Machine Utilization Example

Amazon SimpleDB measures the machine utilization of each request and charges based on the amount of machine capacity used to complete the particular request (QUERY, GET, PUT, etc.), normalized to the hourly capacity of a circa 2007 1.7 GHz Xeon processor. Machine utilization is driven by the amount of data (# of attributes, length of attributes) processed by each request. A GET operation that retrieves 256 attributes will use more resources than a GET that retrieves only 1 attribute. A multi-predicate QUERY that examines 100,000 attributes will cost more than a single predicate query that examines 250.

Source: http://groups.google.com/group/cloud-computing/browse_thread/thread/e6e985a4acb279bc?q=#3b48b85ee8fd116d 

In the response message for each request, Amazon SimpleDB returns a field called Box Usage. Box Usage is the measure of machine resources consumed by each request. It does not include bandwidth or storage. Box usage is reported as the portion of a machine hour used to complete a particular request. The cost of an individual request is Box Usage (expressed in hours) * $0.14 per Amazon SimpleDB Machine hour. ( http://www.utilitystatus.com/images/FIG_2.jpg ) The cost of all your requests is the sum of Box Usage (expressed in hours) * $0.14.

For example, if over the course of a month, the sum of the Box Usage for your requests uses the equivalent of one 1.7 GHz Xeon processor for 9 hours, your charge will be:

9 hours * $0.14 per Amazon SimpleDB Machine hour = $1.26.

( http://www.utilitystatus.com/images/FIG_6.jpg ) Nice, metering in actual elapsed CPU time. How else are you going to measure software usage ?

More info:  http://www.utilitystatus.com/features/

It is a story of rise and fall for the stock market as you know these days, ok ok more fall than rise. Does what goes up must go down also apply in the other direction as well? I think so, it can only get better from here. There was a system outage on Google Finance ticker for GOOG on September 30th 2008, the stock went to zero for 20 minutes, see slides below. GOOG stock holders could'nt see what was happening.

During that time and the begining of the financial crisis I was implementing a Monitoring-as-a-Service (TM) solution at the Depository Trust and Clearing Corporation for thier private cloud.

Here are a few snapshots of the slides I will be presenting tomorrow in San Francisco, CA. Your welcome to join here is the invitation.

Date: Nov 20th, 2008
Meeting time: 12:00 PST
Meeting location: 425 Market Street 19th floor, Room 201 San Francisco, CA
Meeting topic: ITCAM Implementation Case Study by Jason Meiers of CAM Solutions

Are you still monitoring your apps using log files or building custom JMX monitoring apps? You got to see this presentation. 

Wouldn't it be nice to be able to consume utility computing services and tie them to a business process with chargebacks based on actual CPU elapsed time to your SaaS application or Cloud Computing Infrastructure.

This example shows how a business process is created by consuming utility computing services for a StockQuoteService. The SaaS provider offers two types of service, one a free service that has delayed stock quotes and the other a premium utility service that offers real-time quotes.

Build SaaS applications how your customer needs them based on changing business requriements. Built in chargback models for services are key to ensure sucess of running a productive business. You haven't built your SaaS application to run as Beta for always you know. i.e Facebook, Linkedin, Yahoo Utilities, SAP WebServices...

Amazon EC2 running Zen virutalization has proven to be an affordable approach to cloud computing and many small to large enterprise leverage their technology. There are a few concerns today about EC2 shared CPU architecture and security holes in the Zen platform. Many cases customers would rather not share CPU cycles with different applications or surly not compete with other companies for CPU cycles.

VMWare fully supported and commercial product hosted in data centers to offer most features virtualization has to offer as well as production scenario examples how to keep SaaS applications available. Easly build virtual machies from your desktop and deploy images directly to staging and production servers.

AIX with it's built in virtualzation enables not only a smaller footprint for the virtual images that both Zen and VMWare but also includes a nice feature for micropartitioning. The technology for LPAR virtualization has not only been around for a while but has proven to be the virtualzation for large financial institutions as well as mission critical government applications as well.

Although iDataplex is not virtualization I would like to include it in this list as well has interesting features as well and is a interesting solution for solving typical scalability challenges for SaaS applications. Each MAC address is known thoughout the iDataplex cloud regardless of its growth challenges. This enables a scalability model in a true sense of "add as you grow".

Here is the link for the article how Microsoft is dipping it's toes on how to improve cloud computing service level agreements. [link]

This describes how service levels need to be improved or even added as most software-as-a-service vendors do not provide SLA's at all not to mention report monitoring and availability metrics. Most interesting is the importance of violations of a service level and how to deal with incidents.

A refund of a service outage is the improvement and an addtional step from where we are today for software-as-a-service. In example, see a few sommers back with traditional electrical companies service outages in New York City. Cloud computing at Microsoft, Google and Amazon is internet centric computing simular to the traditional electric company i.e. PG&G. If an outage at PG&G occurs your are not billed for actual unsed service. Today in Software-as-a-Service your are billed monthly regardless if the service is available or if the service is being used at all.

Are you upgrading to 6.2.1? This release is worth it just alone because of the multi-threaded capability. Here are some more reasons to upgrade to the current version.

Highlights:

• - Dynamic threasholds (baseline)
•    - adaptive theresholds (learn)
     - local or remote management of dynamicthreasholds
     - extension to the situation editor
     - override expressions
• - Green Maps
• - Sub-node agents
• - Script and log Editor
•    configure log file and script response monitoring with visual editor to configure parsing.
  
• - Agent/agentless monitoring
• - Multi-threaded agent deployment
• - CLI Enhancements,
• goal to have all commands from graphical also avalable as CLI
• - Graphing zoom function
• - Security
•  Universal Agent IPv6 Support, SSO with Java WebStart Client, FIPS compliance fror TEP/S
• - Netcool agent deployment
•   update, config changes with TEMS from ITM
• - Improved TPM / TCM integration with ITM
• - Editor for Baroc classes
• - Watchdog for agents
•   monitor the availability of end-points and self-heal if agents is unavailable

Source: Tivoli User Group

Google is offering today access to the first search index 10 years ago. Out of curiosity I entered the term "Cloud Computing"

http://www.google.com/search2001/search?q=%22cloud+computing%22&hl=en&btnG=Search

MD19.2 Intermediaries in Cloud-Computing: A New Computing Paradigm Ramnath Chellappa --- Univ. of TX, Ctr. for Res. on Elect. Comm., MSIS Dept., Grad. Sch. of Bus., Austin, TX 78712 , (ram@cism.bus.utexas.edu)
    Computing has evolved from a main-frame-based structure to a network-based architecture. While many terms have appeared to describe these new forms, the advent of electronic commerce has led to the emergence of 'cloud computing.' This work aims at analyzing the role of agents and intermediaries enabling this framework.

*Download and evaluate free trial versions of AppScan 7.6.
http://www.watchfire.com

*For the latest whitepapers
http://www.watchfire.com/news/whitepapers.aspx

bluehouse.lotus.comIBM announced it's cloud services today here is the link for more details about the press release. (http://www-03.ibm.com/press/us/en/pressrelease/25341.wss)

Products include Bluehouse (SaaS Collaboration Services, https://bluehouse.lotus/ ), AppScan (Application Security)  and SametimeUnyte ( http://www.sametimeunyte.com/ ). Sofar I have been really impressed with the products I have used sofar. Recently I had the opportunity to participate in a 3 hour hands on demonstration for AppScan here in San Francisco. The presentation exceeded my expectations from a traditional hands on traning from different vendors. Sure there was the introduction to the enterprise version of the tool but also when someone asked if AppScan was also available as SaaS, the normal look of "say what?" was replaced with sure SaaS, we got it you can signed up via computer with no server software to deploy and manage.

AppScan includes a living library of application security processes and solutions for keeping your online applications secure from online hacking threats. A lot of times when your focus is on architecture and systems engineering you loose site of the new security threats that are at risk. Thats where AppScan came in to help with my business application security process. For small businesses without the budget for a Chief Security Officer and crowd the SaaS version of the tool is a no brainer. Security on demand. 

First look at BlueHouse was pretty impressive, this tool operates simular to LinkedIn that has an enterprise Webex built in, and the url reminded from alot of the enterprise colloboration tools from Lotus http://bluehouse.lotus.com/ . After the simple signup process there is the immediate opportunity for hosting a meeting.

BlueHost worked perfectly on the first shot, once you hit join a meeting the meeting session begins and you can add email address to whoever needs to join. Also the URL's have improved compared to Webex or other collaboration tools. The URL to my meeting is one you can actually remeber, which is pretty handy when traveling or putting it on a business card. May even improve the value for you business card.

Hope this helps get a short view of the new IBM cloud offerings.

Jason Meiers

CAM Solutions

https://bluehouse.lotus.com/meetings/join?id=415-135

Description

Utility computing is the realization of "on-demand" computing which provides services for Cloud Computing, IT Departments and Mobile Networks. Services provided by utility computing providers are charged on actual elapsed CPU time used, simular to traditional electricity companies that charge in actual kilowatt used. Service level agreements between the utility computing provider and consumer for the service are mandatory for establishing reliable networks for business transactions. Without services level agreement contracts in place the availability of the utility computing resource cannot be guaranteed. Each utility service provider contributes as a paid service to the success of the overall transaction of the consumer. The utility computing company offers a list of services in form of a service catalog from which the consumers can purchase computing utilities. The service catalog contains services offered by the utility computing company.

 Benefits

- The availability and reliability of computing resources enables access to innovative computing services that generate new opportunities for small, medium and enterprise businesses.

- Intellectual property is protected by providing software, hardware and processes as a managed services.

- The cost effectiveness to purchase utility computing resources based on actual elapsed CPU time used for a service, enables access for small, medium, and enterprise businesses. This includes savings in travel cost and implementation cost for the consumer.

Exclusion

This section describes what utility computing is not:

- Cloud Computing

- Grid Computing