Finding a size level per partition
-
2012年4月15日 下午 07:50
Hi
Working on a project, where users can create events, and others can attend the events. To model this in Table Storage, my first idea was to model each Event as a seperate partition as this would spread the load on each event. But as each event might have between 10-100 users, this might be a too small partition size.
If I on the otherhand put all events in the same partition, it will not be scaleable or am I missing something?
/Rasmus
Developer
所有回覆
-
2012年4月15日 下午 08:27Tiny partitions may not be efficient but are not incorrect and certainly better than too large a partition. So, partition key = event id is quite reasonable.
You may want to also think about deleting events older than a certain date. In that capacity having the event id as a combination of a chronological identifier like year or year + month or date and a base event id could be of value. -
2012年4月15日 下午 08:52I will continue in this thread, as the answer might be the same: http://social.msdn.microsoft.com/Forums/en/windowsazuredata/thread/a3172f1d-9188-4651-9907-897a168339a2
Best regards Rasmus Christensen
-
2012年4月16日 上午 03:50版主
Hi,
Please refer to another thread you've posted on this forum.
Thank you.
Please mark the replies as answers if they help or unmark if not. If you have any feedback about my replies, please contact msdnmg@microsoft.com Microsoft One Code Framework
- 已標示為解答 Rasmus Christensen 2012年4月22日 下午 01:03

