Revision as of 18:20, 1 December 2010

mClock: Handling Throughput Variability for Hypervisor IO Scheduling

Notes to Group

We might as well work directly on the main page

I think I've moved all the existing text to the main page, as things were being edited in 2 places. Hope that's okay. --Dagar 23:01, 30 November 2010 (UTC)

that's fine! hey guys are we good with the information that we gathered? i think we should find external links on mClock if possible. --npatel1

I can't any external links on mClock

Group Members

Please leave your name and email address if you are in the group

Daniel Agar - dagar@scs.carleton.ca
Xi Chen - xintai1985@gmail.com
Niravkumar Patel - npatel1@connect.carleton.ca
Tuan Pham - tpham3@connect.carleton.ca
Aaron Leblanc - aellebla@connect.carleton.ca
Nisrin Abou-Seido - naseido@connect.carleton.ca

Layout

Paper

Authors:

Ajay Gulati VMware Inc. Palo Alto, CA, 94304 agulati@vmware.com

Arif Merchant HP Labs Palo Alto, CA 94304 arif.merchant@acm.org

Peter J. Varman Rice University Houston, TX, 77005 pjv@rice.edu

Link to the paper:

mClock: Handling Throughput Variability for Hypervisor IO Scheduling

Background Concepts

Explain briefly the background concepts and ideas that your fellow classmates will need to know first in order to understand your assigned paper.

Research problem

What is the research problem being addressed by the paper? How does this problem relate to past related work?

Related work
- http://www.fortuitous.com/docs/whitepapers/Linux2.6-Perf-Intro.pdf
   -Linux schedulers (CFQ, SFQ)

Contribution

What are the research contribution(s) of this work? Specifically, what are the key research results, and what do they mean? (What was implemented? Why is it any better than what came before?)

Storage IO allocation is hard

mClockcontributions

 •Supports reservation, limit and shares in one place
 •Handles variable IO performance seen by hypervisor
 •Can be used for other resources such as CPU, memory & Network IO allocation as well

Future work

 •Better estimation of reservation capacity in terms of IOPS
 •Add priority control along with RLS
 •Mechanisms to set R, L,S and other controls to meet application-level goals

Critique

What is good and not-so-good about this paper? You may discuss both the style and content; be sure to ground your discussion with specific references. Simple assertions that something is good or bad is not enough - you must explain why.

Tuan,i think the term PARDA was explained in the article. It stands for Proportional Allocation of Resources in Distributed storage Access. It was basically a priority queue for the storage devices and VMs.--Aaron Leblanc 22:47, 30 November 2010 (UTC)

I see, I'll have to read it over again