DV-ARPA：面向累积应用中大数据处理的数据多样性感知资源调配
DV-ARPA: Data Variety Aware Resource Provisioning for Big Data Processing in Accumulative Applications
摘要：
在云计算中，所使用的资源配置方式对处理成本有很大的影响，特别是在用于大数据处理时。由于数据的多样性，虚拟机（VM）的性能可能基于数据块的内容而不同。数据多样性无关分配导致虚拟机的性能降低，并增加了处理成本。因此，可以通过将VM与给定数据块匹配来降低作业的总成本。我们使用一种数据多样性感知的资源分配方法来降低所考虑的作业的处理成本。针对这个问题，我们将输入数据划分为若干数据块。我们定义每个数据块的重要性，并根据它选择合适的虚拟机以降低成本。为了检测每个数据部分的显著性，我们使用特定的采样方法。此方法适用于累积应用程序。我们使用一些著名的基准测试和配置的服务器进行评估。基于结果，我们的供应方法提高了处理成本，与其他方法相比高达35%。
英文摘要：
In Cloud Computing, the resource provisioning approach used has a great impact on the processing cost, especially when it is used for Big Data processing. Due to data variety, the performance of virtual machines (VM) may differ based on the contents of the data blocks. Data variety-oblivious allocation causes a reduction in the performance of VMs and increases the processing cost. Thus, it is possible to reduce the total cost of the job by matching the VMs with the given data blocks. We use a data-variety-aware resource allocation approach to reduce the processing cost of the considered job. For this issue, we divide the input data into some data blocks. We define the significance of each data block and based on it we choose the appropriate VMs to reduce the cost. For detecting the significance of each data portion, we use a specific sampling method. This approach is applicable to accumulative applications. We use some well-known benchmarks and configured servers for our evaluations. Based on the results, our provisioning approach improves the processing cost, up to 35% compared to other approaches.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
