This discussion is rapidly becoming irrelevant... we were originally talking about MySQL and now see where we've ended up...
On Sun, Mar 22, 2009 at 2:02 AM, fooler mail <[email protected]> wrote: > On Sat, Mar 21, 2009 at 10:36 PM, Orlando Andico <[email protected]> > wrote: >> > > now we are dealing another topic to debate but ill answer you... > >> Teradata isn't used for transactional at all.. > > teradata database specializing in data warehousing but their database > is capable of transactional too.. it can do transaction rollback (eg. > begin transaction (BT) and end transaction (ET) commands)... > > look at their customers section at wikipedia... > > http://en.wikipedia.org/wiki/Teradata > > bank of america.. you know how highly critical monetary transactions > are there... So we're back to quoting Wikipedia. The existence of a transactional feature does not mean that it works well. You can do transactions in SybaseIQ for example but they are painfully slow. Most banks do have data warehousing loads, so just because BofA uses Teradata doesn't mean they use it for transactional loads. There's not much point really bandying Wikipedia-based arguments around. From my own personal knowledge, which isn't that vast I admit, .. >> amusingly, all of these are from MySQL Cluster's wikipedia entry.... >> but I digress. > > but that proves that they used mysql clustering and they are happy > with it without a problem on their side... No, you are reading too much into a Wikipedia entry. Not all references are "happy" and don't have problems. >> And yes... for the commodity market, everybody is still going with >> SMP. My point is that MPP is either exotic or expensive. Expertise is >> harder to come by. > > the argument here is scalability... when SMP cant scale.. the next > step is MPP.. No, my argument was that SMP (or shared-all) can scale, to a certain extent. And the extent of that scaling still exceeds most use case requirements. >> Shared-nothing is a well-known scalable technology, but for the vast >> majority of the market, it's still exotic. > > exotic is meaningless if you need to scale as your workload grows... Yeah... but I personally know of no use cases that absolutely require MPP-class scaling. Of course there will always be some use case which requires massive scaling. On all the things I've worked on, I've only seen one which requires more scaling than traditional shared-all can provide. .. >> Which is my entire point -- shared-nothing requires more rocket >> science skills than shared-everything. Of course we all know the >> limitations and performance plateau of shared-everything: but that >> plateau is still high enough for most use cases. > > orly because your mindset wont adapt to the next level...your mindset > is just like an old people that dont want to touch that computer or > cellphone as it is hard for them to learn... Oh, so I've got an obsolete mindset now. Please go look up Oracle Coherence. I've been workign with it for a couple years now. Unfortunately most people DO have a conventional mindset and nobody's deploying this solution in ASEAN. .. >> But again, that's not something we on PLUG generally do as part of our >> daily lives. > > but someday they do and others are on it... True. But it's irrelevant to the original topic of this discussion anyway. Saying "others are on it" is like saying "some people have a Benz S500." It's irrelevant for most people. -- Orlando Andico +63.2.976.8659 | +63.920.903.0335 _________________________________________________ Philippine Linux Users' Group (PLUG) Mailing List http://lists.linux.org.ph/mailman/listinfo/plug Searchable Archives: http://archives.free.net.ph
