number of business and individual users of cloud services have been increased and larger volumes of data is being processed and stored in it. As a consequence, the data breaches in the cloud services are also increasing day by day. Due to various security vulnerabilities in the cloud architecture; as a result the cloud environment has become non-resilient. To restore the normal behavior of the cloud, detect the deviations, and achieve higher resilience, anomaly detection becomes essential. The deep learning architectures-based anomaly detection mechanisms uses various monitoring metrics characterize the normal behavior of cloud services and identify the abnormal events. This paper focuses on designing an intelligent deep learning based approach for detecting cloud anomalies in real time to make it more resilient. The deep learning models are trained using features extracted from the system level and network level performance metrics observed in the Transfer Control Protocol (TCP) traces of the simulation. The experimental results of the proposed approach demonstrate a superior performance in terms of higher detection rate and lower false alarm rate when compared to the Support Vector Machine (SVM).

Impact of control loop performance on energy use, air quality, and thermal comfort in building systems with advanced sequences of operation

Maintaining control system performance over the lifespan of a building offers great potential for increasing system operation efficiency. Programming the building monitoring system with control loop performance assessment (CLPA) indices provides a way to identify poorly performing loops. This work further advances building control monitoring by developing an approach to help prioritize control problems based on the severity of their system-level impact. CLPA indices were added to a Modelica-based small office building model programmed with advanced heating, ventilating, and air conditioning control sequences. An extensive set of unique simulations of different levels of loop detuning were implemented to generate a database that contains both system-level performance metrics and CLPA indices. A regression model was then developed that combines individual loop performance to assess the impact on system-level outputs. Loops of the zone with higher heat gains and the air handling unit supply air temperature loop produced the greatest system-level impact.

Octans: Optimal Placement of Service Function Chains in Many-Core Systems

Network Function Virtualization (NFV) offers service delivery flexibility and reduces overall costs by running service function chains (SFCs) on commodity servers with many cores. Existing solutions for placing SFCs in one server treat all CPU cores as equal and allocate isolated CPU cores to network functions (NFs). However, advanced servers often adopt Non-Uniform Memory Access (NUMA) architecture to improve the scalability of many-core systems. CPU cores are grouped into nodes, incurring performance degradation due to cross-node memory access and intra-node resource contention. Our evaluation shows that randomly selecting cores to place NFs in an SFC could suffer from 39.2 percent lower throughput comparing to an optimal placement solution. In this article, we propose Octans, an NFV orchestrator to achieve maximum aggregate throughput of all SFCs in many-core systems. Octans first formulates the optimization problem as a Non-Linear Integer Programming (NLIP) Model. Then we identify the key factor for problem solving as evaluating the throughput drop of an NF caused by other NFs in the same SFC or different SFCs, i.e., performance drop index, and propose a formal and accurate prediction model based on system level performance metrics. Finally, we propose two online algorithms to quickly find near-optimal placement solutions for one-time and incremental deployment. Extensive evaluation on a prototype implementation shows that Octans significantly improves the aggregate throughput comparing to two state-of-the-art placement solutions by 27.1 ~ 45.2 percent for one-time deployment and by 20.9 ~ 38.1 percent for incremental deployment, with very low prediction errors. Moreover, Octans could quickly find a near-optimal placement solution with tiny optimality gap.

Key Device and Materials Specifications for a Repeater Enabled Quantum Internet

Entangled photons can be used to create a truly secure communication link between two parties. However, the distance over which this can be achieved is limited by the transmission losses associated with optical fibers. One potential solution is using quantum repeaters (QRs) where initial entanglement is created over short distances and then extended via entanglement swapping. The system-level performance metrics (data rate, fidelity of entanglement, etc.) impose demands upon the hardware components (of a QR), which can be used to guide applied materials and device design toward these objectives. This has become increasingly important with an expanding list of candidates for quantum technologies, as the physical realization of each qubit or detector technology brings its own set of advantages and disadvantages. In this article, we present a framework that uses a modular QR and highlights the tradeoffs that exist between technological component modules. Using reported values, we take a near-term perspective and show the achievable range of rates as a function of distance and the corresponding requirements on matter qubit properties.

On Developing Techniques for Sharing Satellite Spectrum with Indoor Small Cells in 5G

In this paper, we present two spectrum sharing techniques for a multisystem, incorporating an integrated satellite-mobile system and an autonomous terrestrial-mobile system (iSMS/aTMS), namely orthogonal spectrum sharing (OSS) and non-orthogonal spectrum sharing (nOSS) techniques. aTMS consists of numerous small cells deployed in several buildings, and iSMS consists of a satellite station integrated with complementary ground component (CGC) stations deployed within buildings. By exploiting the high external wall penetration loss of a building, the iSMS spectrum is shared with small cells per building in OSS, and small cells per 3-dimensional (3D) cluster per building in nOSS. An interference management scheme, to avoid interference in apartments with collocated CGC stations and small cells, was developed and an optimal number of almost blank subframes (ABSs) per ABS pattern period (APP) was defined. System-level capacity, spectral efficiency, and energy efficiency performance metrics were derived. Furthermore, we present an algorithm for both OSS and nOSS techniques. With extensive simulation and numerical analysis, it is shown that the proposed nOSS significantly outperforms OSS in terms of spectral efficiency and energy efficiency, and both techniques can meet the expected spectral efficiency and energy efficiency requirements for the fifth-generation (5G) mobile networks.

Basics of Clock and Data Recovery Circuits: Exploring High-Speed Serial Links

The choice of clock and data recovery (CDR) architecture in serial links dictates many of the blocklevel circuit specifications (specs). Block-level specs ultimately determine the energy efficiency of the system. Therefore, to design energy-efficient serial links, it is important to understand the basics of CDR operation, CDR's main performance metrics, and the relationship between circuit-level parameters and system-level performance metrics.

A Tactic for Architectural Exploitation of Indoor Small Cells for Dynamic Spectrum Sharing in 5G

This paper presents a tactic to realize numerous in-building small cell base station (SBS) architectures for dynamic spectrum sharing by varying the number of physical transceivers as well as the number, amount, and characteristics of operating spectrums per SBS. Each SBS architecture is defined as a Type and the spectrum sharing mechanism of each Type of architecture is detailed and mathematically analyzed to define existing dynamic spectrum sharing techniques suggested in the literature. The high external wall penetration loss of a building and the eICIC technique are exploited to manage co-channel interference generated due to sharing the spectrum of one system to another. System-level capacity, spectral efficiency, and energy efficiency performance metrics are derived for each SBS architecture to show the relative outperformance of one to another. It is found that, unlike energy efficiency (EE), the spectral efficiency (SE) response of an SBS architecture is directly affected by the channel characteristics as well as the number and amount of operating spectrum bands. However, the number of transceivers of an SBS does not have a noticeable impact on both SE and EE so long as the feature of an operating spectrum is not altered. Finally, we show that each Type of the proposed SBS architectures can achieve the prospective SE and EE requirements for fifth-generation (5G) mobile networks.

Countrywide Mobile Spectrum Sharing with Small Indoor Cells for Massive Spectral and Energy Efficiencies in 5G and Beyond Mobile Networks

In this paper, we propose a technique to share the licensed spectrums of all mobile network operators (MNOs) of a country with in-building small cells per MNO by exploiting the external wall penetration loss of a building and introducing the time-domain eICIC technique. The proposed technique considers allocating the dedicated spectrum Bop per MNO only its to outdoor macro UEs, whereas the total spectrum of all MNOs of the country Bco to its small cells indoor per building such that technically any small indoor cell of an MNO can have access to Bco instead of merely Bop assigned only to the MNO itself. We develop an interference management strategy as well as an algorithm for the proposed technique. System-level capacity, spectral efficiency, and energy efficiency performance metrics are derived, and a generic model for energy efficiency is presented. An optimal amount of small indoor cell density in terms of the number of buildings L carrying these small cells per MNO to trade-off the spectral efficiency and the energy efficiency is derived. With the system-level numerical and simulation results, we define an optimal value of L for a dense deployment of small indoor cells of an MNO and show that the proposed spectrum sharing technique can achieve massive indoor capacity, spectral efficiency, and energy efficiency for the MNO. Finally, we demonstrate that the proposed spectrum sharing technique could meet both the spectral efficiency and the energy efficiency requirements for 5G mobile networks for numerous traffic arrival rates to small indoor cells per building of an MNO.

Dynamic availability and performance deficiency of common bus systems with imperfectly repairable components

Common bus performance sharing (CBPS) abounds in diverse applications such as transportation, power supply and collaborative computing to facilitate efficient utilization of limited system resources. This paper models such a CBPS system with repairable components performing the main system function and repairable bus lines for redistributing surplus performance of some system components to components undergoing performance deficiency. Both time-to-failure and time-to-repair of system units (including components and lines) are random and may follow any arbitrary types of distributions. A probabilistic model based on discrete-state continuous-time stochastic processes is proposed for evaluating instantaneous availability of these repairable system units. The proposed model addresses the general repair policy encompassing minimal repair, perfect repair and imperfect repair. The universal generating function technique is further implemented for evaluating the system level performance metrics, including instantaneous system availability, instantaneous expected performance deficiency, expected system availability and total expected unsupplied demand during a specified mission time. Examples are provided to illustrate the proposed evaluation methodology and its application in prioritizing maintenance improvement actions.

Genetic Programming for Energy-Efficient and Energy-Scalable Approximate Feature Computation in Embedded Inference Systems

With the increasing interest in deploying embedded sensors in a range of applications, there is also interest in deploying embedded inference capabilities. Doing so under the strict and often variable energy constraints of the embedded platforms requires algorithmic, in addition to circuit and architectural, approaches to reducing energy. A broad approach that has recently received considerable attention in the context of inference systems is approximate computing. This stems from the observation that many inference systems exhibit various forms of tolerance to data noise. While some systems have demonstrated significant approximation-versus-energy knobs to exploit this, they have been applicable to specific kernels and architectures; the more generally available knobs have been relatively weak, resulting in large data noise for relatively modest energy savings (e.g., voltage overscaling, bit-precision scaling). In this work, we explore the use of genetic programming (GP) to compute approximate features. Further, we leverage a method that enhances tolerance to feature-data noise through directed retraining of the inference stage. Previous work in GP has shown that it generalizes well to enable approximation of a broad range of computations, raising the potential for broad applicability of the proposed approach. The focus on feature extraction is deliberate because they involve diverse, often highly nonlinear, operations, challenging general applicability of energy-reducing approaches. We evaluate the proposed methodologies through two case studies, based on energy modeling of a custom low-power microprocessor with a classification accelerator. The first case study is on electroencephalogram-based seizure detection. We find that the choice of two primitive functions (square root, subtraction) out of seven possible primitive functions (addition, subtraction, multiplication, logarithm, exponential, square root, and square) enables us to approximate features in 0.41 <inline-formula><tex-math notation="LaTeX">$mJ$</tex-math></inline-formula> per feature vector (FV), as compared to 4.79 <inline-formula><tex-math notation="LaTeX">$mJ$</tex-math></inline-formula> per FV required for baseline feature extraction. This represents a feature extraction energy reduction of 11.68 <inline-formula><tex-math notation="LaTeX"> $\times$</tex-math></inline-formula> . The important system-level performance metrics for seizure detection are sensitivity, latency, and number of false alarms per hour. Our set of GP models achieves 100 percent sensitivity, 4.37 second latency, and 0.15 false alarms per hour. The baseline performance is 100 percent sensitivity, 3.84 second latency, and 0.06 false alarms per hour. The second case study is on electrocardiogram-based arrhythmia detection. In this case, just one primitive function (multiplication) suffices to approximate features in 1.13 <inline-formula><tex-math notation="LaTeX">$\mu J$</tex-math> </inline-formula> per FV, as compared to 11.69 <inline-formula><tex-math notation="LaTeX">$\mu J$</tex-math></inline-formula> per FV required for baseline feature extraction. This represents a feature extraction energy reduction of 10.35 <inline-formula><tex-math notation="LaTeX"> $\times$</tex-math></inline-formula> . The important system-level metrics in this case are sensitivity, specificity, and accuracy. Our set of GP models achieves 81.17 percent sensitivity, 80.63 percent specificity, and 81.86 percent accuracy, whereas the baseline achieves 82.05 percent sensitivity, 88.12 percent specificity, and 87.92 percent accuracy. These case studies demonstrate the possibility of a significant reduction in feature extraction energy at the expense of a slight degradation in system performance.

Packaging and Miniaturization of a 2–18 GHz UWB Radar for Measurements of Snow and Ice: Initial Results

Abstract We report recent progress towards miniaturizing and packaging an ultra wideband (UWB) radar developed for measurements of ice and snow over the 2–18 GHz range. The initial objective of this work is the close integration of the radar's RF section (composed of a receiver and a low-power transmitter) while maintaining or enhancing its electrical performance. First, we describe a connectorized reference system developed to serve as a performance benchmark for transmit and receive signal chains. Next, we discuss the use of drop-in modules to achieve a significant size reduction of the RF front-end. These modules are based on packaged chips and arranged in a planar configuration. We present system-level performance metrics of the reference system justifying the need for further integration. We briefly discuss other potential applications of the miniaturized radar and the direction of future packaging efforts as we work toward our future goal of integrating the complete radar “from antenna to bits” by using a system-in-package approach.

Cloud-Based Scalable Software for Optimal Long-Range, Network-Level Bridge Improvement Programming

The current state of the practice in bridge management highlights a growing need to develop scalable optimization software tools to support the development of truly optimal bridge improvement programs and ensure that limited financial resources are optimally allocated. The heuristic project selection approaches employed in today’s bridge management systems are not capable of generating optimal programs. Agencies that rely on suboptimal programs may inadvertently direct a significant portion of their budget to the wrong projects, leading to an increase in maintenance backlogs and overall system risk levels. The subjective project selection criteria may also hinder the ability to quantify project benefits or justify projects to funding agencies and stakeholders. This paper presents a novel dynamic programming–based multiobjective optimization approach that is capable of generating global optimal network-level, long-range bridge improvement programs. The algorithm considers three objectives: the minimization of system-level risk, the maximization of system-level condition, and the minimization of life-cycle costs, subject to agency-defined constraints and planning scenarios. The algorithm efficiently explores the enormous search space to find optimal project lists for each year in the planning horizon under any given scenario. Alternative planning scenarios are defined to quantify the impact of different investment levels on system-level performance metrics and to determine the investment required to achieve the desired performance and risk targets.

Performance Indicators for Structural Systems and Infrastructure Networks

AbstractEstablishing consistent criteria for assessing the performance of structural systems and infrastructure networks is a critical component of communities’ efforts to optimize investment decisions for the upkeep and renewal of the built environment. Although member-level performance and reliability assessment procedures are currently well-established, it is widely recognized that a member-oriented approach does not necessarily lead to an efficient utilization of limited resources when making decisions related to the management of existing deteriorating structures or lifeline systems, especially those that may be exposed to extreme events. For this reason, researchers have renewed their interests in developing system-level assessment methods as a basis to modern structural and infrastructure performance evaluation and design processes. Specifically, system-level performance metrics and characteristics such as reliability, redundancy, robustness, resilience, and risk continue to be refined. The objecti...

Transceiver Design for High-Density Links With Discontinuities Using Modal Signaling

In high-speed chip-to-chip single-ended signaling links, far-end crosstalk presents one of the dominant noise sources, limiting the performance of the link. Diagonalizing the channel using modal decomposition has been proposed to mitigate the crosstalk, but so far only the transceiver designed for uniform low-loss homogenous media channels has been investigated. Furthermore, a practical circuit implementation of such a transceiver system will unavoidably deviate from the ideal (crosstalk-free) eigenvalue decomposition. In this paper, the impact of modal encoder/decoder coefficient quantization, random common noise, and uncorrelated noise on signal integrity is analyzed in terms of system-level performance metrics, signal-to-noise ratio at the receiver slicer, and bit error rate. Also discussed is the optimal selection of decoder coefficients with respect to the input-referred noise at the receiver. Based on these guidelines, a circuit design of a modal transceiver system over a typical memory bus with discontinuities is presented in detail. The proposed approach is verified using circuit-based link simulation, and the performance compared to several other proposed crosstalk-cancellation techniques.

Application-Based Congestion Control Policy for the Communication Channel in VANETs

This letter proposes a novel proactive congestion control policy for vehicular ad-hoc networks, in which each vehicle's communication parameters are adapted based on their individual application requirements. Contrary to other approaches, where transmission resources tend to be assigned based on system-level performance metrics, the technique proposed in this paper aims to individually satisfy the target application performance of each vehicle, while globally minimising the channel load to prevent channel congestion.