General question: why are there certain limitations w/ MySQL cluster?

[mike]

I've always been confused as to why MySQL cluster has so many limitations or "fine print" compared to the normal server. Is there no way to implement the clustering architecture with all the same features? Things like VARCHAR support *just* being introduced make me nervous about considering this for my application. I'm not sure why it couldn't have been done the same way the other storage engines are done?

Perhaps it was because it had to be stored in memory, but now with the disk-based storage coming in 5.1, shouldn't it be getting pretty close to matching all the functionality of the existing (at least) MyISAM format?

I do understand probably the biggest things are auto_increment and transaction support since those will cause data discrepancy and/or race conditions. That seems like the #1 thing stopping a general ability to setup N number of MySQL servers and have them all interact properly together. The auto_increment issue can be resolved in some other ways, as well.

I just wanted to bring this up, it may be a duplicate topic, but it would help me understand why there are some of the limitations there are and a general feeling of where it's going.

Thanks,
mike

please reply to my thread about mysql usage - i'd like to collect some data! thanks!
http://forums.mysql.com/read.php?11,57080



Edited 1 time(s). Last edit at 11/11/2005 09:46PM by mike .