1,720,970 research outputs found
Resource Placement and Scheduling for Distributed Systems
No full textApplying distributed systems is a typical solution for data intensive applications to collect large computational power to handle the enormous data. To enhance overall performance of the distributed systems, we need to address two important groups of problems about how to manage the distributed resources. The first group is how to place the resources at the proper locations of the network to achieve load balance, and the second one is how to schedule the requests of the shared resources to reduce the overhead caused by the requests that share the same resources.n the first problem group, we investigate the I/O server placement and data replica placement. Parallel I/O techniques can help solve the serious bottleneck of performance caused by I/O. However, switch-based clusters of workstations/PCs and distributed systems typically adopteneral topologies to allow the construction of scalable systems with incremental expansion capability. These general topologies lack many of the attractive mathematical properties of regular topologies, whichakes optimizing parallel I/O performance on general networks a difficult task. Therefore, we optimize server placement for parallel I/O in switch-Based clusters to balance the workload among the I/O servers. In addition, data replication is a typical strategy for improving access performance and data availability in distributedystems with data intensive applications (especially in Data Grids). The existing works usually focus on the infrastructure for data replication and the mechanism of replicas creation and deletion, but the important problem of choosing suitable locations for placing replicas has not been fully studied. Thus, we also address replicalacement problem in Data Grids.n the second problem group, we discuss parallel I/O scheduling and multicast scheduling. The lack of global information about I/O traffic between computing nodes and I/O servers impose new challenges in optimizing parallel I/O for distributed systems. Therefore, we developwo distributed algorithms for parallel I/O scheduling with non-uniform data sizes. Moreover, multicast is an important communication pattern, with applications in collective communication operations, and theandwidth limitation of the links in the routing tree for general topologies make multicast scheduling critical. Thus, we propose an agent based multicast algorithm that guarantee contention free multicast by exploiting the properties of routing tree for general network.ajor contributions of this dissertation are summarized as follow. First, in I/O server placement, we formulate the problem as a weighted bipartite matching with the goal of balancing the workload on the I/O servers, and we propose an efficient algorithm to find an optimal solution. To minimize link contention among the subclusters connecteds a general topology, we devise a tree-based heuristic algorithm to assign servers among subclusters. Our simulation results demonstrate that our best algorithm is near-optimal in some cases. Second, in replica placement in a Data Grid, we propose a placement algorithm thatinds optimal locations for replicas so that the workload among the replicas is balanced, and we also propose an algorithm that determines the minimum number of replicas when the maximum workload capacity of each replica is given. Third, in parallel I/O scheduling problem, weropose distributed scheduling algorithms, and our experimental results indicate that our algorithms yield parallel performance within 6% of the centralized solutions. We also compare the performance of ourlgorithms with a distributed Highest Degree First method, which divides non-uniform data transfers into units of fixed-sized blocks. The experimental results show that our algorithms require less scheduling and data transfer time. Finally, in multicast scheduling for general networks, our experimental results demonstrate that our agent-based algorithm outperforms the most efficient algorithm reportedn existing literature.1 Introduction 1 Related Works 6.1 Placement Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.1 I/O server placement . . . . . . . . . . . . . . . . . . . . . . . 6.1.2 Replicas placement . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Scheduling Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.1 Parallel I/O scheduling . . . . . . . . . . . . . . . . . . . . . . 10.2.2 Multicast on general networks . . . . . . . . . . . . . . . . . . 12 Optimizing Server Placement for Parallel I/O in Switch-Based Clus-ers 14.1 Server Placement in Same-Cost Clusters . . . . . . . . . . . . . . . . 16.1.1 System model . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.1.2 Problem definition . . . . . . . . . . . . . . . . . . . . . . . . 18.1.3 Load balance matching algorithm . . . . . . . . . . . . . . . . 20.2 Server Placement in General Networked Clusters . . . . . . . . . . . . 21.2.1 Up-down routing . . . . . . . . . . . . . . . . . . . . . . . . . 21.2.2 Problem definition . . . . . . . . . . . . . . . . . . . . . . . . 22.2.3 Load balance traversing . . . . . . . . . . . . . . . . . . . . . 24.3 Experiment Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.3.1 Experiments on same-cost clusters . . . . . . . . . . . . . . . . 40.3.2 Experiments on general-networked clusters . . . . . . . . . . . 44 Optimal Replica Placement in Hierarchical Data Grids 52.1 The Unconstrained Model . . . . . . . . . . . . . . . . . . . . . . . . 54.1.1 The Feasible algorithm for the FindR problem . . . . . . . . 56.1.2 Lazy updating . . . . . . . . . . . . . . . . . . . . . . . . . . . 60.1.3 The BinSearch algorithm for the MinMaxLoad problem . . . . 63.2 The Constrained Model . . . . . . . . . . . . . . . . . . . . . . . . . 64.2.1 The PlaceReplica algorithm for the FindR problem . . . . . 66.2.2 The BinSeek algorithm for the MinMaxLoad problem . . . . . 72.3 The Priority List Model . . . . . . . . . . . . . . . . . . . . . . . . . 73.3.1 Definitions of priority list . . . . . . . . . . . . . . . . . . . . 73.3.2 Replica placement problems . . . . . . . . . . . . . . . . . . . 74.4 The Sibling Tree Model . . . . . . . . . . . . . . . . . . . . . . . . . . 75.4.1 The PlaceR problem . . . . . . . . . . . . . . . . . . . . . . . 77.4.2 The LoadBalance problem . . . . . . . . . . . . . . . . . . . . 87 Efficient Distributed Scheduling for Parallel I/O 89.1 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90.2 Centralized Scheduling Algorithms . . . . . . . . . . . . . . . . . . . 92.2.1 The maximum weight matching algorithm . . . . . . . . . . . 92.2.2 The earliest completion first algorithm . . . . . . . . . . . . . 92.2.3 Effectiveness of centralized scheduling algorithms . . . . . . . 93.3 Distributed Scheduling of Parallel I/O . . . . . . . . . . . . . . . . . 94.3.1 Distributed matching algorithm . . . . . . . . . . . . . . . . . 94.3.2 Distributed greedy algorithm . . . . . . . . . . . . . . . . . . 95.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 An Optimal Agent-Based Scheduling for Multiple Multicast oneneral Networks 104.1 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105.1.1 Routing mechanism . . . . . . . . . . . . . . . . . . . . . . . . 106.1.2 Contention . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106.2 Agent-Based Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . 107.2.1 Single multicast . . . . . . . . . . . . . . . . . . . . . . . . . . 108.2.2 Multiple multicast . . . . . . . . . . . . . . . . . . . . . . . . 110.3 Message Forward Model . . . . . . . . . . . . . . . . . . . . . . . . . 111.4 Forwarding Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 114.4.1 Criticality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115.4.2 The testing algorithm . . . . . . . . . . . . . . . . . . . . . . 115.4.3 Optimal Schedule . . . . . . . . . . . . . . . . . . . . . . . . . 121.5 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Conclusion 128.1 Resource Placement Problems . . . . . . . . . . . . . . . . . . . . . . 128.1.1 I/O server placement . . . . . . . . . . . . . . . . . . . . . . . 128.1.2 Replica placement . . . . . . . . . . . . . . . . . . . . . . . . 129.2 Resource Scheduling Problems . . . . . . . . . . . . . . . . . . . . . . 131.2.1 Parallel I/O scheduling . . . . . . . . . . . . . . . . . . . . . . 131.2.2 Multicast scheduling . . . . . . . . . . . . . . . . . . . . . . . 132.3 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133ibliography 13
Growth Rank and Morphological Attributes of Four Fagaceae Species Seedlings Under Different Relative Light Intensities
No full text中文摘要
本研究之目的在探討四種殼斗科樹種幼苗，毽子櫟、赤皮、錐果櫟及鬼櫟對光環境之適應性。藉觀察三種相對光度下，種子發芽後至九個月生長期間之相對生長率，評估試驗樹種之生長優勢是否會有轉變，並由生物量分配策略及形態表現來瞭解植物如何適應不同光環境。
以相對生長率判定四種樹種之生長優勢轉變，其發生時間約為生長五個月後且保持其優勢度。在高光下（相對光度80%）下，錐果櫟及赤皮相對生長率較高，在高光下較具有優勢；在中低光下（相對光度10% 及5% ）下，則以毽子櫟及鬼櫟有較高相對生長率，此二種樹種在低光下較具有優勢。
光度與相對生長率之迴歸分析顯示，四樹種之相對生長率因光度增加而增加，錐果櫟與赤皮之相對生長率對於光度增加，呈線性反應，而毽子櫟與鬼櫟之相對生長率，則隨光度的增加而趨於平緩。
錐果櫟與赤皮之生物量，在高光下之累積遠大於中低光；對毽子櫟而言，中高光下生物量分配差異小，鬼櫟生物量在不同相對光度下仍有一定之累積量。
錐果櫟及赤皮苗高在相對光度80% 下較高，毽子櫟及鬼櫟苗高則在相對光度10% 下較高，表示毽子櫟及鬼櫟在低光下會以增加苗高而獲取更多光源。
錐果櫟、赤皮及鬼櫟在不同相對光度下，均以葉部生物量分配最多。毽子櫟生物量分配隨時間增加，以全方面發展並趨於一致，根莖葉分配比相近。
四種樹種之葉面積比與比葉面積皆隨著光度減少而增加，表示在低光下適應方式，以葉片變大變薄來增加獲取光源面積。各形態參數相關性分析顯示，相對生長率和葉面積比與比葉面積成負相關。葉部比和莖部比及根部比成負相關但和葉面積比成正相關。根部比和葉面積比和比葉面積成負相關。葉面積比和比葉面積成正相關。相對生長率為植物適應環境綜合表現的結果，此試驗中，並無法藉由形態上表現的參數作為很好的預測，此部分需待進一步研究。
由生長優勢轉變、相對生長率和光度的關係與苗高的反應，均可將四種樹種分為二群:（一）錐果櫟與赤皮在高光下較適合其生長，（二）毽子櫟與鬼櫟在中低光下生長較佳。驗證不同樹種對於光的適應性確實會不同，各樹種在不同光資源下會有生態棲位分化的現象。
關鍵詞: 相對生長率、相對光度、生長優勢Abstract
To investigate whether crossover in seedling growth ranks among four potentially co-occurring Fagaceae species, namely, Cyclobalanopsis acuta var. paucidentata, Cyclobalanopsis longinux var. longinux, Cyclobalanopsis gilva, and Lithocarpus lepidocarpus would occur between low and high irradiance, the relative growth rates (RGR) of seedlings grown under three relative light intensities (5%, 10%, and 80% of full light) were determined over a period of 9 months. In addition, to understand the likely seedling adaptation strategies of the four species under different relative light intensities, seedling biomass allocation patterns and morphological attributes were also investigated.
Results showed that the growth rank patterns were inconsistent among the four species during the first four months. However, from the fifth month and onward, Cyclobalanopsis longinux var. longinux and Cyclobalanopsis gilva had higher relative growth rates under 80% relative light intensity, whereas Cyclobalanopsis acuta var. paucidentata and Lithocarpus lepidocarpus had higher relative growth rates under both 10% and 5% relative light intensities. Thus, crossover in seedling growth ranks did occur among the four species.
For Cyclobalanopsis acuta var. paucidentata and Lithocarpus lepidocarpus, their RGR’s showed a curvilinear response to the three relative light intensities, whereas for both Cyclobalanopsis longinux var. longinux and Cyclobalanopsis gilva, the responses were linear.
For all the species, their seedlings biomass increased with increasing relative light intensity. Cyclobalanopsis longinux var. longinux and Cyclobalanopsis gilva had higher seedling height under 10% relative light intensity. In contrast, Cyclobalanopsis acuta var. paucidentata and Lithocarpus lepidocarpus had higher seedling height under lower relative light intensities. Regardless of relative light intensity, Cyclobalanopsis longinux var. longinux, Cyclobalanopsis gilva allocated more biomass toward the foliar component. In contrast, for Lithocarpus lepidocarpus, Cyclobalanopsis acuta var. paucidentata the biomass were equally partitioned among the foliar, stem, and root components.
For the four species, leaf area ratio (LAR) and species leaf ratio (SLA) increased with decreasing relative light intensity. For all four species, RGR was negatively correlated with SLA and LAR.
Based the responses to the three relative light intensities, the four species could be separated into two groups: Cyclobalanopsis longinux var. longinux and Cyclobalanopsis gilva could be considered as shade-intolerant species since they grew much better under high light environment than under low light environment; whereas Cyclobalanopsis acuta var. paucidentata and Lithocarpus lepidocarpus could be considered as shade-tolerant species since they still had good growth performance under relatively low light environment. This study suggested that one of the reasons that the four species could co-occur was in their abilities to differentiate along the light gradient.
Keywords: Relative growth rate, relative light intensity, growth rank目錄
中文摘要.............................................................................................…... i
英文摘要.............................................................................................…..iii
表目次………………………………………………………………...…vi
圖目次…………………………………………………………………..vii
壹、前言.....................................................................................................1
貳、前人研究…………………………………………………………....3
一、光度對林木個體發育的影響……………………………………3
二、光度對種間生長優勢的影響……………………………………3
三、光度對生物量分配的影響………………………………………6
四、試驗樹種特性及在台灣分佈……………………………………7
參、材料與方法.......................................................................................11
一、試驗樹種…………………………………………………………11
二、試驗設計…………………………………………………………15
三、環境變化…………………………………………………………18
肆、結果…………………………………………………………………24
一、生長優勢轉變…………………………………………………..24
二、生物量分配及形態表現………………………………………..33
伍、討論.............................................................................................…..53
一、種間相對生長率變化…………………………………………...53
二、生物量分配與形態表現………………………………………..56
陸、結論..............................................................................................….64
柒、參考文獻.....................................................................................…..65
表目次
表1、試驗樹種生態特性………………………………………………..8
表2、試驗樹種在台灣分布情形.............................................................10
表3、試驗樹種種子採集地點、時間與試驗播種粒數..........................11
表4、蔭棚內樹種配置情形....................................................................15
表5、蔭棚內溫濕度變化.................……………………………………19
表6、光度與後四期相對生長率分析之結果…………………………30
表7、中低光度與後四期相對生長率分析之結果……………………31
表8、不同相對光度與相對生長率之相關性分析.................................32
表9、相對生長率與生物量分配比之相關分析結果............................52
圖目次
圖1、2004年8月晴天及陰天光度之日域變化………………………20
圖2、2004年8月晴天及陰天溫度隨時間之日域變化………………21
圖3、2005年1月晴天及陰天光度隨時間變化之日域變化…………22
圖4、晴天及陰天溫度及濕度隨時間變化之日域變化………..……..23
圖5、各樹種前四期在不同光度下之相對生長率變化……….……...27
圖6、各樹種後四期在不同光度下之相對生長率變化……….……...28
圖7、光度與相對生長率後四期之關係……………………….……...30
圖8、中低光度與相對生長率後四期之關係………………………….31
圖9、四種樹種隨時間平均生物量變化……………….………………34
圖10、四種樹種在不同光度下苗高生長……………………..………37
圖11、四種樹種在不同光度下苗高生長比較……………………..…38
圖12、毽子櫟之各部位葉、莖、根比隨時間變化……………..………40
圖13、錐果櫟之各部位葉、莖、根比隨時間變化………………………41
圖14、赤皮之各部位葉、莖、根比隨時間變化……….…….…………43
圖15、鬼櫟之各部位葉、莖、根比隨時間變化………………..………44
圖16、各樹種葉莖根比在不同光度下分配比之比較…………………47
圖17、各樹種葉面積比隨時間變化......................................................49
圖18、各樹種比葉面積隨時間變化.....................................................50
圖19、各樹種於不同光度下葉面積比與比葉面積的表現…………..5
An Optimal Scheduling Algorithm for an Agent-Based Multicast Strategy on Irregular Networks
No full textAn Optimal Scheduling Algorithm for an Agent-Based Multicast Strategy on Irregular Networks
No full text
- …
