Seismic time
Earthquakes are tectonic events. Was this an attempt at irony?
Intel, the maker of the Itanium family of server processors, and Hewlett-Packard, the main customer for the chips, have fired back at an Oracle announcement late yesterday that it was stopping software development on the Itanium chip. The PR folks at Intel rousted Paul Otellini, Intel's president and chief executive officer, …
I have always been amazed that DB purchasers are willing to deal with 'spoiled brat' ellison. He is always throwing his toys out the pram and spitting his dummy out. Now he has his mate with him, who single handedly managed to alienate and suck out any morale that was left in the HP workforce when He gave himself a $40M bonus and either cut or tried to cut every employees salary between 5% and 25%. He is to leadership what butter is to personal armour.
Hopefully the sun will start setting on Oracle (no pun intended). I for one do my best not to purchase anything from them - I always look for alternatives and typically find something better and/or cheaper. Hopefully more purchase decision makers will start looking for alternatives too.
"Oracle continues to show a pattern of anti-customer behavior as they move to shore up their failing Sun server business," Dave Donatelli, general manager of HP's Enterprise Servers, Storage, and Networking group, said in his statement."
Oracle of course have reacted and can confirm for questioning, that they are able to offer a fabulous in-house range of novel customer-centric solutions to make brilliant use of revitalised and reconfigured [born again and debugged] Sunny business servers ..... for quite Titanic Quarters into Big Business Recycling ...... Power Elite Great Game Plays for Commanders and Controllers of CHAOSystems Operating Advanced Hosting Clouds for Provision of Facilities and Utilities Supplying Virtual Reality Remote Instruction Set Configuration.
Or is that A.N.Other Party altogether, sailing into Safe Harbour Ports Astutely Championing Anonymous Legions in Advanced Intelligent Stealth Programs, Projects and Protocols?
"If HP had any sense, it would buy a software company that can emulate Oracle databases. How much do you think EnterpriseDB is worth?" ..... It is worth practically nothing if it can't do at least that, methinks. Buy a company that can capture and improve upon metadatabase structures/virtually intelligent architectures, and the EnterpriseMDB worth sky rockets out of orbit and any normal reach and into Deep CyberSpace and COSMIC Quantum Intelligence Fields of Need to Know you may not be presently equipped, and therefore enabled to know.
However, that is easily remedied with a comprehensive simple instruction pack delivering every answer to all pertinent, and even impertinent, questions.
Is not even having to run `make install clean` or `aptitude install`ever again. And you are threatening them that they may have to maintain their own software?
Documentation, replication, pretty chart makers, roadmaps, promises (however vague) and handholding is what many customers are only too happy to pay for. Oracle has good products and excellent sales people who are good are giving customers the warm and fuzzy feeling needed to open corporate wallets.
Much as I like Postgres and it is nice to see Enterprise DB and others picking up business from smaller installs, I don't think Oracle really gives a shit. SaaS is, however, something that does threaten the very core of their licensing business.
Agreed completely. We really like the Oracle DB product even though we know there are alternatives. Problem is, we also really like hp-ux and Integrity because they also deliver easy use and that sense of comfort. At the moment, this is like having your two best friends falling out, all you hope for is that they can patch things up so you don't have to start avoiding the one of them.
Don't confuse Intel's commitment to HP-UX with its commitment to to Intainium...obviously they are trying to get HP to port HP-UX to Xeon or do an Aries emulator again.
"We remain firmly committed to delivering a competitive, multi-generational roadmap for HP-UX and other operating system customers that run the Itanium architecture."
Allison
Doesn't IBM stuff scale up to mainframe monsters?
I mean yes, there is a need for asm, hand tweaked code but not to the point of losing the entire compatibility of the product with future CPU arch. World's largest server CPU vendor, who has been always known for perfect roadmaps say "I am keeping and enhancing this CPU". Even Apple changed an entire architecture making it easy/possible for hackintosh (their worst nightmare) just because Intel is a vendor who doesn't speak sh*t when it comes to roadmaps.
I know Oracle is World's second largest software vendor so it really amazes me how they can be managed that way.
I gave the ./configure example just to show power of opensource against these spoiled monopolies. Not like postgres etc. are exactly garage code maintained by some hippies (!) in their free time :)
"If HP had any sense, it would buy a software company that can emulate Oracle databases."
Like this one headed by Foster Hinshaw, Netezza's founder?
http://www.dataupia.com
Looks like large-scale Oracle support is in there:
http://www.dataupia.com/pr20100223_largest_oracle_olap_db.pdf
OK, I'll be the dummy that asks the obvious question... because I genuinely don't know and I suspect a lot of others don't either
What development does Oracle have to do for a particular hardware platform? I thought the OS handled all that? So Oracle for Windows/Linux/Solaris on x86 (or HP-UX/OpenVMS on Itanium) would be very different, but Oracle for Windows on x86/Itanium or Linux on x86/Itanium/dead badger should be pretty much the same.
If you can't resist flaming the dummy, go ahead. But I'd appreciate a (link to) a simple explanation
Although operating systems promise some kind of hardware independent heaven for application developers the truth is that this is rarely the case. Sometimes you can just recompile and run but often you have to tweak the code a lot to get any real benefit.
The whole reason behind different flavours of hardware is horses for courses and for CPUs this means aligning the software algorithms with the CPU microcode in the compile step as the goal is to get your software running as natively (MMX, Altivec and friends) as possible on the chip with as few calls to the application for rescheduling, run on the integer, fpu and memory allocation. This requires having good tools in the chain (compiler) and good software engineers who know how to tweak the right code the right way for the right hardware.
In a theoretical world, you'd be 100% correct.
For most cross-platform RDBMS's I've worked with (MySQL, Sybase, etc.) you'd be maybe 95% right.
But what is often said about emacs applies more so to Oracle, IMHO: "OK RDBMS ('editor' for emacs), but a really terrific O/S!" People pay Larry the big bucks for scalability, which is good, because frankly, that's all Oracle offers. All other offerings are cheaper, nearly all are easier to administer, develop against, etc. If you can keep Windows running underneath even MS offers a more feature rich database product these days, but if you want to scale to millions of TPS with your data safely on replicated disks after each commit, you're probably looking at Oracle.
To achieve that level of scalability, you end up with an installer that actually links your binaries together on the target host.
It is more discriminating about hardware, O/S, etc. than any cross-platform database engines I've ever worked with in the last 20 years or so.
It costs Oracle to maintain development on any platform. The thrust of Oracle's announcement is that, unless HPUX is transitioned over to x64, the Itanium cut-off will curtail HPUX new-release work.
I just wonder if there is some reason why HP is having so much trouble with its traditional 'friends' on technology (RH, MSFT, CSCO, and now ORCL). It seems unlikely that ORCL will get a windfall of new SPARC business from this...
Allison, I think you need to learn english!
"We remain firmly committed to delivering a competitive, multi-generational roadmap for HP-UX and other operating system customers that run the Itanium architecture."
The clue is the fact they mention Itanium in the quote. If they just said HP-UX then maybe you had a point. I read that quote as they are willing to keep the Itanium development going for as long as there were operating systems that wanted to use it.
As the worlds most high tech ink seller HP just sells servers as a loss leader. As Matt B says its not like anybody is going to go the garbage architecture that is SPARC. All this will due is move more people to x86_64. Nobody buys itanic except legacy VMS/VAX Alpha shops anyway so this much to do about nothing.
With Xeon scaling up to 2048 cores, in a single Linux instance, then Itanium is looking increasingly pointless. Scale-up x86 also helps avoid the RAC premium that Oracle would like all x86 customers to pay.
And you can even run 256 cores of Windows if you're feeling brave :)