A waste of time

For six years there has been no entries to the cluster version of the TPC-C benchmark. And for a very good reason. There is no interest in these systems at all. Anyone can combine a large number of servers and produce any result they wish. But these systems have very marginal use cases and do not provide the flexibility customers are looking for.

Exadata machines are just of no practical value. Like a wedding ring they only symbolize the current crush Ellison has for a particular hardware vendor. Exadata v1 was supposed to be the killer offspring of the Oracle/HP partnership. That lasted a full year. Luckily only a handful of customers were left with the abortion called Exadata1. I guess the market will have that in mind when Oracle now announces their next creation of hurriedly scrapped together SUN parts.

And now IBM has entered this silly contest.

@Matt

Matt, they can try... now, if it plays out, that's a different story and yet to be seen.

Re:Victor 2

You wrote:

"The single traffic cop is one massive point of failure. On Oracle's Exadata, any component on any node, or even a whole node (be it an Exadata or a Database server) can fail and the system will continue running as if nothing happened"

You described RAC as being able to continue runnning witout any disruption with a node failure. That is simply not true. And neither would any DB2 on POWER product. Hence my comment.

"Big deal, so has Sun... It doesn't mean THIS particular case of creating a DB cluster had to be done using Infiniband... why not 10Gb?... Let's see what IBM does to address the flash accelerator part, but why use Infiniband if you are going to use standard and SLOW hard drives?"

I didn't say that SUN hadn't used Infiniband for years, only that your comment about IBM copying SUN by using infiniband is a not true.

And you don't get it. The external IO drawer is connected to the server via Infiniband. And one main reason to use Infiniband on POWER is that you have the option of plugging it into the GX bus, which means lower latency and higher throughput.

"Oh, oh... so, you say using flash is flawed, yet IBM will come up with something flashy too?.. contradiction alert..."

Now, seriously, who told you you had to replace flash memory every month?... remember also, that Sun/Oracle are NOT using SSD drives, they are arrays of memory modules. Ellison emphatically stated so.

If you do ALOT of write IO to the same block, you will very quickly burn through a flash memory chip. This is the same for all.

"I'll grant you that, yet, how does it compare to Exadata v2 grand total of 15TB of data when compressed?"

Better AFAIK, compressions isn't really something I normally use.

"SPARC roadmap has hardly changed over time, ROCK was set to be a parallel architecture to that of CMT and Fujitsu's SPARC.. new scheduled projects can get canceled, that's why they are called "projects"."

Call it what you want."

I call it p***** on it's customers. If you hype a product that much, yes it was hyped way way to much, and then just cancel it.. you must expect that customers (like where I work) gets p*ssed.

Just look at these few links:

http://news.cnet.com/Sun-burnishes-next-gen-Sparc-chips/2100-1006_3-5561693.html?

http://blogs.sun.com/jonathan/entry/rock_arrived

http://www.theregister.co.uk/2009/01/28/rock_sparcs_on_track/

http://www.physorg.com/news97329357.html

http://sun.systemnews.com/articles/107/3/news/17539

http://www.infoworld.com/t/hardware/suns-rock-rolls-further-along-898

http://sun.systemnews.com/articles/111/1/news/18046

http://sun.systemnews.com/articles/114/4/news/18516

http://research.sun.com/spotlight/2007/2007-08-13_transactional_memory.html

....

It's one of the most hyped products in recent years, and it turned out to be just hot air.

"SPARC is alive and well, thanks for asking."

Yes it is alive but barely. Don't get me wrong I WANT SPARC to be there, we need competition and good SPARC products. But on the APL line nothing is happening. And to be honest fujitsu isn't really moving toward being more of a server vendor, than they are today. They would rather buy the PC part of the joint Fujitsu/Siemens vendor, than buy SUN. That basically leaves us with the TX which is terrible for workloads that cannot be parralized, in the extreme.

One of my wife's friends who is a very competent Oracle DBA, she quit her job after they went from APL to T5XXX'es where she worked. Simply cause the DBA's had to take the blame for the bad performance, massive increases in install times.

And that is what Larry don't understand when he praises Niagara, he is going to p*** off much of his loyal base which is the Oracle DBA community, when they get forced to rethink much of what they today consider best practice.

"Then, don't use for other things than what it was intended too... You can't buy a toaster and expect it to serve you coffee.

Oh, and Power isn't all that cheap either."

Jup, and I am not using it for anything that requires just decent single threaded throughput.

Power boxes are expensive, and you should always batter your IBM sales person into giving you some decent discounts. But the savings in software licenses, and the pure utilization that you can get these boxes to run with. Means that it is actually a good deal IMHO.

"No, look closely... why is that "virtualization" so similar to solaris 10 "virtualization"?"

Ehh... Solaris 10 virtualization is containers, and yes AIX 6.1 had WLM extended a little bit to make Workload Partitions, that had more functionality than containers, but I would hardly call that virtualization, rather workload management and user and process isolation.

Now when an AIX admin talks about virtualization (s)he talks about what the power hypervisor can do for you. And here SUN is no match.

"So, you want to use Oracle and not pay for it?... execution compatibility is not al there is to Oracle products."

No, but I don't want to pay more than I have to, I don't want to pay for 24 licenses if I can design a solution that can run on 8 licenses, with the same throughput.

It gives me no value NO VALUE what so ever, to pay for more licenses. And if it comes down to the buy cheap hardware with expensive software, I'd take the better hardwarewhich saves on the license cost any day. Good hardware gives you value.

Note that this also goes for any software, not just oracle.

"Leave that part to the people that knows."

Yes, I am. Myself and I'm good at it.

// Jesper

It's all about balance

In my experience with both IBM,SUN and HP. It's all about balance.

On Sun I get great parallel throughput but to be honest crap single threaded performance, on IBM I get great single threaded performance and often bad parallel performance especially if I have a number of heavy single threaded jobs running.

HP often falls in between, with the best and worst of both worlds.

Now solaris on x86, has in most cases given me the best bang per buck, good single threaded performance and pretty good parallel performance, specially with 6core (AMD) or multi thread (intel).

In specialist cases (i.e. OLAP) I generally go for and IBM solution or for web type transactional loads I've gone for Sun

Sun+ParAccel Exadata-killer: Where did it go?

Last year, perennial TPC under-achiever Oracle sticks a low-tech query filter and gobs of DRAM in front of the disk-pool (IBM circa 1990's, EMC circa 1980's), adds Infiniband (IBM circa 2002), and clusters it up into something dubbed "Exadata"...and pronounces it "revolutionary".

Meanwhile, Exadata performance on all but the simplest data warehouse schema turns out to be only marginally better than Oracle's (famously poor) data-warehousing performance WITHOUT Exadata. Meanwhile, Oracle's "bizarre" TPC configuration tweaks have been noticed...but not explained...

http://www.dba-oracle.com/news_2009_hp_exadata_tpc_h_benchmark.htm

Now this year, Oracle switches from (hp)Intel to (sun)Intel servers, adds some Flash, and once again it's "revolutionary"! To prove it, Oracle promises a world-beating TPC benchmark to show how magical these Oracle/Sun synergies really will be.

But strangely and for no apparent reason, Sun's ALREADY world-beating June 2009 TPC result on X4540 platform and ParAccel Analytic Database has just got "disappeared" from the TPC website. Sun's 30TB result was not only the world's fastest, it was also the world's largest, and (among large systems) the world's best cost/performance -- 1/16th the cost of #2. The Sun/ParAccel results vanished just the week before Oracle's premature mediajaculation about a "world's best" TPC on Exadata2, leaving (get this...) an Oracle/HP solution as the lone result in that category.

Ostensibly the Sun result was "withdrawn" on protest of an anonymous TPC member, even though Sun's Full Disclosure report went through two revs, and neither Sun nor ParAccel has ever had a problem with their numerous previous TPC results.

Maybe the amazing Sun-ParAccel result -- over 1,000,000 QpH on a 30TB warehouse at only $2.86 per QpH -- was "withdrawn" because it would have made the upcoming Exadata2 results look sick. Consider: Oracle's previous result on a 30TB warehouse was 150,000 QpH at $47 per QpH. That's only 15% of the performance of Sun/ParAccel, and at 16x higher cost. Furthermore...even with the advantage of a much smaller 1TB warehouse -- and with the further advantage of having 2TB of DRAM to hold it all, the best that Exadata1 could manage was $5.42 per-query-per-hour.

So...the Sun/ParAccel "world's best" result vanishes just before the "new world's best" result is supposed to come out on Sun/Oracle's Exadata2. Anybody smell fish?

If EU regulators REALLY want to know if Oracle's buying Sun is anti-competitive, they should find out what happened to the Sun/ParAccel solution.

When/if the Oracle/Sun TPC results do come out, the important number will be COST/performance, not absolute performance. Any major vendor today can throw enough "brute-force" (aka hardware) at a problem to overcome even the crappiest software. Good architecture on the other hand is demonstrated via the best COST/performance and performance/energy ratios. By this measure, the Exadata1 been a total failure.

I guess we'll wait for the numbers on Exadata2, and then we'll wonder how it might perform against the competition that got disappeared.

(for anyone interested in digging in to the story, the disappeared Sun results have been de-linked at the TPC website, but are still available here):

Original:

http://www.tpc.org/results/FDR/tpch/sunfire_x4540_paraccel_tpc-h_30tb.fdr.pdf

Revised:

http://www.tpc.org/results/FDR/tpch/sunfire_x4540_paraccel_tpc-h_30tb_V2-090721.fdr.pdf

(for anyong interested in discounting ParAccel vs. Oracle based on it's columnar approach, check here first):

http://database.cs.brown.edu/sigmod09/benchmarks-sigmod09.pdf