Service Level Agreement (SLA) is one of the most important characteristics when discussing about Cloud scheduling. Modeling Cloud scheduling under SLA constraints represents one of the main research areas in the field...
详细信息
ISBN:
(纸本)9781538618394
Service Level Agreement (SLA) is one of the most important characteristics when discussing about Cloud scheduling. Modeling Cloud scheduling under SLA constraints represents one of the main research areas in the field of Cloud computing. One of the main ways to simulate the scheduling in the Cloud is by using a simulator that is verified by the academia and by the industry as one of the most accurate and capable of reproducing real scheduling situation. In this paper, we model scheduling in the Cloud under SLA constraints. Our model also attempts to achieve profit optimization. The model of the scheduling under SLA constraints and profit optimization is implemented and verified by using the features of the CloudSim simulator, a well-known Cloud simulator.
Assessing the performance of scheduling heuristics through simulation requires to generate synthetic instances of tasks and machines with well-identified properties. Carefully controlling these properties is mandatory...
详细信息
ISBN:
(纸本)9783662480960;9783662480953
Assessing the performance of scheduling heuristics through simulation requires to generate synthetic instances of tasks and machines with well-identified properties. Carefully controlling these properties is mandatory to avoid any bias. We consider the scheduling problem consisting of allocating independent sequential tasks on unrelated processors while minimizing the maximum execution time. In this problem, the instance is a cost matrix that specifies the execution cost of any task on any machine. This paper proposes a measure for quantifying the heterogeneity properties of a cost matrix. An analysis of two classical methods used in the literature reveals a bias in previous studies. A new method is proposed to generate instances with given heterogeneity properties and it is shown that they have a significant impact on several heuristics.
High-performance input queued switches achieve good performance with low cost. However, with the appearance of optical techniques, the line rate is much higher than before. scheduling algorithms require not only good ...
详细信息
ISBN:
(纸本)076951961X
High-performance input queued switches achieve good performance with low cost. However, with the appearance of optical techniques, the line rate is much higher than before. scheduling algorithms require not only good performance in delay and stability but fast speed and simple implementation as well. A variety of scheduling algorithms for Virtual Output Queuing (VOQ) packet switch architecture are proposed Round-robin scheduling algorithms are fast and simple to implement in hardware. In particular, a group of fully desynchronized round-robin scheduling algorithms- SRR (static round robin matching), proposed recently, achieve pretty good delay performance while easy to implement. However, they are not stable under non-uniform traffic. Randomized algorithms are stable under any admissible traffic, however their delay is high and hardware implementation is complex. Based on the concept of randomized algorithms and SRR, we propose a group of new scheduling algorithms, DRDSRR, the improved version of DRDSRR, ARDSRR and the variations of ARDSRR. They not only ensure stability but also have good performance and simple implementation. We have proved stability in the paper.
The possibility of providing mobile devices with fast and affordable Cloud computing (CC) services is a major motivation for Cloudlet technology. However, for this happen, a suitable scheduling algorithm that minimize...
详细信息
ISBN:
(纸本)9781728132891
The possibility of providing mobile devices with fast and affordable Cloud computing (CC) services is a major motivation for Cloudlet technology. However, for this happen, a suitable scheduling algorithm that minimizes the consumer's (i.e. Mobile devices) waiting time under various Cloudlet scenarios needs to be determined. Hence, determining the suitability of a scheduling algorithm with respect to consumer's waiting time is the main objective of this study. To this end, we first present a brief review of existing scheduling algorithms as a road map towards evaluating the performance of each algorithm. Next, we present the performance evaluation of three popular scheduling algorithms under various Cloudlet scenarios. Specifically, an EdgeCloudSim simulator is used to evaluate the performance of MM1, Priority (PR) and Round Robin (RR) with respect to consumer's waiting time. Results gotten from the simulation show a noticeable difference in the performance of scheduling algorithms for different mobile applications as the number of mobile devices fluctuates.
A downlink wireless system is considered where a single base station serves two users simultaneously based on the information available at the PRY and NUC layers. Two new quality of service (QoS) oriented scheduling a...
详细信息
ISBN:
(纸本)9781424404681
A downlink wireless system is considered where a single base station serves two users simultaneously based on the information available at the PRY and NUC layers. Two new quality of service (QoS) oriented scheduling algorithms are proposed that consider both channel state and direction of arrival information. available at the PHY layer and queue length information available at the NUC layer to schedule users. The performance of the proposed algorithms is compared with existing scheduling algorithms, that consider either the channel state information, delay information or the combination of the two, in terms of SINR and exceedance probability of delay and queue length. The results obtained indicate that the proposed algorithms perform better than existing algorithms.
This paper evaluates how distinct real-time task scheduling algorithms impact power consumption and timing performance of embedded systems. A design space exploration methodology is proposed in order to adjust the sys...
详细信息
ISBN:
(纸本)1581139470
This paper evaluates how distinct real-time task scheduling algorithms impact power consumption and timing performance of embedded systems. A design space exploration methodology is proposed in order to adjust the system's power consumption by tuning the CPU frequency according to the scheduling algorithm and to the temporal requirements of the embedded application. The goal is to find an optimized configuration, selecting the right combination of a scheduling policy with a CPU frequency, so as to consume less power without missing any deadline in the application. Experiments based on a synthetic workload that simulates realistic applications demonstrate that considerable power savings can be obtained. Moreover, the paper defines guidelines to be used by system designers in order to find a configuration that best matches the design constraints and requirements.
This paper studies the classical problem of online scheduling of deadline-sensitive jobs with partial values and investigates its extension to Electric Vehicle (EV) charging scheduling by taking into account the proce...
详细信息
ISBN:
(纸本)9781538625422
This paper studies the classical problem of online scheduling of deadline-sensitive jobs with partial values and investigates its extension to Electric Vehicle (EV) charging scheduling by taking into account the processing rate limit of jobs and charging station capacity constraint. The problem lies in the category of time-coupled online scheduling problems without availability of future information. This paper proposes two online algorithms, both of which are shown to be (2-1/U)-competitive, where U is the maximum scarcity level, a parameter that indicates demand-to-supply ratio. The first proposed algorithm is deterministic, whereas the second is randomized and enjoys a lower computational complexity. When U grows large, the performance of both algorithms approaches that of the stateof- the-art for the case where there is processing rate limits on the jobs. Nonetheless in realistic cases, where U is typically small, the proposed algorithms enjoy a much lower competitive ratio. To carry out the competitive analysis of our algorithms, we present a proof technique, which is novel to the best of our knowledge. This technique could also be used to simplify the competitive analysis of some existing algorithms, and thus could be of independent interest.
The input-queued switch architecture is widely used in Internet routers due to its ability to run at very high line speeds. A central problem in designing an input-queued switch is the scheduling algorithm that decide...
详细信息
ISBN:
(纸本)9781424410460
The input-queued switch architecture is widely used in Internet routers due to its ability to run at very high line speeds. A central problem in designing an input-queued switch is the scheduling algorithm that decides which packets to transfer from ingress ports to egress ports in a given timeslot. It is desirable that such algorithms be iterative (so as to be pipelineable), distributed (allowing flexibility in hardware implementation) and are able to deliver high performance (in terms of throughput and delay). In practice, implementable algorithms have so far had limited success in combining all of the above properties. For example, the popular iSLIP [1] algorithm is known to perform suboptimally, but it is commercially deployed mainly because it is iterative and distributed. The main contribution of this paper is the design and systematic analysis of two algorithms which, to the best of our knowledge, are the first high-performance iterative and distributed scheduling algorithms with possibility of efficient implementation. We first present an iterative, distributed and low-delay maximal throughput algorithm based on the celebrated "Auction Algorithm" [2], [3]. This algorithm can be seen as a natural extension of iSLIP when queue-size information is allowed to be exchanged. The standard auction algorithm can take an unbounded number of iterations to converge in the worst case. However we show that under admissible Bernoulli i.i.d. traffic, our algorithm takes O(n(2)) iterations, where n is the number of ingress/egress ports in the switch. Moreover for a switch with finite buffer-size, the algorithm allows for a graceful trade-off between running time and performance, which we verify by representative simulation results. Next, we propose and analyze a throughput-optimal, iterative and distributed scheduling algorithm influenced by Max-Product Belief Propagation [4], [5]. Recently the problem of efficient transmission over multi-hop wireless networks has been formu
LTE is a promising generation of mobile communication, characterized by high data rate, lower latency in control and user plane. In the study simulation model was developed for LTE system throughput estimation for dif...
详细信息
ISBN:
(纸本)9781467391115
LTE is a promising generation of mobile communication, characterized by high data rate, lower latency in control and user plane. In the study simulation model was developed for LTE system throughput estimation for different scheduling algorithms. Conclusion of the algorithms advantages and weaknesses were made.
This paper presents the evaluation results of a simulation-based study on the fairness of packet scheduling algorithms to support QoS in IP networks. The simulation study aimed at validating the performance of our pro...
详细信息
ISBN:
(纸本)0889863911
This paper presents the evaluation results of a simulation-based study on the fairness of packet scheduling algorithms to support QoS in IP networks. The simulation study aimed at validating the performance of our proposed scheme for QoS provisioning for real-time traffic in the context of Differentiated Services. We assume that traffic classification based on the application QoS requirements is an indispensable feature for providing QoS guarantees to a wide variety of Internet multimedia applications such as voice over IP (VoIP) and video conferencing due to the scalability issues. We evaluate the performance of our proposed scheme using extensive network simulation in comparison to other approaches. The results from the simulation studies indicate that the packet scheduling mechanism we propose ensures both the required bandwidth fairness and end-to-end network delay bounds for QoS-sensitive applications according to the specified weight ratios under various traffic and network conditions.
暂无评论