Feeds

back to article Oracle fudges touts Sparc SuperCluster prowess

Oracle has said that sales of its Sparc T series of servers are growing in the "double digits" in its most recent quarter, and that it expects this to continue through the remainder of its fiscal year. And while Oracle has not been precise about what is selling and what isn't selling, what is clear is that Oracle wants to peddle …

COMMENTS

This topic is closed for new posts.

Page:

Alert

Of course there is spare CPU cycles...

That is how Oracle rolls... make sure each CPU is highly under-utilized... because how will Larry ever be able to buy a new super-yacht if people onlu license software on the ACTUAL CPUs that are required?

When I read 20% utilized CPU, I see 80% waste and likely a machine that is imbalanced for the workloads... needing more CPUs to address internal system bandwidth issues...

The SPARC T4-4 is a competitive server... and if they took a high road and compared similar systems, they could argue the fine details to claim their's is better... but Oracle is constantly making marketing claims that prey on ignorance of the reader.

My job is to continue to call bull$h17 whenever I see it.

6
5
Anonymous Coward

Re: Of course there is spare CPU cycles...

Agree, it's Oracle. They do not care how many slap on the wrist fines they rack up for these fraudulent benchmarks. They are going to keep on making these absurd claims. For every five people that laugh at these benchmarks and make an Oracle joke, they know one person is going to accept it.

Oracle HAS to be cooking the books on their T sparc (doesn't deserve capitalization) numbers as well. They actually expect people to believe that legacy Sun has dropped from $1.2 billion to $500 million per quarter in three years, IDC has them losing 15% in Unix sales per quarter... yet T sparc is growing by double digits. Think about it. That is faster than Xeon and way faster Unix as a whole. I have a feeling that when pressed they will say something like "T4 sales are growing by double digits (from $0 last year)... we did not take previous T sparc sales into account."

4
4
Anonymous Coward

Re: Of course there is spare CPU cycles...

David14 --- "When I read 20% utilized CPU, I see 80% waste..."

When I see Java based enterprise applications, I normally see 100% utilization as lack of memory in the data center. Add memory, watch that CPU utilization dive down... garbage collection is often at fault.

When I see Database based applications, I normally see 100% utilization as lack of memory in the data center. Add memory, watch that CPU utilization dive down... not enough available shared memory.

David14 --- "needing more CPUs to address internal system bandwidth issues"

20% CPU needs more CPU's???

It is important for customers to understand how much headroom they have for an application. It is good to see there is processing headroom for expansion, backups, hardware failures, etc. This is a very real-world metric.

2
0
WTF?

Re: Of course there is spare CPU cycles...

Maybe there are just too many ignorant readers? If you read the press release and understood what it says, you'd see the following: http://www.oracle.com/us/corporate/press/1851992

"The new solution helps customers meet service level agreements (SLAs). A half-rack SPARC SuperCluster easily supports hourly throughput of approximately 200,000 Oracle Financials transactions with 2,000 concurrent users in the application tier domain with 80 percent additional capacity available for higher loads or further consolidation"

So * 20 per cent utilization in the application tiers of the cluster* just means theres a ton of headroom left to support even more transactions (during peak demand for example), x more users or run/consolidate other workloads to consume the remaining CPU cycles.

1
1
Devil

Re: Of course there is spare CPU cycles...

It's effing Oracle. What the hell do you expect?

0
0

SPARC, not Sparc !

While I no longer work for Sun or Oracle (long story), still it is spelled SPARC and not Sparc. Trademarks are supposed to be observed, even across the pond. Hmm. I guess I should have written "spelt".

Besides, SPARC means something specific. And Sparc just looks like some chic way of spelling Spark.

Whatever. But your legal department should be worried. (Does The Reg HAVE a legal department?)

You'll notice that even the previous commenter got it right. David +1, The Reg -100 (because it's been at least a hundred misuses by now, and probably my 100th email on the subject.)

SPARC rules! Just spell it right.

Cheers.

6
11
Meh

Re: SPARC, not Sparc !

Well, IANAL but I fail to see how El Reg's use can be considered trademark violations. They're clearly referring to the party that holds the SPARC‰® trademark. I would've thought observing trademarks just means that you cannot use the mark to refer to something else. I didn't think trademark law forced everybody to always spell everything correctly.

4
0
Anonymous Coward

Re: SPARC, not Sparc !

Legalities aside, I think the point was it's stupid to call it Sparc when it's not. It's SPARC. Just like the guy said, it's not a clever way of spelling "spark". (And if you're young and/or ignorant enough to not know that then maybe you should not be writing about the stuff...)

2
4
Silver badge
Happy

Re: SPARC, not Sparc !

Shirley, that should be SPANKed! A much more accurate description of both the lack of performance AND how Oracle DB customers feel?

2
2

Re: SPARC, not Sparc !

We all call it SPANC since every processor out there has been SPANCing it for about 10 years.

3
1

This post has been deleted by its author

Anonymous Coward

Re: SPARC, not Sparc !

> I didn't think trademark law forced everybody to always spell everything correctly.

Well, stop just 'thinking" and look it up. SPARC is a trademark, sparc is not. When referring to the processor chips from SPARC International the correct term is SPARC, anything else is a trademark violation.

0
5
Silver badge
Thumb Down

Re: SPARC, not Sparc !

I'd guess that it's nothing to do with legal concerns over trademarks, it's just a reporter that is too lazy or rushed to override a spell-checker which automatically "corrects" excessive capitalization. What that says about the level of care taken over the rest of the article is left to the reader...

4
2

Re: SPARC, not Sparc ! (@AC 14:03)

Well it so happens that I'm seriously retarded, so much that I just simply don't know where to look that up. Can you please provide some pointer? I hear you saying that using a name that is not a trademark ("Sparc" or "sparc") when referring to something trademarked ("SPARC") is a trademark violation? So calling the Sparc a "CPU" is wrong too? (Yeah, the lack of capitalization is entirely not on purpose, I just stopped thinking.)

3
0

Re: SPARC, not Sparc !

Would refering to it as 'piece of crap' be a trademark violation? I think that as long as you're not labeling an actual chunk of feces as a SPARC(tm)- ie. "Man, I feel better. Just laid down a nice big SPARC(tm) in the toilet.", you're in the clear.

2
1
Anonymous Coward

Sparc is fine Re: SPARC, not Sparc !

You want to know why Oracle spells it SPARC? Because long-ago, some dim-witted marketing people got the brilliant idea that if their product names were ALL CAPS, they would stand out in articles. Seriously. Pretty soon, vendors were CAPITALIZING EVERYTHING even if it wasn't an acronym and their PR people were kept busy sending off chiding notes to those who got it wrong. Most journalists took the attitude of "screw that, we're not here to function as an extension of your PR" and ignored or ridiculed the practice.

So you have a choice. You can act like a pedantic fourteen year-old who just discovered the Chicago Manual of Style, and accept that you're acting as an unpaid shill for one of the stupidest PR practices in our industry. Or you can act like an adult and just spell it like a normal word. Up to you.

3
1
Anonymous Coward

Re: Sparc is fine SPARC, not Sparc !

Oracle spells it SPARC because SPARC International, which owns the IP, tells them too, since the word is derived from Scalable Processor ARChitecture, and is legally protected as a trademark. Seriously, are the commenters here really too lazy to look beyond their own nose?

0
2
Anonymous Coward

XIV ?

IBM XIV Gen-3 with SSD Cache would have been the better IBM Storage array to put against this.

DS8K is way overkill for this, and no one would actually do it.

I work with hundreds of customers running AIX with XIV, Gen-2 and Gen-3.

2
1
Anonymous Coward

Re: XIV ?

Agree, IBM would recommend XIV Gen3 as the sub system in a real world situation. I would like to see an independent group take some high end T or M Sparc servers with whatever Oracle's best sub system is (probably Pillar, not even a joke) and compare it to Power 795 with XIV Gen3. Oracle would get destroyed in an Oracle DB benchmark.... Much less taking a Power 795 filled with SSD internal storage, which would be more apporpriate.

Oracle also claims that Power is IBM 'highest end system".... Ah, no, System z is their high end system. Oracle doesn't have a comparable system, but that is not IBM's fault. If they really want to compare highest end to highest end, obviously Oracle isn't concerned about real world configs, they should but their Exa license ploy up against IBM BlueGene/L supercomputer... see how that works out for them if they want to compare highest end to highest end.

2
1

Re: XIV ?

Regarding cpu performance on the IBM system Z, they are very bad. Any decent high end x86 cpu is roughly twice as fast. So cpu performance wise, the new Z12 is very bad.

POWER7 on the other hand is a very good cpu and it is much much much faster than the IBM z12 or Z196 cpu. So I would not call the System Z, the high end. An IBM P795 packs much more cpu punch than a dinosaur z12 or z196.

0
4
Anonymous Coward

Re: XIV ?

System z isn't very bad in terms of CPU. True, it is not a supercomputer or a Power 795, but it is not very bad either. System z is all about I/O, designed to run bank transactional systems and the like, not crunch human genomes.... Oracle EBS financials module, described in the Oracle ad, is I/O intensive, not CPU intensive... just a pile of Java. Oracle has nothing that can mess with System z - Linux in performance *or* cost with the high end M (if it still exists), T Sparc or Exa systems.

1
0

Re: XIV ?

"...System z isn't very bad in terms of CPU. True, it is not a supercomputer or a Power 795..."

Well, the Z12 is a very slow cpu. No doubt about it. Where are all the IBM benchmarks? Nowhere. Why? For a reason. Have you ever wondered why IBM never have released benchmarks of Mainframe cpus? There are none. Zipp. Zero. Zilch. Nada. Why? You tell me.

On the other hand, we have loads of POWER7 cpu benchmarks everywhere of all different kinds. Well, the POWER7 happens to be a really good cpu, so we understand when IBM wants to tell the world how fast the POWER7 is. But regarding the Mainframe cpu? Well, it is "the worlds fastest cpu", but no evidence nor proof of that. Just some marketing material:

http://www-03.ibm.com/press/us/en/pressrelease/32414.wss

At least IBM could have offered ONE benchmark vs the POWER7? But no. The Mainframe cpus all get obliterated by the POWER7.

Some would call extravagant claims without any proofs to be FUD. Which coincidentally, happens to be IBM's master dance:

http://en.wikipedia.org/wiki/Fear,_uncertainty_and_doubt#Definition

"FUD was first defined with its specific current meaning by Gene Amdahl the same year, 1975, after he left IBM to found his own company, Amdahl Corp.: "FUD is the fear, uncertainty, and doubt that IBM sales people instill in the minds of potential customers who might be considering Amdahl products."

0
5

Re: XIV ?

Actually there are quite a few Mainframe benchmarks out there. You just have to look for them.

Try in some of the same places as Oracle's T4 benchmarks.

// Jesper

0
0
Holmes

Oracle press releases full of FUD? Who would have thunk it...

Isn't it always like this? Instead of letting their achievements stand for themselves, they always have to sour the news with some disingenuous comparison against 2 year old hardware from a rival vendor. It's the Oracle way.

Another Oracle way is to claim all the glory for someone else's results:

http://www.oracle.com/us/corporate/press/1853058

This was a Cisco benchmark on Violin Memory storage, but Oracle glosses over the former and neglects to mention the latter entirely.

3
1

This post has been deleted by its author

Silver badge
Boffin

Superdome2 flash.

When we were asking for a demo unit earlier this year we were told flash with the SD2 had to be external. This wasn't a great surpise as the design philosophy with the old Superdome was all storage was external. With the hp x64 blades you have the options of either SSDs instead of hot-swap disks, or you can use mezzanine slots for plug in SSD cards. Seeing as the SD2 blades don't have any onboard disk slots we guessed the first option was a goner, but we did hope hp would get their mezz SSD option working with hp-ux so we could put them directly onto the blade (both SD2 and the Inetgrity blades), but we were coyly told that "was not available at launch". We also asked if hp might put PCI-e SSD accelerator cards into the SD2 IOX bays but we were told that was not on the roadmap. So the only supported option does appear to be external flash in something like a 3PAR, EVA or MSA 2000.

0
1
Silver badge
Meh

Vendor benchamrk? Meh!

Unless your business is doing exactly what Oracle did in their benchmark (and it almost definately isn't), then it would be a lot better to get a solution designed for you requirement and then tested in your environment with your data. Anything else should be considered a possible indicator of the best performance likley to be achieved, and definately not what you can expect to get in a real environment. Same goes for all vendor benchmark sessions.

2
1
Anonymous Coward

IBM do it too...

I saw an IBM benchmark comparing Power to x86 which had a much lower TCO for the Power solution using LPARs, primarily because the x86 solution wasn't allowed to use virtualisation.

When I was trying to evaluate ZFS vs Veritas Filesystem a few years ago, I could only find two comparison papers. One from Sun, which showed ZFS was faster than VxFS and one from Veritas/Symantec which showed that VxFS was faster/more efficient. Go figure.

Vendors stack their benchmarks in odd ways which favour their product.

0
0

Re: IBM do it too...

Here is a Oracle ZFS benchmarks vs NetApp:

http://www.theregister.co.uk/2012/04/20/oracle_zfs7420_specsfs2008/

Here are some more

https://blogs.oracle.com/si/entry/7420_spec_sfs_torches_netapp

0
1

Re: IBM do it too...

Another one. IBM XIV vs Oracle T4 ERP. "Without any optimizations Oracle was 2.5x faster".

https://blogs.oracle.com/si/entry/zfssa_smashes_ibm_xiv_while

0
2
Anonymous Coward

Re: IBM do it too...

That is using a partially populated Gen2 XIV against a very heavily cached ZFS array.

Gen-3 brings SSD cache to the mix, and a much faster and lower latency back end.

Gen-2 is over 4 years from GA date, and while it has had some minor improvements over those years, it just would not show well against a more cache heavy array.

0
0
Headmaster

Re: IBM do it too...

The key here is reading the small print.

It's a brand new Oracle system.. versus a POWER5 system.. which was GA 2004.......

IMHO this just shows how desperate Oracle is.. they need to find and pick a fight with solutions that are not just current generation -1 from their competitors.. but systems that are up to 8 years old.

// Jesper

3
0
WTF?

Re: IBM do it too...

Did you actually read (and understand) the fine print or just want to FUD Oracle regardless?

Heres the actual text from http://www.oracle.com/us/corporate/press/1851992

"Up to 56 percent lower acquisition cost versus new IBM POWER7 and HP Superdome2-based configurations(1)."

Footnotes

(1) Comparison based on current published list pricing of a half rack SPARC SuperCluster solution: one SPARC SuperCluster half-rack with eight 3.0 GHz SPARC T4 processors, eight channels, eight cores each with built in storage (36 × 600GB) and one Exadata Storage module versus IBM configuration example of the IBM POWER 780: one IBM Power 780 with eight 3.8 GHz POWER7 processors, 1 channel, 8 cores, and 512GB memory and IBM storage DS8700 (384 × 146GB FC, one base cabinet and one expansion cabinet) and HP Superdome2: two Superdome2 each with sixteen 1.73 GHz Itanium 9350 processors, two channels, eight cores, and 1TB memory with HP EVA8000 storage (182 × 300 GB FC, one base cabinet and one expansion cabinet). Actual deployments and configurations will vary.

And the second claim which is about *consolidation*, which in my dictionary, means older systems, usually 3-5 years old. So clearly, that means Power6 systems.

"Up to 40 percent lower five-year TCO when consolidating existing IBM and HP infrastructure, while reducing rackspace footprint up to 73 percent(2)."

0
1
Devil

Re: IBM do it too...

@Phil 4.

I was replying to Kebbart's link:

https://blogs.oracle.com/si/entry/zfssa_smashes_ibm_xiv_while

Try to read my comment having that link in mind.

The Registers reply function doesn't always make it clear what you are replying to.

// Jesper

0
0
Anonymous Coward

Re: IBM do it too...

If it is for Oracle DB, that is probably true. VMware is technically not supported for Oracle DB, only Oracle VM (Xen). PowerVM, AIX's hypervisor, is supported and allows you to hard cap licenses. If you run any Oracle on a VMware server, you license every core... and every core on the stand-by VMotion server. If you are taking Oracle DB licenses into account, Power is less costly than x86 - VMware if you hard cap.

Power - Linux (RHEL or SUSE) is not supported by Oracle, only AIX is supported, but the prices for Power - Linux are the same as x86 - Linux. They peg the Power - Linux prices to x86.

0
0
Anonymous Coward

Re: IBM do it too...

If Oracle say it, then it must be true?.... It is not though the entire point of the article you are commenting on is that Oracle benchmarks are pure fiction.

1
0

Re: IBM do it too...

@Jesper Frimann

"...I was replying to Kebbart's link:

https://blogs.oracle.com/si/entry/zfssa_smashes_ibm_xiv_while..."

It was never clear that Oracle compared vs an POWER5 system. It is no merit to be faster than an old generation system. I will never post that benchmark again, because it is unfair.

Thanx for informing me.

1
1

The wonders of statistics.

If I sell or do ten things one year, and twenty the next... I have experienced triple digit growth.

Looking at Quickbooks... This week I have earned over a thousand times what I made last week. (I sold an old printer to a friend last week, didn't work at all.)

0
0
Anonymous Coward

What you'd need...

... is some third party that sets a goal and gives each vendor a set time to come up with (and tune) a suitable configuration. Then the third party runs the benchmarks, we have a round of re-tuning, and the third party runs the benchmarks again. And it's the third party that publishes the benchmarks.

But what vendor would put money&effort into such a vehicle? Much more "interesting" to buy a sack of kit from the competitors and be as stupid as you can with them, then publish "test results" that are utterly and completely useless, but make your kit shine. Nothing like that false sheen of second hand car salesmen style marketeering.

2
0
FAIL

Oracle make stuff up? No! Really?

As a number of people here have already said, vendor benchmarks are always suspect and none more so than our friends at Oracle (Larry appears to think "talking $h1t" and "marketing" and inter changable concepts

I will point out that the company I work for, which may or may not be HP, has stopped doing benchmarks for exactly this reason - no one beleives what we say, and given that these configs can cost $1m+ to buy, build, configure, test and publish it seems somewhat insane to keep doing it.

I'd always recommend you get with your vendors technical boys and girls and get something built and tested for your specific environment with your data....and don't forget that HP, IBM et al have gurt big DCs set aside for customer tests if you just want a benchmark - they'll happily run one for you, using your data if you want it

2
0
Anonymous Coward

Re: Oracle make stuff up? No! Really?

I thought the reason HP had stopped publishing benchmarks was that the DB vendor has to sign the form as well as the HW vendor and Oracle refuse to sign anything to do with HP HW.

1
0

Re: Oracle make stuff up? No! Really?

It may come as a shock but there are other DBs out there :)

0
0
Silver badge
Facepalm

Re: Oracle make stuff up? No! Really?

"It may come as a shock but there are other DBs out there....." Here's the funny bit - up until the Oracle-hp case settled in hp's favour, we were getting from hp's salesgrunts a refreshing new diet of MS SQL and PostgreSQL fluffing. Now that Oracle has been tied to supporting hp-ux on Itanium the hp salesgrunts have largely switched back to their preferred "I luv Oracle" message!

0
4
Anonymous Coward

Comparable systems

Snoracle box 2TB of RAM,

HP box 1TB of RAM

IBM box 512GB

hmmmmmmmmmm, conventional TPC wisdom would say that the system with 4 times as much RAM should score 4 times the DB performance, if it doesn't, it sux.

4
1

Re: Comparable systems

It is not obvious that more RAM helps a certain benchmark. In this case, is the current benchmark RAM starved? If it is, then Oracle should not publish such benchmarks where the RAM differs. But if the RAM amount does not matter, above a certain threshold, then the RAM amount does not matter.

I remember a SAP benchmark where the Sun x86 machine had 256GB RAM but slow RAM, and the Linux x86 server had 128 GB RAM, but faster RAM - and still the Solaris server was slightly faster. It turned out that the Linux HP server used 128GB RAM, because the faster RAM sticks came in smaller size. So HP could choose to equip their Linux server with 256GB RAM or 128 GB RAM, and HP chose the faster RAM sticks. In this SAP benchmark, it did not matter if you had 128GB or 256GB - because had it mattered, then HP would have gone with 256GB RAM.

Thus, it is not obvious that this benchmark would have been helped with more RAM.

.

.

But I agree that this benchmark seems a bit fishy from Oracle. It needs to be investigated a bit further. For instance, is RAM crucial? Is it better to have 2TB RAM, or does it not matter? If it does not matter, then it doesnt matter and the criticism is only IBM fanboyism.

On the other hand, when IBM does fishy benchmarks, then the author Timothy Prickett does not write an article about it. Clearly IBM biased.

1
4
FAIL

timw@splhi.com

I think it's a pretty good bet that, unless the people doing the benchmark were incompetent idiots, they put in the appropriate amount memory in their system (since clearly that affects the price) and so, yes, to put in 1/2 or 1/4 in the "rival" systems makes the whole comparison extremely suspect.

Hey, my "inexpensive family care of choice" beats the "expensive sports car of choice" if I remove the spark plugs from all but one of the cylinders!

0
0
Anonymous Coward

Re: Comparable systems

> It is not obvious that more RAM helps a certain benchmark.

If the RAM wasn't helping the Oracle system then they wouldn't have put it in as its effect would just be to make there system more expensive.

But if you're an Oracle sufferer then try looking in a AWR report's Buffer Pool Advisory on the impact of increasing available cache space on the expected number of physical read operations. These the little buggers that normally limit the performance of database. The redo log responds well to write caching in the storage system.

1
0
Holmes

It's hilarious

I must admit I stand with TPM. What Oracle is claiming in pricing against HP and IBM is actually ... hilarious.

Now I have nothing against people saying.. I want to have a good comfortable chair, but I don't want to spend money on a quality crafted designer chair, that will last a lifetime, but I'll go get one from from a place like IKEA cause it's good enough for me.

But when they then go and buy a mass-produced industry standard chair from Snurakle that costs the same as the quality crafted designer chair, they need their heads examined.

Basically what they are doing is taking some ... well lets call it lowend midrange servers and storage, that have been put together into a clustered solution, which they have smacked a 'highend' sticker on.

A system that does not integrate with anything else. neither horizontal or vertical in the solution stack.

Furthermore it is just as proprietary and closed as an IBM 80'es mainframe.

And at last they put a serious price tag on it. A price tag of such a magnitude that it will only make remotely sense to clients if they (Oracle) try to compare it against REAL highend solutions from other vendors, and even HighEnd solutions that have been neutered and purposely made suboptimal solutions.

Again, if I were to compare a HP or IBM scale up solution (cause if you compare scale out versus scale out for the different vendors the machines are quite different) against a

For HP you are kind of stuck with the SD2, again if you were scale out I'd go with a HP itanium blade, but using 2 x SD2-16 is ridiculous, one is quite enough. And as others have stated then the storage system is.. well.. why do you need to compare a highend storage system (that isn't even sold anymore) , that is designed to have many many host attached to it, with a buildin proprietary system ?

For IBM again as it is with HP why go with a highend system ? And also the server chosen is a highend server, why use a highend POWER 780 when you really only need half a midrange POWER 770 ?

Now if you were to compare these realistic HP and IBM configurations against the Oracle solution, you'd see that both the HP and the IBM solutions have been cut by more than 56% of the price.

So what would you buy.. the lowend midrange eq. bundle from Oracle, or the high quality gear with all it's RAS from HP or IBM ?

// Jesper

3
2
Facepalm

Re: It's hilarious

Jester, Its clear that you’ve been drinking too much of that blue coolaid like TPM, but what I find most amusing that you don't seem to have any clue on what SPARC SuperCluster is all about, but yet you have so much to say against it. Probably a good idea to do some research first before making such uneducated responses.

First of all, SPARC SuperCluster is an engineered system with servers, storage, networking, flash, infiniband, etc that’s all integrated into one rack, all engineered to run any Oracle SW and the 1,000's of apps that support Solaris faster. It cant compare directly to high end SMP systems like SD2, M-Series or Power systems without adding all the other components into the mix. The closest comparison would be IBM's Puresystems which includes a mixture of 2-CPU or 4-CPU nodes, networking and storage. However, Puresystems does nothing about faster performance, just faster delivery times.

Theres nothing more proprietary than running IBM Power which locks you into running AIX or IBM mainframe or running HP-UX which locks you into running Itanium, a dead architecture.

A single SPARC T4-4 (there are up to 4 of them in a SPARC SuperCluster) outperforms Power 750's, Power 770's and even the turbo-core Power 780 on several public benchmarks running DB's, Java and Weblogic environments, and a SPARC T4-4 outperformed a SD2-16 by quite a high margin, so comparing a SPARC SuperCluster which includes 4 x SPARC T4's to a SD2-16 or single Power 780 is really whats hilarious!.

A SPARC SuperCluster with 4 x SPARC T4's, which has a total of 4TB of RAM, is a fraction of the cost of a fully loaded Power770 or Power780 which doesn’t include any additional storage or networking for that matter!

I dare you or anyone else reading this post to do a side by side price comparison of a SPARC SuperCluster versus anything equivalent (from a like for like spec and performance perspective) and prove me wrong - no one offers a better price/performance solution than SPARC SuperCluster for running Oracle SW and even SAP for that matter!

1
4
Big Brother

Re: It's hilarious

@Phil 4

So you now refute to name calling.. That is nice and mature, but kind of shows your true colours.

I am fully aware of what Oracle SPARC SuperCluster is made up of. I get bombarded with Business Partner information all the time, and invitations and presentations and and and...

And where as I really Really really like the Oracle Database, and I have a profound respect for the SUN systems of old days, where I (and also on hp9000) as an young computer science major learned the craft of *NIXing, I have no problem calling a pig with lipstick for what it is.. a pig.

As for benchmarks, then basically the only industry benchmarks where the T4's have really gone against other systems are TPC-H and SPECjEnterprise2010.

Now TPC-H is there cause Oracle cracked that benchmark some years ago, and even the OLD M9000 systems were able to beat other systems that they normally wouldn't have a chance against in other benchmarks.

SPECjEnterprise2010, cause that it what the T-processors were really designed for with their many threads and hardware accelerators, and they also threw a lot of 'auxiliary' hardware at the benchmark also, as price isn't a factor. But the accelerators is really what makes the machines shine, cause they do that, in this benchmark. Just to bad for Oracle that POWER7+ will now have the same type of accelerators, so lets see what happens then.

But that is basically it, there aren't really any other datapoints, besides the prelaunch comparisions to the T3, and they really don't support your claims.

Now as for price.. then sure a comparison is a good idea.

What you don't seem to get Phil, is the magicians move that oracle is doing.. waving one hand to catch your attention while putting the other one deep into your pocket. Oracle will give you the the damn hardware for free, if you buy their bloody software.

Now the list price of a brand spanking new POWER 770 with POWERVM and AIX in enterprise edition Hypervisor mirroring, Memory compression bla bla. will set you back 2MUSD in list price including 3 years worth of HW maintenance etc etc.

Now a full rack SuperCluser will set you back ~1.8-1.9MUSD, with 3 years of support.

Now surely the POWER 770 does not include storage.. which it should.. so .. lets.. nahh ... lets just look at the software.cost, we can always put in a few SDD disks later. (they are real cheap)

Now the SuperCluster needs some software licenses, if you will run Oracle on it.

You'll need Oracle EE, Oracle RAC, Oracle Partitioning, with 3 years of swma, for the SuperCluster and only Oracle EE for the POWER 770, if you need to run the full Oracle monty.

Now that is 8.711.680 USD for the Oracle Supercluster versus 5.046.400 USD for the POWER machine

Which basically gives you 2.5 Million USD to buy some storage for the POWER 770. And surely I could go with the 48 core 6 cores per chip POWER 770 rather than the 64 core version, cutting away 25% of the Oracle prices.. while only loosing around 10% in throughput and no memory.

If you compare the RAS features on the machines, you'll find on the T4-4 specs looking like this:

Hot-pluggable disk drives

Redundant, hot-swappable power supplies and fans

Environmental monitoring

Extended ECC, error correction and parity checking memor

Easy component replacement

Integrated disk controller with RAID 0 and 1

Electronic prognostics

Compared to the POWER 770 (or the SD2 for that matter) It's pathetic.

A machine like the POWER 770 will do:

Redundant hotswap powersupply, fans and adapters.

Concurrent firmware updates.

Processor instruction retry

Alternate processor retry of instructions.

Live Partition Mobility and Live Application Mobility.

Redundant service processor and system clock with hot failover of both.

Hot node repair.

Memory/processor sparing.

Mirrored Hypervisor memory.

Again.. the T4-4 won't even hot-swap an adapter .. you have to take that whole server down. You basically need a n+1 configuration that you won't need with the POWER 770 or a SD2 or a M9000.

So what would normally be something that is a standard simple change, all of a sudden becomes a more complex issue with the Supercluster.

Again that is what all the Infiniband and storage nodes and and is all about, addressing the shortcomings of a cluster with added hardware components and complexity. And downtime and service windows.. cost a LOT of money.. again.. TCO.

Furthermore

A machine like the POWER 770 will do memory compression, out of the box, for the Supercluster it's 11.5KUSD per license + 22% in yearly swma. Which will set you back 1.2 MUSD on 4 node cluster machine. Sure DB compression is better than compression in HW or on OS level... but 1.2 MUSD ? brrrr..

A machine like the POWER 770 has shared pool processor virtualization (so does the SD2), where as on a T4-4 server you'll have to turn to both Zones and LDoms for some of the same functionality. Which basically means that it'll run with a good deal better utilization in a RL virtualized environment, as you will basically be forced to do partitioning in a multi tenant environment, on a T4.

I always get amazed when what is basically v-par/LPAR in HP/IBM terms is touted as virtualization.

A machine like the POWER 770 also does infiniband, actually the x12 adapter plugs directly into a GX++ mezanine bus. No need to fight with other adapters over the bandwidth to the processor. And with this being a single system, there isn't even the need for an infiniband switch to get involved.. it can just be a tube directly to a sh*tload of SDD drives.

And what is better than about 12TB of SDD storage directly in an IO drawer and then virtualized through the Hypervisor, for maximum IO speed ? Wroom Wroom.

You do know the SPC-1 benchmark where an old POWER6 machine(that is 2007 technology with PCI-X ) beats out a pair of Oracle Sun ZFS Storage 7420c Appliance, by a factor of 2+.. with half the latency...and ?

It's like.. why setup a compex storage setup with infiniband and storage servers and bits and bolts that break and needs firmware updates and and and and when you can achieve the same in software inside a single high RAS machine ?

Then there is the Datacenter cost picture. A T4-4 uses a massive 2.7 KW when 100% active compared to the 4.5KW of the POWER 770 (both number using Oracle and IBM power/heat calculators), that is double the heat density for the T4, so much for coolthreads. Furthermore the max operating temperature of a Supercluster is 32C versus 35C for the POWER 770. This basically means that you need to be able to handle a heat density that is almost twice as dense with at T4 than you need with a POWER 770. And a SuperCLuster rack uses 15KW max power.

It does require a rather modern datacenter, which.. costs a lot of money.. Again .. tco...tco...

Well.. I could keep going.. ... but.. I'll let Matt handle the SD2 vs SuperCluster, he's most likely better skilled to do that than I am.

... and try to cut down on the Snoracle Cool aid Phil :)=

// Jesper

3
1

Page:

This topic is closed for new posts.