Hybrid storage systems can fully leverage the advantages of multiple devices to achieve better performance. However, current systems are designed primarily for a slow disk with an expensive fast device at high costs. ...
详细信息
ISBN:
(纸本)9798350342918
Hybrid storage systems can fully leverage the advantages of multiple devices to achieve better performance. However, current systems are designed primarily for a slow disk with an expensive fast device at high costs. They ignore device features and workload status while placing data. The issue of cache pollution is affecting their data hotness identification. Besides, inconsistent load status in multiple devices is overlooked during migration. These shortcomings constrain the overall performance of the system. To solve this, we propose HyF2FS, a hybrid storage filesystem based on F2FS. HyF2FS features a cache-tiering integrated architecture that stores hot data and metadata in an accelerator while asynchronously migrating cold data to the SSD, which provides cost-effective opportunities to optimize device parallelism. HyF2FS uses multidimensional scores to place data on the appropriate device to achieve high bandwidth. To improve data hotness identification, HyF2FS proposes two-level counters. Besides, a migration window is employed to minimize the impact of migration on foreground I/O. By implementing these scheduling algorithms, HyF2FS can fully exploit the parallelism of both fast and slow devices. Experimental results demonstrate significant improvements in throughput (116%-244%) and latency reduction (49%-64%) compared to F2FS and other hybrid storage systems.
OpenMP is widely used for shared memory parallel programming and is especially useful for the parallelisation of loops. When it comes to task parallelism, however, OpenMP is less powerful and the sections construct la...
详细信息
ISBN:
(数字)9783540693031
ISBN:
(纸本)9783540693024
OpenMP is widely used for shared memory parallel programming and is especially useful for the parallelisation of loops. When it comes to task parallelism, however, OpenMP is less powerful and the sections construct lacks support for dependences and fine grained tasks. This paper proposes a new work-sharing construct, tasks, which is a generalisation of sections. It goes beyond sections by allowing unbalanced and finer grained tasks with arbitrary dependence structure. A proof-of-concept compiler has been implemented for the new directives, which includes a state-of-the-art scheduling algorithm for task graphs. Experiments with a large set of programs were conducted using the new directives. The results demonstrate that the new approach can efficiently exploit the task parallelism inherent in the code, without introducing any additional overhead.
Real-time scheduling algorithms for multiprocessor systems have been the subject of considerable recent interest. For such an algorithm to be truly useful in practice, support for semaphore-based locking must be provi...
详细信息
ISBN:
(纸本)9780769529752
Real-time scheduling algorithms for multiprocessor systems have been the subject of considerable recent interest. For such an algorithm to be truly useful in practice, support for semaphore-based locking must be provided. However for man), global scheduling algorithms, no such mechanisms have been proposed. Furthermore, in the partitioned case, most prior semaphore schemes are either inefficient or restrict critical sections considerably In this paper a new flexible multiprocessor locking scheme is presented that can be applied under both partitioning and global scheduling. This scheme allows unrestricted critical-section nesting, but has been designed to deal with the common case of short non-nested accesses efficiently.
To propose a reliable cloud computing task deployment algorithm for the optimization theory. The current research on cloud computing task deployment mainly only focuses on one of the two goals: reliability and optimiz...
详细信息
We investigated how operating system design should be adapted for multithreaded chip multiprocessors (CMT) - a new generation of processors that exploit thread-level parallelism to mask the memory latency in modem wor...
详细信息
ISBN:
(纸本)1931971277
We investigated how operating system design should be adapted for multithreaded chip multiprocessors (CMT) - a new generation of processors that exploit thread-level parallelism to mask the memory latency in modem workloads. We determined that the L2 cache is a critical shared resource on CMT and that an insufficient amount of L2 cache can undermine the ability to hide memory latency on these processors. To use the L2 cache as efficiently as possible, we propose an L2-conscious scheduling algorithm and quantify its performance potential. Using this algorithm it is possible to reduce miss ratios in the L2 cache by 25-37% and improve processor throughput by 27-45%.
Fog computing extends the traditional centralized cloud-computing model to the edge of the network to provide cloud services such as computation and networking distributed closer to the end device. For Smart Farming a...
详细信息
ISBN:
(纸本)9781728132891
Fog computing extends the traditional centralized cloud-computing model to the edge of the network to provide cloud services such as computation and networking distributed closer to the end device. For Smart Farming applications, Fog computing can enable real-time analysis of crop and environmental behavior to improve the production of the agricultural sector. However, being at the edge results in limited physical underlying infrastructure thus lessens fog nodes (Fog node) resources compare to the cloud resulting in resource overconsumption problem and network performance downgrade if not managed. Therefore, to address resource enhancement problem. In this paper, we propose a K-Means based clustering algorithm for a Smart Farming application, compared network performance with first come first serve (FCFS) algorithms. Simulation results demonstrate that K-Means outperforms FCFS by 1.80 % on energy consumption, 1.2 % lesser on network usage, and 27 % lesser on end-to-end delay.
Task scheduling and resource scheduling are the core issue in cloud computing. Pointing at the premature problem in the scheduling algorithm of particle swarm, we propose a scheduling algorithm of cloud task particle ...
详细信息
Global scheduling algorithms are very promising for application in embedded real-time systems using multi-core controllers. In this paper we want to make a first step forward to apply such scheduling methods on real e...
详细信息
ISBN:
(纸本)9788026101666
Global scheduling algorithms are very promising for application in embedded real-time systems using multi-core controllers. In this paper we want to make a first step forward to apply such scheduling methods on real existing systems. Especially a new resource model is necessary to avoid deadlocks, as this goal can not be achieved by using the standard OSEK Priority Ceiling Protocol when shared global resources are in use. We also introduce the new metric mean Normalized Blocking Time in order to be able to compare locking mechanisms according to the timing effects of their blocking behavior. Finally we give a simulative application example of the new metric by the use of two different kinds of semaphore models and an example task set typical for existing embedded real-time systems in the automotive powertrain environment.
With mobile edge computing (MEC), the data compression at the edge devices can effectively improve the communication efficiency by transmitting the compressed data. In this paper, we construct a joint data compression...
详细信息
ISBN:
(纸本)9781538680889
With mobile edge computing (MEC), the data compression at the edge devices can effectively improve the communication efficiency by transmitting the compressed data. In this paper, we construct a joint data compression and transmission scheduling framework to optimize the system throughput with the limited transmission resources. Different to most of the existing works, we consider the interaction between the data compression and data transmission to achieve the optimal throughput. Specifically, to explore the effect of data compression, we construct a queue system through constructing the mapping between the original data queues and the compressed data queues under different compression schemes (including the uncompressed queues). We design the transmission scheduling algorithm based on Lyapunov optimization according to the original data queues. Due to the nature that the data compression does not change the original data queue length directly, we choose the optimal data compression scheme considering the achieved utilities when the compressed data are transmitted, which can be estimated via Q-learning. In addition, we theoretically prove the queue stability under our proposed joint data compression and transmission scheduling algorithm. The simulation results show that the proposed algorithm has better delay performance than the conventional schemes.
The paper proposes air architecture for a scheduling algorithm, to be integrated in IEEE 802.11 Access Points (AP), able to take into account, besides the transport service class required by the destination user, the ...
详细信息
ISBN:
(纸本)0769523420
The paper proposes air architecture for a scheduling algorithm, to be integrated in IEEE 802.11 Access Points (AP), able to take into account, besides the transport service class required by the destination user, the channel quality experimented by the destination mobile STAtion (STA). The relevance of this topic is due to the observation that when one or more STAs experiment poor radio channel conditions, they increase the time spent to transmit a single packet, due to the retransmission of corrupted frames and the adoption of a transmission techniques at lower bit rate, leading to an inefficient use of the shared medium. These phenomena have as a consequence the worsening of the performance of all the STAs sharing the wireless medium independently of their radio channel conditions. The adoption of a scheduling algorithm able to manage information on channel quality permits to reduce these effects, not penalizing the STAs experimenting good channel condition and, as a consequence, their experimented throughput. As a result, only the STAs in bad channel conditions experiment a reduction of throughput. A prototype of an AP equipment implementing the proposed architecture is then presented;it has been obtained modifying the Hierarchical Token Bucket (HTB) and has been indicated as Wireless Hierarchical Token Bucket (WHTB). The performance of the presented prototype are experimentally evaluated and compared with those obtained with standard scheduling algorithm, which do not take into account information on channel quality.
暂无评论