달력

92019  이전 다음

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  •  
  •  
  •  
  •  
  •  

LINK : http://portal.acm.org/citation.cfm?id=1064984

ACM/Usenix International Conference On Virtual Execution Environments
Proceedings of the 1st ACM/USENIX international conference on Virtual execution environments, Aravind Menson, Jose Renato Santos, Yoshio Turner, G Janakiraman, Willy Zwaenepoel ( HP Labs )

Abstract
Virtual Machine (VM) environments (e.g., VMware and Xen) are experiencing a resurgence of interest for diverse uses including server consolidation and shared hosting. An application's performance in a virtual machine environment can differ markedly from its performance in a non-virtualized environment because of interactions with the underlying virtual machine monitor and other virtual machines. However, few tools are currently available to help debug performance problems in virtual machine environments.In this paper, we present Xenoprof, a system-wide statistical profiling toolkit implemented for the Xen virtual machine environment. The toolkit enables coordinated profiling of multiple VMs in a system to obtain the distribution of hardware events such as clock cycles and cache and TLB misses. The toolkit will facilitate a better understanding of performance characteristics of Xen's mechanisms allowing the community to optimize the Xen implementation.We use our toolkit to analyze performance overheads incurred by networking applications running in Xen VMs. We focus on networking applications since virtualizing network I/O devices is relatively expensive. Our experimental results quantify Xen's performance overheads for network I/O device virtualization in uni- and multi-processor systems. With certain Xen configurations, networking workloads in the Xen environment can suffer significant performance degradation. Our results identify the main sources of this overhead which should be the focus of Xen optimization efforts. We also show how our profiling toolkit was used to uncover and resolve performance bugs that we encountered in our experiments which caused unexpected application behavior.

필요 문맥 정리
[해결 방안 제시 - abstract]
 We user our toolkit to analyze performance overheads incurred by networking applications running in Xen VMs
[문제 제기를 위한 예 제시]
 2. Motivating example
[해결 방안 제시 및 설명]
 3.2 OProfile
 OProfile is a system-wide statistical profiling tool for Linux System ( 대상을 명확히 한다. )
[진행 방법 설명]
 Profiling with OProfile operates as follows : 1.. 2.. 3.. ( 번호를 매겨 순서대로 대상을 명확히 정의 )
[비교 대상 정의]
 We first compare the performance of the receiver in th Xen-domain0 configuration with its performance in a baseline Linux system.
[비교 대상 정의 이름 통일화]
 guest 는 모두 Xen-Guest0 와 같이 정의 (명확히 함)
[테이블 설명]
 Table 5 gives the breakdown of instruction counts across the guest and driver domains and Xen for the three configuration.
[결과 정리]
 In summary, 1... 2... 3... 진행 방법 설명과 유사. 번호를 매겨서 확실히 함. 비교가 쉬움.

Posted by Teshi
LINK : http://portal.acm.org/citation.cfm?id=1555384
ACM
Proceedings of the eleventh international joint conference on Measurement and modeling of computer systems 2009, Thomas Sandholm, Kevin Lai Hewlett-Packard Laboratories, Palo Alto, CA, USA

Abstract

We present a system for allocating resources in shared data and compute clusters that improves MapReduce job scheduling in three ways. First, the system uses regulated and user-assigned priorities to offer different service levels to jobs and users over time. Second, the system dynamically adjusts resource allocations to fit the requirements of different job stages. Finally, the system automatically detects and eliminates bottlenecks within a job. We show experimentally using real applications that users can optimize not only job execution time but also the cost-benefit ratio or prioritization efficiency of a job using these three strategies. Our approach relies on a proportional share mechanism that continuously allocates virtual machine resources. Our experimental results show a 11-31% improvement in completion time and 4-187% improvement in prioritization efficiency for different classes of MapReduce jobs. We further show that delay intolerant users gain even more from our system.

[측정 방법 설명]
To measure this effect, we introduce a total system efficiency metric that is based on th average ratio of actual application performance in a shared system to the application performance in a dedicated system.

[시나리오 설명]
2. Usage scenario
This section describes the usage scenario for the system described in this paper.

[질문을 제시하고 설명]
(1)How much do I want to spend?
.
(2) How do I want to spend?
.
(3)Should I spend more or less?

[수식 설명]
A rprivider allocates resource share qi to user i at time t as follows:[수식]

Posted by Teshi

Ganesha: Black-Box Diagnosis of MapReduce Systems
Link:http://portal.acm.org/citation.cfm?id=1710115.1710118
Xinghao Pan, Jiaqi Tan, Soila Kavulya, Rajeev Gandhi, Priya Narasimhan   Carnegie Mellon University, Pittsburgh, PA
SPECIAL ISSUE: Special issue on the workshop on hot topics in measurement & modeling of computer systems (HotMetrics 2009)

Abstract -
Ganesha aims to diagnose faults transparently (in a black-box manner) in MapReduce systems, by analyzing OS-level metrics. Ganesha's approach is based on peer-symmetry under fault-free conditions, and can diagnose faults that manifest asymmetrically at nodes within a MapReduce system. We evaluate Ganesha by diagnosing Hadoop problems for the Gridmix Hadoop benchmark on 10-node and 50-node MapReduce clusters on Amazon's EC2. We also candidly highlight faults that escape Ganesha's diagnosis.

[문제 설명]
There are manu challenges in problem localization and root-cause analysis.
[목표를 설명할 때]
Ganesha aims to diagnose faults transparently in MapReduce systems, by analyzing OS-level metrics.
[설명]
We describe Ganesha, our black-box diagnostic approach that we apply to diagnose such performance problems in Hadoop.
[메뉴 배열]
3.1 Hypotheses, 3.2 Goals, 3.3 Non-Goals, 3.4 Assumptions
[수식 설명]
Thus, the true-positive and false-positive ratios are computed as : [수식]

Posted by Teshi
Paper Link : ACM Portal

Thomas Karagiannis, Konstantina Papagiannaki, Michalis Faloutsos
ACM SIGCOMM Computer Communication Review,
Volume 35 ,  Issue 4  (October 2005), SESSION: Security, Pages: 229 - 240,Year of Publication: 2005, ISSN:0146-4833

 In this paper, they thought traffic classifiers of the future will need to classify traffic "in the dark" by application trends and the increasing use of encryption. So they suggest different approach to classifying traffic flow according to the applications that generate them. In contrast to pervious methods, there approach is based on observing and identifying patterns of host behavior at the transport layer. They suggest analyze these patterns at three levels, social, functional and application level.

 1. Taking into account empirical application trends ant the increasing use of encryption, we conjecture that traffic classifiers of the future will need to classify traffic "in the dark".
2. Furthermore, our approach has two important features.
3. In contrast to previous methods, our approach is based on observing and identifying patterns of host behavior at the transport layer.

Posted by Teshi
Hive: a warehousing solution over a map-reduce framework
Ashish Thusoo, Joydeep Sen Sarma, Namit Jain, Zheng Shao, Prasad Chakka, Suresh Anthony, Hao Liu, Pete Wyckoff, Raghotham Murthy 
August 2009   Proceedings of the VLDB Endowment , Volume 2 Issue 2 

Page Link

HIVE 의 생성 배경과 어떠한 구조로 동작하는지에 대해 설명하고 있다.
SQL Query 문을 지원하는 Component 를 Hadoop 상위 레벨에 두어 Hadoop의 단점인 Low Level, 어려운 프로그래밍을 해결하였다.
Posted by Teshi