IT Capacity Planning for and Economics of Cloud Computing Part 2 - Examples Ivan L. Gelb Gelb Information Systems Corp. Phone: CMG 2010 – Monday, Dec. 6, 2010, Orlando, FL Gaylord Palms Hotel, “Sunday” Workshop Session

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Attention CMG 2010 Workshops Attendees  This presentation is prepared many weeks before the CMG 2010 Conference in Orlando, FL.  Everyone knows that in the field of Cloud Computing technology, the only certain guarantee is ongoing change  This version of the “Capacity Planning for and Economics of Cloud Computing” is subject to updates before the actual presentation date.  We encourage CMG 2010 Workshops registered attendees to request the latest version of this presentation via an to

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information TRADEMARKS and COPYRIGHT  Trade or service marks found in this presentation belong to their respective owners. Any omissions are purely unintended.  © 2011 Gelb Information Systems Corp. URL: Phone: http://  No part of this material can be reproduced by any means without prior written permission from the author and with proper attribution displayed.

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information MOTHER OF ALL DISCLAIMERS (MOAD ) 2  Samples and examples in this seminar are for education and discussion purposes. The samples from various vendors are in no way to be construed as: (a) the only way to do the task, and (b) the author’s recommendation of the products.

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Contents  What Are We Trying to Avoid?  Free Cloud Monitoring by Hyperic  Amazon EC2 Monitoring by Hyperic  Not free from Hyperic  “Guarantee” Architecture by Oblicore-CA  One Slowdown Analysis  Is This a Capacity Problem?  Total CPU by Workload  Total CPU: 15 Min Intervals

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information What Are We Trying to Avoid? More on today's Gmail issue Tuesday, September 01, 2009 | 6:59 PM Posted by Ben Treynor, VP Engineering and Site Reliability Czar Gmail's web interface had a widespread outage earlier today, lasting about 100 minutes. We know how many people rely on Gmail for personal and professional communications, and we take it very seriously when there's a problem with the service. Thus, right up front, I'd like to apologize to all of you — today's outage was a Big Deal, and we're treating it as such. We've already thoroughly investigated what happened, and we're currently compiling a list of things we intend to fix or improve as a result of the investigation. Here's what happened: This morning (Pacific Time) we took a small fraction of Gmail's servers offline to perform routine upgrades. This isn't in itself a problem — we do this all the time, and Gmail's web interface runs in many locations and just sends traffic to other locations when one is offline. However, as we now know, we had slightly underestimated the load which some recent changes (ironically, some designed to improve service availability) placed on the request routers… Note: To see the entire blog item, please visit the source cited below. Source: cited in A Measured Approach To Cloud Computing: Capacity Planning and Performance Assurance by Annie Shum at BSMReview.comhttp://gmailblog.blogspot.com/2009/09/more-on-todays-gmail-issue.htmlAnnie Shum

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Free Cloud Status Monitoring by Hyperic Source: “Free real-time health and performance monitoring of Amazon and Google cloud services Hyperic's CloudStatus BETA is the first service to provide an independent view of the health and performance of the most popular cloud services on the web. CloudStatus gives businesses that use the cloud an independent perspective into the performance of cloud-based services, and a framework for determining the cause when the performance of their cloud-hosted applications changes. CloudStatus is a free service that provides real-time reports and weekly trends on infrastructure metrics-including service availability, response time, latency, and throughput-that affect application availability and performance. “

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Amazon EC2 Monitoring by Hyperic

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Not Free from Hyperic

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information “Guarantee” Architecture by Oblicore-CA

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information One Slowdown Analysis During this slowdown: (1)One Virtual Machine’s (VM) CPU utilization increased by about 29% - from 38 to 50%; (2) VM-1’s CPU share was limited during this slowdown at 50% because the other VM-2 also needed its 50% CPU share, and (3) VM-1’s DASD IO/SEC. rate increased by 23%. May look like cause of slowdown, but it is normal activity associated with increased business activity rate. Fact verified with other peak load measurements.

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Is This a Capacity Problem?

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Sample: Total CPU by Workload

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Sample: CPU Time in 15 Min Intervals

© 2011 Gelb Information Systems Corp Think Faster with Gelb Information Time for…

Speaker Information Ivan Gelb © 2010 Gelb Information Systems Corp. (GIS) Phone: Your questions and comments are always