It is hard to do such tests on production servers,
But i found similar tests on the internet.
http://www.mysqlperformanceblog.com/2010/05/25/flashcache-tpcc-workload/
PS: You can buy separate BBU (battery backup unit) for ssd (in order not to lose info in case of power fail)
Btw google have BBU in every server
PPS: also it is a good idea to use raid1 and/or master/slave configs
ssd can (and will eventually) fail
PPPS: you should tune your scheduler, some system params to get most of ssd. (do noatime and nodiratime at least)
other ideas: ext4 / nojournal, | XFS nobarrier
echo deadline > /sys/block/sda/queue/scheduler
echo 1 > /sys/block/sda/queue/iosched/fifo_batch
always keep 50%+ of free space on device.
check: http://www.slideshare.net/matsunobu/ssd-deployment-strategies-for-mysql
parf ()