SQL Anywhere On Demand can be licensed by the "perpetual core" or by the "core hour." Here is the license definition of core hour:
The start and stop times - are those times when the server starts and stops, or when CPU usage by the database process starts and stops? For example, if my server is running for 100 hours on one core at an average usage by SAOD of 1%, am I using 100 core hours or 1 core hour?
Your example scenario would be billed at 100 core-hours.
The calculation of core-hours do not take into account the utilization of each core. Instead, it is based on the server's capacity to do work on a core. The start and stop times indicated are the start and stop times of a server that has the ability to use a given core.
The help section of the SQL Anywhere, on-demand Usage Server contains a number of example scenarios that may help to provide some clarity: https://saode.sybase.com/help/
This method of billing is consistent with the way that many core/cpu based utility computing charges are calculated. For example in the case of running virtual machines on public clouds such as Amazon EC2 or Microsoft Azure, you must pay for any running machine, whether or not that machine is actually doing any work.
SQL Anywhere, on-demand allows you to dynamically start and stop servers, as well as restrict servers to only running on certain cores. This allows you to dynamically adjust the number of cores that are being used by the servers. Running on the minimum amount of hardware is good practice in a cloud SaaS application. In your example, the 99 excessive cores of capacity are each consuming power, cooling, and maintenance (and may be being billed themselves at an hourly rate). Reducing the amount of hardware needed to run your database is just a piece of the general practice of reducing the amount hardware needed to run your whole application.
answered 23 Aug '12, 10:57