Poster INGRID 07

1
Grid Economics and Business Models EC-Funded
Project 1st July 2006 30th April 2009
The Monitoring ServiceA System for Monitoring
Computing Resources and Applications in the
Computing Cloud

• Component Interaction
• The following sequence diagram illustrates the
  interaction of the monitoring service in greater
  detail. It indicates all the components of the
  GridEcon Platform that are involved.

• Characteristics
• The monitoring service supports the user in
• Starting an application in a Cloud computing
  environment
• Monitoring the resource usage
• Scaling up or scaling down the application on
  runtime, depending on the monitoring data.

• Scenario Description
• A provider places an Ask into the GridEcon
  Marketplace. A buyer places a Bid into the
  GridEcon Marketplace
• If the GridEcon Marketplace matches the Bid and
  Ask, it also produces a resource allocation
  that is scheduled for execution sometime in the
  future
• The user can monitor the application and the
  computing resource usage
• The user can add more machines to the cluster,
  when there is a high demand
• The user can remove machines from the cluster,
  when the demand decreases.

• Monitored Parameters
• The parameters can be grouped into two types of
  parameters
• Computing Resource Parameters
• Machines CPU usage
• Memory usage
• Communication.
• Application Parameters
• Number of reads operations
• Number of write operations
• Number of update operations.

The consortium would like to thank the European
Commission who have funded the project under
Framework Programme 6. The members of the
consortium are
For more information, please visit
http//www.gridecon.eu
2
Grid Economics and Business Models EC-Funded
Project 1st July 2006 30th April 2009
The Monitoring ServiceA System for Monitoring
Computing Resources and Applications in the
Computing Cloud
Step 2 Run your application
Step 1 Place a bid and ask an ask

• Place Bid and Ask for 5 machines
• Place bid and ask for 5 machines
• As soon as the ask is matched and the start
  time
• has reached the application is started.

Example Stock Monitoring Application
Step 3 Monitor the machines
Step 4 Increase Computation Power

• Use the clusters Web page to navigate to the
  following Web pages
• Stock Demo ltGrid_Manager Host Name gt/StockDemo/
• Monitoring ltUI Host Namegt/ganglia
• GS management tool ltUI Host Namegt.

Add one more machine to increase the performance
of the application
Step 5 Redistribute The processes
Step 6 Decrease Computation Power

• Move one of the processes to the newly added
  machine
• In the GS management tool, drag one of the
  stockSpaces and drop it into the new machine

Reduce the number of allocated machines by
removing one of the cluster machines
The consortium would like to thank the European
Commission who have funded the project under
Framework Programme 6. The members of the
consortium are
For more information, please visit
http//www.gridecon.eu