Performance Evaluation最新文献

We consider a Multi-access Edge Computing (MEC) system with a set of users, a base station (BS) with an attached MEC server, and a cloud server. The users can serve the service requests locally or can offload them to the BS which in turn can serve a subset of the offloaded requests at the MEC and can forward the requests to the cloud server. The user devices and the MEC server can be dynamically configured to serve different classes of services. The service requests offloaded to the BS incur offloading costs and those forwarded to the cloud incur additional costs; the costs could represent service charges or delays. Aggregate cost minimization subject to stability warrants optimal service scheduling and offloading at the users and the MEC server, at their application layers, and optimal uplink packet scheduling at the users’ MAC layers. Classical back-pressure (BP) based solutions entail cross-layer message exchange, and hence are not viable. We propose virtual queue-based drift-plus-penalty algorithms that are throughput optimal, and achieve the optimal delay arbitrarily closely but do not require cross-layer communication. We first consider an MEC system without local computation, and subsequently, extend our framework to incorporate local computation also. We demonstrate that the proposed algorithms offer almost the same performance as BP based algorithms. These algorithms contain tuneable parameters that offer a trade off between queue lengths at the users and the BS and the offloading costs.

FlexRay is a high-bandwidth protocol that supports hard-deadline periodic and sporadic traffic in modern in-vehicle communication networks. The dynamic segment of FlexRay is used for transmitting hard deadline sporadic messages. In this paper, we describe an algorithm to minimize the duration of the dynamic segment in a FlexRay cycle, yielding better results than existing algorithms in the literature. The proposed algorithm consists of two phases. In the first phase, we assume that a sporadic message instance contends for service with only one instance of each higher-priority message. The lower bound provided by the first phase serves as the initial guess for the number of mini-slots used in the second phase, where an exact scheduling analysis is performed. In the second phase, a sporadic message may contend for service with multiple instances of each higher-priority message. This two-phase approach is efficient because the first phase has low overhead and its estimate greatly reduces the number of iterations needed in the second phase. We conducted experiments using the dataset provided in the literature as well as the SAE benchmark dataset. The experimental results demonstrate superior bandwidth minimization and computational efficiency compared to other algorithms.

The interleaved regulator (implemented by IEEE TSN Asynchronous Traffic Shaping) is used in time-sensitive networks for reshaping the flows with per-flow contracts. When applied to an aggregate of flows that come from a FIFO system, an interleaved regulator that reshapes the flows with their initial contracts does not increase the worst-case delay of the aggregate. This shaping-for-free property supports the computation of end-to-end latency bounds and the validation of the network’s timing requirements. A common method to establish the properties of a network element is to obtain a network-calculus service-curve model. The existence of such a model for the interleaved regulator remains an open question. If a service-curve model were found for the interleaved regulator, then the analysis of this mechanism would no longer be limited to the situations where the shaping-for-free holds, which would widen its use in time-sensitive networks. In this paper, we investigate if network-calculus service curves can capture the behavior of the interleaved regulator. For an interleaved regulator that is placed outside of the shaping-for-free requirements (after a non-FIFO system), we develop Spring, an adversarial traffic generation that yields unbounded latencies. Consequently, we prove that no network-calculus service curve exists to explain the interleaved regulator’s behavior. It is still possible to find non-trivial service curves for the interleaved regulator. However, their long-term rate cannot be large enough to provide any guarantee. Specifically, we prove that for the regulators that process at least four flows with the same contract, the long-term rate of any service curve is upper bounded by three times the rate of the per-flow contract.

Packet processing in current network scenarios faces complex challenges due to the increasing prevalence of requirements such as low latency, high reliability, and resource sharing. Virtualization is a potential solution to mitigate these challenges by enabling resource sharing and on-demand provisioning; however, ensuring high reliability and ultra-low latency remains a key challenge. Since bare-metal systems are often impractical because of high cost and space usage, and the overhead of virtual machines (VMs) is substantial, we evaluate the utilization of containers as a potential lightweight solution for low-latency packet processing. Herein, we discuss the benefits and drawbacks and encourage container environments in low-latency packet processing when the degree of isolation of customer data is adequate and bare metal systems are unaffordable. Our results demonstrate that containers exhibit similar latency performance with more predictable tail-latency behavior than bare metal packet processing. Moreover, deciding which mainboard architecture to use, especially the cache division, is equally vital as containers are prone to higher latencies on more shared caches between cores especially when other optimizations cannot be used. We show that this has a higher impact on latencies within containers than on bare metal or VMs, resulting in the selection of hardware architectures following optimizations as a critical challenge. Furthermore, the results reveal that the virtualization overhead does not impact tail latencies.

Massive deployment of IoT devices raises the need for energy-efficient spectrum-efficient low-cost communications. Ambient backscatter communication (AmBC) technology provides a promising solution to achieve that. Moreover, incorporating AmBC with cognitive radio networks (CRNs) achieves better spectrum efficiency; however, this comes with performance drawbacks. In this work, we investigate the security and reliability performance of an underlay CRN with AmBC, where the backscattering device (BD) exploits the radio frequency (RF) signals of the secondary transmitter (ST), and both the ST and the BD share a common receiver. Different from previous work, we consider an ST with multiple antenna. The ST employs a transmit antenna selection (TAS) scheme to enhance the ST performance and overcome the performance degradation caused by the BD interference. TAS exploits multiple antenna diversity with lower hardware complexity and power consumption. Considering the Nakagami-$m$ fading model, closed-form expressions are derived for the outage probability (OP) and intercept probability (IP) of both the ST and the BD transmissions at the legitimate receiver and the eavesdropper. Moreover, the asymptotic behavior of OPs and IPs is also investigated in the high signal-to-noise ratio regime and the high main-to-eavesdropper ratio regime, respectively. Monte Carlo simulations are performed to validate the derived closed-form expressions. Numerical results show that employing TAS enhances the ST and BD reliability performance by percentages up to 98% and 80%, respectively, at high primary user interference threshold values. Moreover, it results in a better security-reliability trade-off for the ST and the BD.

The emerging paradigm of batteryless intermittent sensor networks (BISNs) presents new challenges for researchers of low-power wireless systems and protocols. The nature of these challenges exacerbates the difficulty of evaluating networks of physical sensor nodes, making simulation an even more important component in evaluating performance metrics, such as communication throughput and delay, for BISN designs. To our knowledge, existing simulators and analytical models do not meet the unique needs of BISN research; therefore, we have created a new open-source BISN simulator named Lure. Lure is designed from the ground-up for simulation of batteryless intermittent systems and networks. Written in Python, Lure is powerful, flexible, highly configurable, and supports rapid prototyping of new protocols, systems, and applications, with a low learning curve. In this paper, we present Lure and validate it with experimental data to show that Lure can accurately reflect the reality of BISNs. We then demonstrate the process of applying Lure to research questions in select case studies.

Due to the proliferation of inference tasks on mobile devices, state-of-the-art neural architectures are typically designed using Neural Architecture Search (NAS) to achieve good tradeoffs between machine learning accuracy and inference latency. While measuring inference latency of a huge set of candidate architectures during NAS is not feasible, latency prediction for mobile devices is challenging, because of hardware heterogeneity, optimizations applied by machine learning frameworks, and diversity of neural architectures. Motivated by these challenges, we first quantitatively assess the characteristics of neural architectures (specifically, convolutional neural networks for image classification), ML frameworks, and mobile devices that have significant effects on inference latency. Based on this assessment, we propose an operation-wise framework which addresses these challenges by developing operation-wise latency predictors and achieves high accuracy in end-to-end latency predictions, as shown by our comprehensive evaluations on multiple mobile devices using multicore CPUs and GPUs. To illustrate that our approach does not require expensive data collection, we also show that accurate predictions can be achieved on real-world neural architectures using only small amounts of profiling data.