Le 12 déc. 2009 à 18:52, Stuart McCulloch a écrit : > 2009/12/12 Niclas Hedhman <[email protected]> > Gang, > I just want to provide a general "warning" for EC2 charge model. > > From the EC2 marketing page; "On-Demand Instances let you pay for > compute capacity by the hour." > > But that is NOT CPU hour, but "wall clock" hour as they say on > http://stackoverflow.com/questions/158765/amazon-ec2-cost > > reminds me of mobile phone plans (at least before they were simplified) > so hard to compare different plans and find the right one for your usage >
I run a server on EC2 since 6 months, no real surprise regarding the bill (almost always around 75$/month). Testing and customizing a disk image takes time and you will probably reboot / relaunch many times. Since any hour/instance started is paid in full, you may end up with surprises when testing load balancing on three nodes since updating/restarting the disk image three times in an hour will actually cost you nine. But once you are ready for production, you can opt for a 'reserved instance' 3 year term for 40 $ / month Or, 3 instances with load balancing and 10GB/months traffic cost around 150 $ /month I did not find any offer able to compete with that (until know). Right now I am planning to use the S3 EntityStore, once JClouds is updated and Infinispan is released, on EC2 with load balancing and auto scaling. Infinispan will be configured in distribution mode using jGroups S3_PING. I will test this setup in the near future and keep you posted about its efficiency. Philippe > > -- > Cheers, Stuart > _______________________________________________ > qi4j-dev mailing list > [email protected] > http://lists.ops4j.org/mailman/listinfo/qi4j-dev
_______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
