CQ switch performance analysis from the point of buffer size and scheduling algorithms

In this study, a crosspoint‐queued (CQ) switch that features a buffer size larger than one and employs the work‐conserving random scheduling algorithm under uniform Bernoulli i.i.d. traffic is investigated through both theoretical analysis and simulation. The effect of the crosspoint buffer state on switch performance is modelled and quantified. A novel method for calculating the throughput and delay of a CQ switch with a given switch size, input load, and buffer size is proposed. Compared with existing models, the proposed calculation model offers a simpler and more accurate result, and can serve as a theoretical guide in designing high‐speed switch fabrics.

Bluetooth is a short-range TDD (time division duplex) wireless network that supports both circuit- and packet-oriented applications. A piconet is composed of a device configured as master and at most seven other devices acting as slaves. At medium access control (MAC) layer, the master can select a slave to send a data packet and until then, the slave is not allowed to transmit. Round-robin (RR) and exhaustive round-robin (ERR) are two elementary MAC scheduling schemes that are both simple and efficient. This paper proposes RR-FCFS, a simple MAC scheduling scheme that has the same advantages as RR and ERR. RR-FCFS acts as RR if the master's queue is empty and starts transmitting packets in first-come-first-serve order otherwise. The simulation results show that RR-FCFS's performance in terms of packet delay and queue length is comparable with those of RR and ERR.

Quality of service provision in combined input and crosspoint queued switches without output queueing match

Performance of ATM switch fabric using cross-point buffers

Load balancing is a widely adopted strategy in modern distributed systems because it distributes workloads across servers, mitigating overload and improving overall performance. However, the rapid growth of such systems has created a need for more adaptive strategies to ensure optimal utilization and responsiveness of resources. Traditional algorithms such as Round Robin (RR) and Weighted Round Robin (WRR) assign requests without considering server states or request characteristics. We implement a machine learning (ML)–based predictive load balancer, forecasting the latency of a request based on the request itself and container parameters, specifically the average latency of the last 50 requests and the count of active requests, and evaluate it against RR and WRR. For the experiment, synthetic data were generated to replicate real-world requests by creating random URL and method combinations, attaching a task size in Million Instructions (MI), and distributing them among three containers with varying resources according to the load balancing strategies described above. Under the conditions tested, the ML approach achieved the worst performance, trailing both RR and WRR in terms of throughput and average latency, although the model accuracy was sufficiently high (R2 = 0.8+). Post hoc analysis indicates that limited and occasionally stale runtime features caused the load balancer to direct all requests to a single container until the next statistics update, since that container was considered the ‘best’ during that interval.

As applications scale to increasingly large processor counts, the interconnection network is frequently the limiting factor in application performance. In order to achieve application scalability, the interconnect must maintain high bandwidth while minimizing variation in packet latency. As the offered load in the network increases with growing problem sizes and processor counts, so does the expected maximum packet latency in the network, directly impacting performance of applications with any synchronized communication. Age-based packet arbitration reduces the variance in packet latency as well as average latency. This paper describes the Cray XT router packet aging algorithm which allows globally fair arbitration by incorporating age in the packet output arbitration. We describe the parameters of the aging algorithm and how to arrive at appropriate settings. We show that an efficient aging algorithm reduces both the average packet latency and the variance in packet latency on communication-intensive benchmarks.

A frame‐based packet scheduling scheme, the pinwheel scheduling (PWS) scheme, is proposed and investigated at the medium access control (MAC) layer in broadband wireless networks. The objective of the proposed scheduling scheme is to provide low delay and low jitter for real‐time traffic. We have demonstrated in this paper that the proposed PWS scheme not only satisfies the packet‐level QoS requirements but also has low implementation cost. The PWS scheme gives the highest priority to CBR connections in a service cycle to minimize their delay and jitter. For ABR traffic, a number of slots are allocated to fit their MCR. The remaining time slots are then allocated to VBR traffic according to their PCR. Thus, the VBR traffic may suffer large delay and jitter when the number of CBR connections increases. The PWS with modular strategy (PWS/MS) is proposed to improve the performance of the VBR traffic. We also introduce three different slot‐reuse strategies named real‐time traffic first (RTF), non‐real‐time traffic first (NRTF) and longest queue first (LQF), respectively, to improve resource utilization. In order to demonstrate the performance of the PWS and PWS/MS schemes, comparisons are made with existing schemes such as the round‐robin (RR) and weighted‐round‐robin (WRR) by using OPNET software. Simulation results show that the proposed schemes are capable of maintaining the lowest delay and jitter for VBR and CBR traffic while not sacrificing the available bit rate (ABR) traffic.

Combined input-crosspoint buffered (CICB) packet switches have been of research interest in the last few years because of their high performance. These switches provide higher performance than input-buffered (IB) packet switches while requiring the crosspoint buffers run at the same speed as that of the input buffers in IB switches. Recently, it has been shown that CICB switches with one-cell crosspoint buffers, virtual output queues, and simple input and output arbitrations, provide 100% throughput under uniform traffic. However, it is of general interest to know the maximum throughput that a CICB switch, with no speedup, can provide under admissible traffic. This paper analyzes the throughput performance of a CICB switch beyond uniform traffic patterns and shows that a CICB switch with one-cell crosspoint buffers can provide 100% throughput under admissible traffic while using no speedup.

Shuffleout is a blocking multistage asynchronous transfer mode (ATM) switch using shortest path routing with deflection, in which output queues are connected to all the stages. This paper describes a model for the performance evaluation of the shuffleout switch under arbitrary nonuniform traffic patterns. The analytical model that has been developed computes the load distribution on each interstage link by properly taking into account the switch inlet on which the packet has been received and the switch outlet the packet is addressing. Such a model allows the computation not only of the average load per stage but also its distribution over the different links belonging to the interstage pattern for each switch input/output pair. Different classes of nonuniform traffic patterns have been identified and for each of them the traffic performance of the switch is evaluated by thus emphasizing the evaluation of the network unfairness.

An efficient iterative scheduling algorithm for input-queued switches, called round robin with longest queue first (RR/LQF), is proposed in this paper. RR/LQF consists of three phases: report, grant, and accept. In each phase, only a single-bit message per port is sent for reporting a packet arrival, granting an input for packet sending, or accepting a grant. In both the grant and accept phases, scheduling priority is given to the preferred input–output pairs first and the longest virtual output queuing (VOQ) next. The notion of the preferred input–output pair is to keep a global RR schedule among all the inputs and the outputs. By serving the preferred input–output pairs first, the match size tends to be maximized. By serving the longest VOQ next, the match weight is also boosted. When RR/LQF is executed for a single iteration (i.e., RR/LQF-1), we show by simulations that RR/LQF-1 outperforms all the existing single-bit-single-iteration scheduling algorithms. When RR/LQF is executed up to $N$ iterations (i.e., RR/LQF- $N$ ), we prove that under any admissible traffic pattern, RR/LQF- $N$ is stable with a speedup of $2-1/N$ , where $N$ is the switch size. To the best of our knowledge, this is the first work showing that an iterative scheduling algorithm is stable with a speedup less than 2. We then generalize RR/LQF to become a class of algorithms that have the same speedup bound of $2-1/N$ . Efforts are then made to further reduce the implementation complexity of RR/LQF. To this end, the pipelined RR/LQF and RR/RR, a simpler variant of RR/LQF, are proposed.

The overhead associated with reconfiguring a switch fabric in optical packet switches is an important issue in relation to the packet transmission time and can adversely affect switch performance. The reconfiguration overhead increases the mean waiting time of packets and reduces throughput. The scheduling of packets must take into account the reconfiguration frequency. This work proposes an analytical model for input-buffered optical packet switches with the reconfiguration overhead and analytically finds the optimal reconfiguration frequency that minimizes the mean waiting time of packets. The analytical model is suitable for several round-robin (RR) scheduling schemes in which only non-empty virtual output queues (VOQs) are served or all VOQs are served and is used to examine the effects of the RR scheduling schemes and various network parameters on the mean waiting time of packets. Quantitative examples demonstrate that properly balancing the reconfiguration frequency can effectively reduce the mean waiting time of packets.

HPC systems have shifted to burst buffer storage and high radix interconnect topologies in order to meet the challenges of large-scale, data-intensive scientific computing. Both of these technologies have been studied in detail independently, but the interaction between them is not well understood. I/O traffic and communication traffic from concurrently scheduled applications may interfere with each other in unexpected ways, and this behavior may vary considerably depending on resource allocation, scheduling, and routing policies.In this work, we analyze I/O and network traffic interference on burst-buffer-equipped dragonfly-based systems using the high-resolution packet-level simulations provided by the CODES storage and interconnect simulation framework. The analysis is performed using realistic I/O workload sizes, a variety of resource allocation and network routing strategies employed in production environments, and a dragonfly network configuration modeled after current vendor options. We analyze the impact of interference on both I/O and communication traffic.We observe that although average network packet latency is stable across a wide variety of configurations, the maximum network packet latency in the presence of concurrent I/O traffic is highly sensitive to subtle policy changes. Our simulations reveal a worst-case single packet latency of 4,700 times the average latency for sub-optimal configurations. While a topology-aware mapping of compute nodes to burst buffer storage nodes can minimize the variation in maximum packet latency, it can slow down the I/O traffic by creating contention on the burst buffer nodes. Overall, balancing I/O and network performance requires careful selection of routing, data placement, and job placement policies.

Event Abstract Back to Event A spiking point of view - Is it possible to predict a neurons morphology from its electrophysiological activity? Anneke Meyer1, 2*, C. G. Galizia2 and Martin P. Nawrot1 1 Freie Universität Berlin, Germany 2 Department for Neurobiology/ Universität Konstanz, Germany Here we investigate neurons of two morphological classes, which reside in the Antennal Lobe (Al) of the honey bee: local neurons (LNs) and projection neurons (PNs). We ask, whether it is possible to conclude the neurons’ morphological class from the neurons’ electro-physiological properties. If this is the case, which combination of features of electrophysiological activity will most effectively describe the difference between LNs and PNs? We analyzed data from 80 intracellularly recorded AL neurons, 33 of which were unambiguously identified as either LN or PN. For each and every neuron we estimated a number of electro-physiological measures. We then explored clustering of neurons based on the principle components of every possible combination of these measures. Using hierarchical clustering with wards-linkage (REF) we determined which combination of measures performed best in separating identified PNs and LNs. Our analyses show that it is indeed possible to separate [class, with high certainty] LNs and PNs on the grounds of electro-physiological features only. We achieved the by far best clustering results (81% separation) based on only two PCs from a combination of six selected features, namely the CV2 (a measure of spiking irregularity), the Fano-Factor (a measure of across-trial spike count variability), the rate change (from spontaneous to stimulus response), the average response latency, the latency variation (across different stimuli), and the baseline power of the membrane potential. Many other feature combinations perform less well but still satisfyingly (70%-76% classification success). Amongst these good combinations, measures of spiking activity were more frequently represented than measures of sub-threshold activity. Neurons in the LN and PN cluster differed significantly for those features which we identified as most potent. Hence, the separation which we achieved in the principle component space can be described statistically on the level of the original features. We conclude that LNs and PNs in the honey bee AL are characterized by different electrophysiological properties, which are represented by measurable features. Still, none of these measures alone is a potent tool to predict a neuron’s morphology. The combination of several classical measures however offers a foundation to tell apart LNs from PNs with a certain probability in cases where morphological data is not at hand. Funding was received from the German Federal Ministry of Education and Research within the Bernstein Focus Neuronal Basis of Learning - Insect Inspired Robots (grant No. 01GQ0941) and from the Landesgraduiertenförderung Baden-Württemberg. Acknowledgements We thank Sabine Krofcik, Bernd Kimmerle, Randolf Menzel, Michael Schmuker, Titziano D'Albis and Jan Sölter Keywords: antennal lobe, clustering, data analysis and machne learning, Electrophysiology, Honeybee, Olfaction Conference: BC11 : Computational Neuroscience & Neurotechnology Bernstein Conference & Neurex Annual Meeting 2011, Freiburg, Germany, 4 Oct - 6 Oct, 2011. Presentation Type: Poster Topic: data analysis and machine learning (please use "data analysis and machine learning" as keyword) Citation: Meyer A, Galizia CG and Nawrot MP (2011). A spiking point of view - Is it possible to predict a neurons morphology from its electrophysiological activity?. Front. Comput. Neurosci. Conference Abstract: BC11 : Computational Neuroscience & Neurotechnology Bernstein Conference & Neurex Annual Meeting 2011. doi: 10.3389/conf.fncom.2011.53.00151 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 22 Aug 2011; Published Online: 04 Oct 2011. * Correspondence: Dr. Anneke Meyer, Freie Universität Berlin, Berlin, Germany, anneke.meyer@uni-konstanz.de Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Anneke Meyer C. G Galizia Martin P Nawrot Google Anneke Meyer C. G Galizia Martin P Nawrot Google Scholar Anneke Meyer C. G Galizia Martin P Nawrot PubMed Anneke Meyer C. G Galizia Martin P Nawrot Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

One of the issues that service consumers face when using cloud-based solutions is the delay in processing large and real-time data. Fog computing has been introduced as a solution to this issue since it allows the data to be processed at the edge of a network while enabling different tasks to be scheduled for processing in fog devices at the edge network. These devices still require cloud resources to give them the capacity of processing real-time applications. However, scheduling these tasks must be made in such a way that they do not consume all the available resources on a fog device. When a task consumes all available resources, it can lead to network breakdown or high latency which is not acceptable for real-time applications. Therefore, to address this problem, this chapter proposes a task scheduling technique for traffic surveillance vehicular network application through smart cameras. We used the iFogSim simulator, where the scenario of vehicle tracking is considered. Simulations are conducted to find an efficient scheduling algorithm among a pool of available ones that can optimize the energy consumption and average delay of our proposed real-time application model. The results show that the First Come First Serve (FCFS) scheduling algorithm outperforms the Short Job First (SJF), Generalized Priority (GP) and Round Robin (RR) counterparts in terms of average latency, energy usage, execution time and network usage.KeywordsTraffic schedulingAlgorithmsFog computingInternet of things (IoT)iFogSim

An output queueing analysis of multipath ATM switches