Typos corrected
"disappointed with the NAD’s decision in this matter, which it believes is unduly broad and will severely limit the ability to lie in advertising, not only for Oracle but for others in the commercial hardware and software industry”
Oracle has pulled an attack advert that claimed its Exadata system is twenty times faster than IBM Power systems - because it isn't, and an advertising watchdog wasn't impressed. The ad compared an IBM Power system database performance at a European retailer with the same databases running on Exadata. The retailer found they …
Oracle is like a redheaded stepchild, who proudly says "look at my pretty dolly", to which the rest of the world is saying "you ripped the head off your dolly, and we don't think that was nice". I am glad IBM came around and shoved them back in the closet. Now if we could just get somebody to lock the door. Maybe that pervy old uncle, Sam, will play with them for a while and make them cry, just a little.....
Really, it says there "up to 20x faster queries". The people in charge of the moneybog DO KNOW that this MEANS "we ultimately managed to find a query, that, if optimized in a very special way (compiler options include at least the options -abcdefgh), performed 15 times faster than a similar query on a nonoptimized software-hardware combination of the competitor which had about half the CPUs and half the RAM of our machine and which ran generic softwar"
Right?
since the customer is buying new kit, that probably means that the comparison is with previous-generation IBM kit.
On the other other hand, the big boys (IBM included) have been doing "case studies" based on competitors' old kit against their own latest-and-greatest for at least the 25 years that I've worked in this business.
Case studies using old kit vs new kit are done by most companies in some way, personally I tend to ignore these unless its against their own products (even then different configs can be used to boost certain stats of newer machines, so again if it looks like they are deliberately making old kit look rubbish I close).
I completely agree. That's kinda Oracles point though. They optimize the stack so you don't have to. There's also some special sauce in there, but the beauty is in the simplicity and consistency. Of course, if it were just better integration then HP and IBM could actually be competitive with Exadata. So far, from the testimonials (read customers), Exadata is blowing everyone away on real customer data. You'd think that IBM would come up with a competitive engineered system if it were just integration of components. It's an engineered system right? IBM has engineers I assume?
if ibm play there cards right they'll win this easyly...
have been involved in a benchmark a decade ago where informix was faster on a 12 cpu sun solaris machine with 12 GB of memory then obstacle on a 64 cpu sun solaris machine with 64 GB of memory.
i guess they bet on the fact that IBM does not put informix against them.... if IBM does put informix in there
this whole thing will blow up on obstacle...........
it would certainly make my whole year ok....
Superboer
The way I interpreted the ad, it seemed a trivial upgrade if the old system was more than a few years old... As long as Moore's Law holds out, we sort of expect every new system to have a hardware advantage.
Having said that, I better ad the disclaimers that I'm in the IBM food chain and I think all ads are extremely dubious these years. What I'd like to see is a progressive tax on idle cash that the corporations are sitting on.
There is no comparison in overall quality, scalability and reliability.
Oracle's Exadata system is an Intel-based cluster of servers, so the database scales horizontally which will work well for some workloads... it also is exceptionally complicated to manage and tune compared to a "regular" database, and the hardware is good for one thing only... the databases.
IBM Power servers are general-purpose computing environments that can scale from 4-core machines up to 256-core machines... and the level of virtualization and the robustness of the environment is second to none!
The kicker for many companies is that they get caught up on the sticker price of the hardware, rather than the big-picture, total cost management view of things which will include software, downtime, risk, management, longevity, etc.
And, for the record, I sell and architect solutions based on Oracle, IBM, and HP hardware... but the winner in the enterprise compute environment is a clear one in my mind and experience.
GS doesn't install Power systems... they implement Oracle and SAP, run IT operations, etc. The VARs implement Power systems, same as ever other provider. Power systems are less costly than Itanium and Sparc. Power with Linux pretty close to the cost of x86 - Linux and it is an integrated system as compared to a bunch of components.
IBM systems were used by basically every country in the world in those days. They also were the systems used to manage US war production and defeat the Nazis. You can't blame the tool maker for how their tools are used. I am sure some computer company supplied systems in the Sudan as well, but I would blame the people using those systems... not whoever happened to make them.
Exadata is fast at queries...that involve lots of where statements, as long as you don't need to do a lot writes, as long as your front and back end aren't too busy with columnar compression CPU time, as long as your database isn't that big, as long as you have all the patches with data loss and unavability bugs loaded, as long as the software raid doesn't have two hard drive failures.
And then you get the HVAC bill. And then the Oracle maintenance bill.
Interestingly, right when Oracle hit the market full press with the second generation hardware, there was a bug in ASM with red hat where it would use the /dev instead of the multipath devices. Strange timing, performance limiting bug at that time.