| 題名 | Estimating CPU Utilization Data in PaaS Environment |
| 著者 | *Chao Luo, Hiroyuki Okamura, Tadashi Dohi (Department of Information Engineering, Graduate School of Engineering, Hiroshima University) |
| ページ | p. 188 |
| キーワード | Markovian arrival process, CPU utilization, Parameter Estimation |
| アブストラクト | PaaS (platform as a service) is a general architecture of cloud computing. It is adoptable for reducing maintenance cost of intra servers. How to estimate the performance of the virtual platforms and the capacity of virtual servers is the important issue before the integration.
This paper deals with the performance evaluation of a cloud server constituted by a PaaS architecture. More precisely, we describe the behavior of CPU in existing server by a simple queueing model, with assuming the input stream is according to a Markovian arrival process (MAP). Then, we represent the CPU behavior of integrated virtual platforms by applying the superposition technique of MAPs. Additionally, we introduce the estimation method for deriving mean CPU utilization from real data. |
| 題名 | Survivability Analysis for MANET with Battery Charge |
| 著者 | *易 志鵬, 土肥 正 (広島大学大学院工学研究科情報工学専攻) |
| ページ | pp. 189 - 190 |
| キーワード | network survivability, MANET, semi-Markov process |
| アブストラクト | Network survivability is a property that network is continually available even if a failure occurs, and is one of the most important concepts to design fault-tolerant computer networks. In the literature [1], we have evaluated a power-aware mobile ad hoc network (MANET) by using a Markov regenerative process (MRGP). However, this model did not consider the situation where the node in low-battery state can be recharged. In this paper, we revisit a power-aware MANET model [1], and compare two network survivability models with and without battery charge. |