Currently, decentralized distributed storage system is widely used in various scenarios, because decentralized data placement algorithms observably improve scalability and robustness of distributed storage system. How...
详细信息
Currently, decentralized distributed storage system is widely used in various scenarios, because decentralized data placement algorithms observably improve scalability and robustness of distributed storage system. However, storage system applied with decentralized data placement algorithm suffers from uncontrollable data migration during expanding the clusters, which lead to dramatic fluctuation of performance. To maintain reliability and stability of storage system, administrators are under great pressure to migrate data during expansion or contraction. Not only the urgent time constraints to expansion but also low interference to foreground application are required during data migration period. It is a difficult challenge for in-experienced administrators to tackle. To address this challenge, this paper introduces Flimm, a fast and low-interference data migration manager that is aware of storage resource utilization and performs data migration with low interference to foreground applications. Flimm mitigates interference by scheduling a subset of migration tasks with careful consideration. Moreover, Flimm adapts to varying foreground application workloads through adaptive migration speed control, providing administrators with convenient and secure parameters for tuning. In our experiments, Flimm demonstrated a 1.2x acceleration in migration speed and achieved a 20% reduction in interference.
Security has been a serious concern in more and more embedded systems. But restricted by the constrained resources in embedded system, security may damage the system real-time performance, we refer to this challenge a...
详细信息
ISBN:
(纸本)9781479974344
Security has been a serious concern in more and more embedded systems. But restricted by the constrained resources in embedded system, security may damage the system real-time performance, we refer to this challenge as the "security real-time gap". To bridge this gap, trusted computing technology can be a good candidate, however, researches focusing on real-time performance in embedded systems employing trusted computing is still blank. In this paper, we investigate the dynamic scheduling of hybrid tasks with trusted computing constraint, and present a hybrid tasks model to describe the tasks which need trusted computing resources and a strategy to manage the trusted requests. Furthermore, an advanced dynamic hybrid task schedule algorithm based on total bandwidth server (TBS) algorithm is presented in this paper, used for enhancing responsiveness performance of aperiodic tasks. In the evaluation by simulation, the proposed method reduced average response times of the aperiodic task by up to 30.4%.
In this paper, we extend and rewrite MapReduce dispatcher and its quasi-real-time schedule algorithm to support operation scheduling in time-limited. MapReduce dispatcher has an evaluation of completion time of tasks ...
详细信息
ISBN:
(纸本)9783037856932
In this paper, we extend and rewrite MapReduce dispatcher and its quasi-real-time schedule algorithm to support operation scheduling in time-limited. MapReduce dispatcher has an evaluation of completion time of tasks in dependence of rate of progress of tasks at hand, and allocated resource dynamically to every task when they are running. Experimental investigation shows that, the algorithm increase the resource utilization of the MapReduce system, and the goals of quasi-real-time MapReduce schedule has been achieved.
Reducing energy consumption is a critical problem in most of the computing systems today. Among all the computing system components, processor and memory are two significant energy consumers. Dynamic voltage scaling i...
详细信息
Reducing energy consumption is a critical problem in most of the computing systems today. Among all the computing system components, processor and memory are two significant energy consumers. Dynamic voltage scaling is typically applied to reduce processor energy while sleep mode is usually injected to trim memory's leakage energy. However, in the memory architecture with multiple cores sharing memory, in order to optimize the system-wide energy, these two classic techniques are difficult to be directly combined due to the complicated interactions. In this work, we explore the coordination of the multiple cores and the memory, and present systematic analysis for minimizing the system-wide energy based on different system models and task models. For tasks with common release time, optimal schemes are presented for the systems both with and without considering the static power of the cores. For agreeable deadline tasks, different dynamic programming-based optimal solutions are proposed for negligible and non-negligible static power of cores. For the general task model, this paper proposes a heuristic online algorithm. Furthermore, the scheme is extended to handle the problem when the transition overhead between the active and sleep modes is considered. The optimality of the proposed schemes for common release time and agreeable deadline tasks are proved. The validity of the proposed heuristic scheme is evaluated through experiments. Experimental results confirm the superiority of the heuristic scheme in terms of the energy saving improvement compared to the most related existing work.
Because of the high attenuation, sound is mainly used as the communication medium in underwater sensor networks instead of light or microwave. MAC protocols designed for radio networks aren't suitable for acoustic...
详细信息
ISBN:
(纸本)9781479987306
Because of the high attenuation, sound is mainly used as the communication medium in underwater sensor networks instead of light or microwave. MAC protocols designed for radio networks aren't suitable for acoustic channels which feature long propagation delay and limited bandwidth. In this paper, we propose a location-based MAC protocol, called LT-MAC. LT-MAC dynamically determines each node's sending time according to the variable offered load, and assigns the time slot to each node based on the minimum waiting time rule. In addition, the protocol combines sending and acknowledgement to overcome low throughput and long end-to-end delay problems. By theoretical analysis and simulation, we show that LT-MAC is a comparatively good solution for single-hop and small-scale underwater sensor networks.
Because of the high attenuation, sound is mainly used as the communication medium in underwater sensor networks instead of light or microwave. MAC protocols designed for radio networks aren't suitable for acoustic...
详细信息
ISBN:
(纸本)9781479987313
Because of the high attenuation, sound is mainly used as the communication medium in underwater sensor networks instead of light or microwave. MAC protocols designed for radio networks aren't suitable for acoustic channels which feature long propagation delay and limited bandwidth. In this paper, we propose a location-based MAC protocol, called ***-MAC dynamically determines each node's sending time according to the variable offered load, and assigns the time slot to each node based on the minimum waiting time rule. In addition, the protocol combines sending and acknowledgement to overcome low throughput and long end-to-end delay problems. By theoretical analysis and simulation, we show that LT-MAC is a comparatively good solution for single-hop and small-scale underwater sensor networks.
As the CPU/FPGA hybrid architecture becomes one of the most important forms, the research of the schedule on hybrid tasks is a hot-spot. The schedule approaches of hybrid tasks can be classified into two categories. O...
详细信息
ISBN:
(纸本)9780769547497
As the CPU/FPGA hybrid architecture becomes one of the most important forms, the research of the schedule on hybrid tasks is a hot-spot. The schedule approaches of hybrid tasks can be classified into two categories. One is about non-preemptive multitasking, the other is preemptive multitasking. For the second one, as it requires architecture support, the existing researches are fewer. This paper presents an architecture which supports preemptive multitasking, an algorithm to select the appropriate migration point and a new schedule algorithm based migration and also introduces naming service to deal with communication problems after tasks migrated. Experimental results show that the algorithm based migration increases acceptance rate for hardware tasks to 12.94%. The architecture proposed can support the preempting well, and improve the performance.
On-board switching has been the development trend of satellite communication. There is electronic bottleneck in the development of on-board ATM switch, so the optical switch is the inevitable direction of future broad...
详细信息
ISBN:
(纸本)9781849195072
On-board switching has been the development trend of satellite communication. There is electronic bottleneck in the development of on-board ATM switch, so the optical switch is the inevitable direction of future broadband multimedia satellite. The paper brings forward a scheme about on-board mixed optical/electronic switching of GEO, analyses the switching mode of the scheme, discusses the key techniques.
In the pull-based wireless network, clients request data items from a server. In order to solve cold and hot data and reduces clients' waiting timer, server considers the hit rate of data items and waiting time of...
详细信息
ISBN:
(纸本)9788955191318
In the pull-based wireless network, clients request data items from a server. In order to solve cold and hot data and reduces clients' waiting timer, server considers the hit rate of data items and waiting time of requests for broadcasting. The hit rate of data items and the waiting time of requests always changed along with time elapsed. Therefore, server has to calculate overall weight value of data items and decides maximum weight value of data items before broadcasting. In order to increase performance of server, we propose an effective dynamically schedule algorithm (EDSA) for wireless broadcast. The EDSA uses weight value of data items and maintains them in a maximum heap tree for quickly broadcasting. Furthermore, we present a indexing for accessing a data item's information faster. The experiment results showed that our method has more effectively than a generally algorithm in the dynamic wireless environment. However, it is still have some problems about our EDSA, therefore we propose an Improve-EDSA to correct the problems.
Channel aware and opportunistic scheduling algorithms exploit the channel knowledge and fading to increase the average throughput. Alternatively, each user could be served equally in order to maximize fairness. Obviou...
详细信息
Channel aware and opportunistic scheduling algorithms exploit the channel knowledge and fading to increase the average throughput. Alternatively, each user could be served equally in order to maximize fairness. Obviously, there is a tradeoff between average throughput and fairness in the system. In this paper, we study four representative schedulers, namely the maximum throughput scheduler (MTS), the proportional fair scheduler (PFS), the (relative) opportunistic round robin scheduler (ORS), and the round robin scheduler (RRS) for a space-time coded multiple antenna downlink system. The system applies TDMA based scheduling and exploits the multiple antennas in terms of spatial diversity. We show that the average sum rate performance and the average worst-case delay depend strongly on the user distribution within the cell. MTS gains from asymmetrical distributed users whereas the other three schedulers suffer. On the other hand, the average fairness of MTS and PFS decreases with asymmetrical user distribution. The key contribution of this paper is to put these tradeoffs and observations on a solid theoretical basis. Both the PFS and the ORS provide a reasonable performance in terms of throughput and fairness. However, PFS outperforms ORS for symmetrical user distributions, whereas ORS outperforms PFS for asymmetrical user distribution. Copyright (C) 2009 Eduard A. Jorswieck et al.
暂无评论