Itanic
Remember when the Reg used to call this chip the "Itanic" every time it was mentioned, and it got funnier every time? I swear, I'm still laughing at that.
HP has told a US court that it should force Oracle to keep supporting its Itanium-based servers for as long as HP sells them. The firm's lawyer said in closing arguments in the case that Oracle was contractually obliged to support Itanium chips based on the Hurd agreement. Oracle stood by its position that the Hurd agreement …
I'm not saying an e-mail contract isn't legal. I'm saying: Who would realisticly negotiate & sign a $4bn contract just via a few e-mails? I'd expect a shed-load of lawyers to be involved with rain-forest amonts of paper work for such a contract.
*Unless you're Mark Zuckerberg, of course. (Which neither Oracle or HP are...)
implying any sort of contract. HP specifically asked for wording about Itanium in the Hurd press release and Oracle specifically told them no in the e-mail exchange. You can say that this is anti-competitive, anti-customer, or whatever on Oracle's behalf, but it is pretty impossible to see this as violating any sort of contract.
Even suppose they win on the contract issue there is nothing stopping Oracle from forking the Db feature set so that Itanic (I love that name too) gets sloppy seconds.
And even if Oracle don't do that there could be plenty of queued bug fixes and performance tweaks that just sit in the request queue.
Rightly or wrongly HP Itanic is staring down the gun barrel on this one.
If I were an Oracle customer with relatively new Itanium hardware, I would be quite unhappy with Oracle and not much comforted by the prospect of my software maintenance being done under compulsion of a court. I might even be open to other options like Postgresql which offer almost all the functionality for almost none of the licensing cost. If I were a large organization like, say the U. S. Department of Defense, I might realize that the cost of a reasonably sized support staff and its training might be far less than Oracle's $15K/CPU license fee and $3K/CPU annual maintenance cost.
If I were Hewlett-Packard, I might think that providing services and support for other-than-Oracle DMBS could be helpful to my loyal customers.
If I were Oracle, I might think about such things but, narcissist that I am, probably would not.
Well that's an easy one, there aren't any recent Itanium customers (have you seen HP's sales figures?)
As much as everyone loves to hate Oracle, they didn't start the abandonment of Itanium, anyone remember how long ago Microsoft dumped Itanium over two years ago?
"....If I were Hewlett-Packard, I might think that providing services and support for other-than-Oracle DMBS...." Just about weekly I get new whitepapers sent through to me from the hp reps, extolling the virtues of PostgreSQL, Sybase and some nastiness called DB2, and how easy it is to extract Oracle from my stack. Truth is we're unlikley to do anything until after the court case.
"Please. I love PostgreSQL too, but it's *NOT* an Oracle replacement!"
Yes, EnterpriseDB is all over the place trying to push Postgre as an Oracle replacement. There are all sorts of technical issues with Postgre vs. Oracle, but the most important shortcoming is that Postgre is not certified or supported for any major commercial applications (SAP, Oracle apps, IBM software, vertical banking apps, obviously not MS, etc). Primarily because their access controls and security features are not enterprise grade. If anyone is migrating off of Oracle, it will be to DB2... which is more or less on par with Oracle and way less costly.
Oracle needs to win this one after their big, big public embarassment in front of Google and the courts.
Perhaps Oracle is legally right and they are able to support or desupport Itanium whenever they please, but like the Google lawsuit, this one is going to harm them big time, regardless of the result.
Scenario one: Oracle wins and they desupport Itanium. Picture IT responsible getting visits from Oracle sales guys telling that their lovely and supported Oracle RDBMS is available, if only they shell out a few thousand bucks for Oracle (formerly Sun) hardware.
Scenario two: Oracle losses and they are forced to support Itanium. Picture IT responsible getting visits from Oracle sales guys telling him/her how Oracle is commited to support their business... as long as a judge forces them to do so.
I don't see happy customers either way, and what I see from this and the Google lawsuit is a company that is agressively trying to squeeze every cent of their Sun acquisition, with total disregard for customers. That strategy could pay for the next boat racer but is not a long term one certainly.
"I don't see happy customers either way,"
Considering the:
1) disappointing performance
2) unjustifiable lack of SQL syntax features (e.g. no cross-table updates without painfully slow sub-selects, usually nested ones, braindead explain syntax with output that doesn't even provide index details and row count estimates, etc.)
3) lack of diagnostic functionality such as logging all queries (possible but the performance overhead is massive, output huge and cluttered and needs to be post-processed to be useful, whereas with MySQL it is pretty much free and gives just the useful part of the output)
[...]
I would say that Oracle have no happy customers in the first place. It was only after I had to use Oracle that I found out just what a shining beacon of SQL database excellence MySQL actually is.
@Wensleydale Cheese
Indeed, but MySQL is GPL-ed and there are forks available should Oracle start to not play ball. So far so good.
PostgreSQL is an _awesome_ database that I use quite extensively, better in many ways than MySQL, but MySQL is just more ubiqutous and more importantly, generally regarded as being crap. This is why I was so disappointed with Oracle when I found it doesn't even measure up to MySQL.
The article has a significant fact-error. Oracle continues to update existing releases and fix problems on Itanium. Oracle WON'T provide its new releases on Itanium, though.
If it didn't do the former, it would have to relinquish maintenance monies for its Itanium base, immediately. Oracle has accountants that advised against THAT.
Yes, Oracle is going to support the current Itanium install base on their current software because they want to wean people off of Itanium... but they are going to do a really poor job of it. They pulled everyone working on Oracle - Itanium development off of Itanium, so it you need a specific patch developed or some other non phone level support, best of luck with that.
This is a typical loose loose situation law suit.
If Oracle loose in court they loose face and have to support a competitor.
If Oracle win in court they just piss off a massive number of highly lucrative customers. Oracle on HP-UX was worth 3 times as much to them as Oracle on Solaris + the Sun HW business, according to other articles on here.
The only winner will be IBM, who must be laughing all the way to the bank.
The lawyers will no doubt win too, I guess.
I really dont understand why you think the customer would be more or less pissed off if Oracle wins?
The customer have had Oracle's message that they wont code new versions on Itanium for about year. I dont see why HP loosing the court case would make said customers more upset with Oracle, i can see why they wouldnt like HP... Oracle winning would just prove that HP didnt have a case to bring the lawsuit agaisnt them.
Perhaps I am missing something...
Not defending Larry as he is one of the world's biggest tools. Just saying comparing Itanium to the dead for years SPARC to make it look relevant is like bringing the fat guy from Borat on a date with you to make you look classy. Both architectures live on out of hubris only and even that will pass soon enough.
Windows isn't supported on Itanic isn't, most (all?) Linux vendors dumped support for it too. What else are customers going to run on it? That crappy OS called HP-UX?! HP ditched their own decent UNIX: Tru64. You know, the one with TruCluster, AdvFS, etc. They had to go begging to Veritas/Symantec of all people to add those to HP-UX for them!
HP has never impressed me but DEC products sure did. Their sales people and executives did a good job (with draconian mandatory maintenance contracts) of keeping their awesome stuff out of the hands of small companies and riff raff like me but Alpha compared to its contemporary peers is still the most revolutionary CPU of my life time. Its a shame to see HP and its Dell like mindset come in shit on that legacy with the sad sorry joke that is Itanium. Itanium doesn't even deserve to be Alpha's fluffer.
Don't blame HP for killing the Alpha. Blame DEC themselves. The have completely failed to promote and sell it and Alpha was very, very sick already even before Compaq bought DEC. Yes Alpha was great and insanely ahead of its time, but DEC has failed everything but its design.
I've been in a 100% DEC shop (mostly VMS, some Ultrix) for many years and they still are the best years of my (long now) career. VMS' tightly integrated clustering is unmatched by any *nix until nowadays as far as I know. Ultrix was OK, a true-blue BSD I liked. Spent several years in the board of our local DECUS chapter. Nostalgy, nostalgy...
Do you actually work on HP-UX or are you basing your judgment on its state in the days of Tru64?
HP-UX 11iv3 actually is a very decent production-level Unix. I have worked on Solaris (a lot), AiX (quite some time), Tru64 (a little bit) as well as many variants of Linux and *BSD and why there are bits of HP-UX I strongly dislike such as the reboots still needed for way too many product updates, I have to say that it has generally become a mature and extremely stable O/S with everything I need: LVM/VxFS, ServiceGuard for clustering etc. On Itanium, its level of support for HP-PA legacy apps is awesome, with excellent perfs and the ability to import whole HP-PA systems as simili-VMs running at close to native speeds.
I do a lot of HP-bashing here so no fanboism.
IBM will be the primary beneficiary as they will have the only viable Unix system after this takes its course. Intel gets to ditch Itanium which is great for them and this will likely drive many people to Xeon - Linux, of those that don't go to IBM - AIX. Oracle also wins. They will probably get a couple of Exas out of this for people that are gluttons for punishment. They will also win on all of those people who go to x86 - Linux. Those that formerly used Serviceguard who will now be using RAC and will now need to buy a bunch of extra Oracle core factors in the VMware soft partition licensing model. Larry knows what he is doing.
This post has been deleted by its author
What is the relative market share positions of Serviceguard on Linux - x86 as compared to Itanium systems? I don't know, but probably something on the lines of 80% on Itanium and less than 5% on x86. They could also use RAC on Itanium systems, but the market is entirely in Oracle's favor on the x86 side and entirely in HP's favor on the Itanium side.