Saturday, June 04, 2011

HP-UX: What now? / E agora?

This article is written in English and Portuguese
Este artigo está escrito em Inglês e Português

English version:

You know I usually stick to Informix related topics, but a few weeks ago something was announced that has to make us think... As we all know by now, Oracle announced that future versions of their software products will not support HP-UX running on Intel Itanium chips.
Why do we have to think about it? Well, from my personal perspective because it raises a lot of serious doubts/questions. Let's see:

  1. First you read in the announcement that Intel is not truly committed to Itanium. The exact words are "Intel management made it clear that their strategic focus is on their x86 microprocessor and that Itanium was nearing the end of its life". Intel denied this in the following days. Who do we believe? The company that owns the chip in question or another (now also) hardware supplier? Is it normal that a company discontinues it's products on a platform that has not announced it's end of life plans?

  2. The announcement was made just before an important HP shareholders meeting... We could of course believe this was just a coincidence

  3. In the last months a lot of news came up referring to personal wars between Oracle's and HP's executives (ex HP CEO is now a very important Oracle executive). Meanwhile ex SAP CEO has been appointed as HP CEO, and Oracle and SAP had an ongoing court trial started at the time he was SAP's CEO. All these an point 2) may lead us to the feeling that while personalities clash, customers suffer. There's nothing wrong with having strong personalities as leaders of large companies, on the contrary. But we would expect their egos not to cause harm to their customers (of course "harm" translates into costs, uncertainty, fear and doubts - typically FUD that competitors like to spread, but in this case competitors don't have to bother...)

  4. Oracle mentions that others (Microsoft and Red Hat) have previously discontinued support for Itanium. This is true, but there is a big difference: The market share that these two companies had in the Itanium market was considered irrelevant. Now, for Oracle, the situation is completely different. They have a large share of Itanium customers (accordingly to this blog, there are around 140.000 Oracle/Itanium customers)

  5. Some people argue that supporting a platform has it's costs. Of course this is true, but the costs are certainly well covered if the above number holds true. That's part of the business. In order to provide a product, you incur in certain costs. And also note that Oracle always said, wrote an publicized that their code base was the same independently of the platform. So using this as an argument is hardly acceptable...

  6. Many people think this is just a commercial move from Oracle. They're trying to weak a competitor position while at the same time they hope to raise their own hardware sales. We can accept that this is normal, but we should keep in mind a few things. First, just a while ago Oracle choose HP as an hardware partner (remember Exadata?). Secondly Oracle CEO has said that it would like to see Oracle becoming the IBM of the sixties (integrated stack where the margins would be bigger), but that it would keep its software running on competitor platforms. Well something has changed :)

  7. Will this be the only case, or will Oracle do the same in the future with other OS? The fact is that the number of OS and Hardware vendors for enterprise computing is shrinking. You currently have Microsof Windows on Intel x86, IBM AIX on System p, HP-UX on Itanium, Solaris Sparc, Linux on Intel x86, and Solaris on Intel (I'm explicitley forgeting other platforms like IBM System Z and HP NonStop). Oracle is doing it's best to eliminate HP-UX/Itanium. Will it stop there? Or will it proceed it's path to become "the IBM of the 1960s" (meaning the closed system that locked customers in)? Note that IBM still suffers with the image it created at the time.

The above are just a list of some important points. I may be missing a few. The announcement must had a significant impact on HP Itanium customers running Oracle software. Imagine that you wake up one morning and find out that the software vendor you choose gave up supporting your platform. Of course you'll have support for your existing products, but you'd really appreciate a roadmap... By the way, wasn't that the same company that 10 years ago accused IBM of not having a roadmap for Informix? It's funny when we put things into perspective...
So, assuming you're one of those customers what will you do? You have a few options:
  1. You jump on the Oracle train and buy a one way ticket... I mean you choose Sparc for your upcoming hardware renovation... You don't know where the train will take you... You don't even know how much it will cost you... Specially because you just bought a one way ticket... Once you're "there" you'll figure what what price the next ride will cost... One thing you know: You'll be traveling with the same company since no other operates in the same region...

  2. You choose another hardware platform, and you really hope the same trick will not be played again

  3. You change your software supplier
I'd say none of the options above looks particularly attractive. But in case you need another database (and you're able to get your application running against it), you should really consider Informix. Here's why:
  1. It's robust, easy to use, reliable, works well in virtualized environments etc., but you should already know that

  2. It has a roadmap and has just completed a decade of improvements after the IBM acquisition

  3. It belongs to a company that will try to sell you it's hardware, because it believes it's good, and not because it tends to be the only option to run it's software

  4. It's already very well integrated with many of the other IBM software portfolio, and this is assumed to be a continuous effort

  5. It has a long history of working well with HP-UX (traditionally on PA-RISC and now with Itanium). A search for "informix" in the HP site will show you several HP documents about integration between Informix and HP-UX
To wrap up this article, I'd like to put here a few links that relate to this topic. I hope they'll allow you to see what's being written about this Oracle announcement, and to form your own opinion about it.




Versão Portuguesa:

Normalmente restrinjo-me a assuntos exclusivamente relacionados com Informix, mas há algumas semanas atrás foi anunciado algo que nos tem de fazer pensar.... Como já todos deveremos saber nesta altura, a Oracle anunciou que futuras versões do seu software não irão suportar HP-UX a correr em chips Itanium.
Porque é que devemos reflectir sobre isto? Bom, na minha opinião pessoal porque isto levanta uma série de dúvidas e questões importantes. Vejamos:

  1. Começamos por ler no anúncio que a Intel não está verdadeiramente empenhada no Itanium. As palavras exactas (tradução pessoal) foram: "... a gestão da Intel deixou claro que o foco da sua estratégia é a linha de processadores x86 e que o Itanium se está a aproximar do fim de vida..." . A Intel negou isto nos dias seguintes. Em quem acreditamos? Na empresa que detém o processador em questão ou noutra empresa (agora também) fornecedora de hardware? Será normal que uma empresa anuncie o fim de desenvolvimento dos seus produtos numa plataforma cujo fim de vida não foi sequer anunciado?

  2. O anúncio foi feito imediatamente antes de um encontro de accionistas da HP.... Podemos claro acreditar que isto foi apenas uma coincidência

  3. Nos últimos meses vieram a público uma série de notícias referentes a guerras pessoais entre executivos da HP e Oracle (o ex CEO da HP é agora um quadro importante na Oracle - o número dois na verdade, logo abaixo do Larry Ellison). Entretanto o ex CEO da SAP foi nomeado CEO da HP, e a Oracle e HP tinham um processo em tribunal que remonta ao tempo em que o mesmo era CEO da SAP. Tudo isto e o ponto 2) podem levar-nos a pensar que enquanto as personalidades se chocam os clientes sofrem. Não há nada de errado em que grandes empresas tenham personalidades fortes na sua liderança, bem pelo contrário. Mas seria de esperar que os egos não prejudiquem os respectivos clientes. O "prejuízo" traduz-se em custos, incerteza, medos e dúvidas - o que em Inglês se chama "FUD - fear, uncertainty and doubt -", que normalmente é espalhado pela concorrência, mas que neste caso nem requer esforço da concorrência pois é feiro pelos próprios.

  4. A Oracle referiu que outros (Microsoft e Red Hat) já tinham previamente descontinuado o suporte para Itanium. Isto é verdade, mas há uma enorme diferença: A quota de mercado que estas duas empresas tinham em Itanium não é comparável à da Oracle. Esta tem uma grande percentagem dos clientes Itanium a usarem os seus produtos (de acordo com este blog, existem cerca de 140.000 clientes Oracle/Itanium)

  5. Algumas pessoas defendem que suportar uma plataforma tem os seus custos. Isto é uma verdade óbvia., mas esses custos são largamente cobertos se os números acima forem correctos. Isso faz parte do negócio. Para fornecer um produto as empresas (de qualquer tipo) incorrem em custos. Note-se ainda que a Oracle sempre disse, escreveu e publicitou que o seu código era o mesmo independente da plataforma. Por tudo isto, o argumento do custo de suportar uma plataforma não me parece aceiável...

  6. Muitas pessoas acreditam que isto é apenas uma manobra comercial da Oracle. Estão a tentar enfraquecer um concorrente ao mesmo tempo que tentam aumentar as suas p´roprias vendas de hardware. Podemos encarar isto como algo relativamente normal, mas devemos manter em mente uma série de factos. Primeiro, apenas há algum tempo atrás, a Oracle escolheu a HP como o seu parceiro de hardware (lembram-se do Exadata?). Segundo, o CEO da oracle disse que gostava de ver (tradução pessoal) A Oracle tornar-se a IBM dos anos sessenta (soluções integradas onde as margens são maiores), mas que manteria o seu software a correr nas plataformas da concorrência. Parece que algo mudou entretanto... :)

  7. Será este um caso único, ou irá a Oracle fazer os mesmo com outras plataformas? A verdade é que as plataformas (SO e hardware) para computação empresarial estão a diminuir. Actualmente temos Microsof Windows em Intel x86, IBM AIX em System p, HP-UX em Itanium, Solaris Sparc, Linux em Intel x86, e Solaris em Intel (estou a omitir outras plataformas como IBM System Z e HP NonStop) . A Oracle está a fazer o seu melhor para eliminar HP-UX em Itanium. Irá parar por aí? Ou irá prosseguir os seu caminho para se tornar a "IBM dos anos sessenta" (neste caso um sistema fechado que prende os clientes)? Note-se que a IBM ainda sofre com a imagem criada nessa altura.

Acima está uma lista de alguns pontos importantes. Posso ter esquecido alguns. O anúncio deve ter tido um impacto significativo nos clientes HP Itanium que utilizam software Oracle. Imagine que acorda uma manhã e descobre que o fornecedor de software que seleccionou, deixou de suportar a sua plataforma. Claro que terá suporte para os produtos que já existem, mas certamente apreciaria a existência de um roadmap... Aliás, não foi esta a mesma empresa que há 10 anos atrás acusou a IBM de não ter um roadmap para Informix? É engraçado quando se colocam as coisas em perspectiva...
Assim, assumindo que é um desses clientes, o que irá fazer? A meu ver tem algumas opções:
  1. Apanha o compboio da Oracle e compra um bilhete de ida... Ou seja, escolhe SPARC para a sua próxima renovação de hardware... Não sabe para onde o comboio o leva... Nem sequer sabe quanto lhe vai custar... Especialmente porque comprará apenas um bilhete de ida... Depois de "lá" chegar logo verá qual o preço da próxima "viagem"... Uma coisa será certa: Irá viajar com a mesma empresa, pois mais ninguém opera na mesma "região"...

  2. Escolhe outra plataforma de hardware e espera ardentemente que o mesmo truque não seja empregue novamente

  3. Muda de fornecedore de softtware
Diria que nenhuma das opções acima parece particularmente atractiva. Mas caso necessite de uma nova base de dados (e possa colocar a sua aplicação a correr nela), deveria considerar o Informix. Eis porquê:
  1. É robusto, confiável, corre bem em ambientes virtualizadoes etc.., mas isto já deverá saber

  2. Tem um roadmap e acabou de completar 10 anos de inovação após a aquisição pela IBMI

  3. Pertence a uma empresa que tentará vender-lhe o seu hardware porque acredita que é bom, e não apenas porque tende a ser a única plataforma onde pode correr o seu hardware

  4. Já está bastante bem integrado com muito do software IBM, e isto é assumidamente um esforço contínuo

  5. Tem uma longa história de bom desempenho em HP-UX (tradicionalmente em PA-RISC e actualmente em Itanium). Uma pesquisa por "informix" no site da HP irá mostrar-lhe vários documentos da HP sobre a integração entre Informix e HP-UX
Para fechar este artigo, gostaria de deixar alguns links relacionados com este assunto. Espero que lhe permitam ver o que tem sido escrito sobre este anúncio da Oracle, e que possa formar a sua p´ropria opinião sobre o tema.

13 comments:

distas said...

Interesting point. Can't aggree on the conclusion for, mainly, two reasons, though.

For one, hardware upgrades are not that "big of a deal" - it has to be done every once in a few years anyway, so there's nothing wrong with switching from HP-UX on Itanium to Unix(Solaris)/Linux on whatever other platform; there are plenty.

For two, - resources. "Sad but true" - there is a kvazillion resources which come up for SQL Server and Oracle from just googling practically anything in regards to either of these RDBMSs, and... almost none for Informix. It seems like the product's been completely abandoned or, otherwise, left behind very profoundly. Well, apparently it is quite a story even to get free ODBC Informix drivers for Windows 7 64-bit - there are simply none out there. There are some third-party ones, but that's another story. You would think there must be any RDBMS-provider ones. Or, if there still are, then why is it such an arduous task to find any?

Too bad, because even from reading bits and pieces of info I managed to find about Informix, it seems like some really good db, with great performance among other advantages.

As both Oracle and later SQL Server DBA - never came across Informix before in 10 years of my experience anywhere, i.e. lots of Oracle, lots of SQL Server, quite a few MySQLs, some DB2 here but... no Informix. So, if the popularity and widespread is not there, who would want to switch to Informix rather than switch just to other hardware, which will have to be upgraded (read changed) anyway? On my strong opinion, Informix, as for now goes exclusively as upgrade for any existing projects, any IBM products db backend or any new projects which are ran by people who come from Informix side. Can't really think of anyone planning to migrate to Informix all over a sudden, especially from such main players as Oracle and SQL Server.

There is also a third reason. It's very much subjective yet seems pretty accurate as soon as many agree on that: IBM support is probably the worst and most definitely one of the worst out there. Dealing with IBM is totally pain in the neck. Even judging by my personal experience with Cognos - before IBM bought it, Cognos Support was awesome, now it's as bad as it can get. IBM for many and many years has been consistently proving itself as highly bureaucratic, awkward and sluggish company. Ending up with IBM being the major, and kind of a single capable resource out there to address any Informix issues, is the factor to reject the very thought of going with Informix.

Fernando Nunes said...

Hi,
I appreciate your comments. Naturally I do not agree with them and that is of course a subjective position. Nevertheless I think some points you make need to be clarified because either I dind't understand you or they're wrong. Forgive me if it's the firts. Let's see:

- Yes. Hardware upgrades tend to be easier than software upgrades. My point was, do you want to keep doing business with a company that discontinues products simply because it doesn't want to compete? Assuming, you do, there is no point in continuing. And of course, you may think this is a legitimate commercial move. I find it hard to accept. But I'm biased of course.

- Having lots of links on a google search doesn't mean you get good information. Just a few days ago while working on a project with another IBM product, I needed info regarding Oracle database. The most useful, and the one I used came directly from Oracle site. And yes, I got a ton of results on google...

- Your observations regarding the ODBC Informix drivers puzzles me. Do you mean "free as in beer" or "free as in speech"? It you're looking for beer, follow the downloads link on this site, or google for "informix odbc driver download site:ibm.com". The first non-sponsored link will get you there. Sincerely, I think I didn't understand you. Is this "hard to find"?

- As for the IBM support quality, I must say I'm biased (again). I work with a lot of people regarding Informix and other IBM products. Most cases I open regarding Informix are though. And I get very good results. It may look presumptuous but I admit I'm not the average Informix "client". On the other hand my quality standards are very high. I have a very clear position regarding software support in general and it's not good. I base this opinion mostly on my experience working and discussing with customers. I hardly see any who is completely satisfied, and I rarely see one that really demands quality.

In any case, if you have further comments feel free to contact me directly (which doesn't mean I won't accept comments here of course)
I took most of your comments as your personal opinions, but couldn't leave some aspects without clarification since they could look as "facts".

Regards!

BeGooden said...

Dear distas,

I will make it short, because Fernando developped very well:

It is true that it has been difficult in the last 10 years, the Informix roadmap has been kind of unclear for the public world, or even inexistant. Some scavengers really took benefit of this situation.

Nonetheless, who has been navigating in the Informix culture at this time has seen a fully dedicated R&D, with a fast and continuous evolution of the product. This evolution is now speeding up because IBM has understood that they have the best ORDBMS on the market.

I agree that practically no advertizing has been made by IBM on Informix for I guess internal reasons, and this has unfortunately left field for others.

Now talk to an Informix user and you will be amazed how he loves this product, and how he will never change to another unless forced to doing this by hierachy.

Regarding the IBM support, did you have an personal experience with the IBM Informix support? I guess not. I did and I can testify that the Tech Support members know and understand the product in depth and are extremely efficient in their assignments. I would not say this about some other DB vendors...

And last, talking about support, just compare how many dbas a company needs to manage an Oracle infrastructure against an Informix infrastructure... Or compare how many years an Informix instance stays up against how many months ( days) for others...

Make your own mind based on your experience with Informix, not on what you have heard or not heard about it. Just try IDS 11.70 and really compare it.

Well, not so short at the end.

Eric, not an IBM employee :-)

distas said...

Fernando,

My point was, do you want to keep doing business with a company that discontinues products simply because it doesn't want to compete?

So, you seriously call it “doesn’t want to compete”. Really?? Let me ask you this: compete where exactly? In some minor market segment which wasn’t supposed to be there at the first place? In a technological segment ever declining from the very beginning it came out? An elephant doesn’t feel the dart pierce, neither it can be full with just couple of bananas, kind of rotten ones too. It’s doesn’t hurt Oracle because there is really nothing to compete for, it’s not even the matter of some competition, but the way HP and IBM try to make it appear, which comes to nothing else but long-running personal dislike of ex-SAP CEO which left to HP some time ago and now tries to bite Oracle. Well, let the guy brake his teeth trying.

HP and IBM, as well as some of their kind followers throw stones at Oracle for moving away from Itanium as a dirty measure, Oracle’s attempt to propagate SPARCs. Well, they just couldn’t possibly come up with more silly and irrational blame. Is it only two platforms on the market, either HP-UX or Sun SPARC and that’s it? Nothing else, right? So, indeed, if Oracle moves away from Itanium, its clients will simply have NO OTHER CHOICE BUT TO SWITCH TO SPARCS. Seriously?! You got to be kidding me. This entire thing is so ridiculous, I don’t even want to bother commenting on this… Simply because I feel kind of stupid elaborating on “2 x 2 = 4”. Let this be my response: any IBM- or HP-86 architecture. There is quite a choice of platforms from both HP and IBM – the ones who is bashing Oracle right now. So, what kind of Oracle’s “dirty games” are we talking about? None, really. At least in this situation. Itanium is just another unsuccessful technology, doomed to disappear, and Oracle just took the very logical and very expectable step in this direction, simply because there are plenty of other more successful alternatives. As simple as this, the rest is just political hysteria of the ones who feel bad of losing their unsuccessful product.


As for “legitimate” or not “legitimate”, it is totally legit if a company doesn’t want to compete in one or another market segment – there is nothing even slightly wrong with it! It’s business, not a sport! Open any business text for the first year Business Admin undergrad degree, - the very first paragraph will tell you any business is about earning money for the owners and shareholders, if there are any. So, I’m not sure what’s your point here. Once again, business is not supposed to play by some commonly established in a general society “rules” – that’s nonsense!

(to be continued)

distas said...

Let me put it in a very simple way in just a single sentence: as far as any business doesn’t brake lows by intentionally cheating and tossing by possible means it’s auditing and stakeholders’ reports (Wait, isn’t this exactly what that Informix ex-CEO did? Oh, yeah, right – that’s exactly the case!), and, of course, is not engage in some “hard-core” criminal activity, this business is doing complete alright. It’s not about “want to compete” vs “doesn’t want to compete”, it’s all about moves. So, yes – it is completely, totally and absolutely a legitimate commercial move. So, yes, there is no any point of continuing.

The main point is rather still here: Informix is much less popular, has much less presence in the market BY ALL MEANS, which makes it considerably intense to work with, as RDBMS, than working with mainstream systems, such as Oracle, SQL Server, and… even MySQL. Yes, even MySQL! BTW, you can propagate the situation with Informix vs Oracle/SQL Server with as MySQL vs Postgre and … can’t even recall now what couple of other distributed under GNU ones are. Well, I hope you got the point. I think it’s pretty explicit and there is really nothing more I could possibly add to this. May be only this.

As for IBM Support quality, once again, your word is just as good as mine. And mine is just as good as yours. I had an awful, I’d rather say disgusting experience, because this is what I exactly ended up with – I got disgusted from it, with IBM Support. I wish it never bought Cognos. Never.

And again to the bottom line of your argument. Informix is not popular. It has much less knowledge base than any other database. Your experience of finding things around from the corporate portals, like you mentioned in your example, is not even close to my experience. I do find, say 90 out of 100 useful stuff from googling it. But once again, it’s subjective. As much as my equally bad experience with IBM knowledge base, which I find awfully organized.

distas said...

BeGooden,

By some scavengers you must be meaning Oracle, of course. And may be Microsoft too, I’m not sure. Well, your attitude towards this situation puzzles me even more than Fernando’s view on how an “honest” business is supposed to operate… Well, let’s imagine for a second the following situation. You’ve been driving in a car on a multi-lane highway with very few other cars driving more slowly and, in their majority on the rightmost lanes. However, there is one car which has been going all along side with you, challenging you, racing you, getting on your nerves. All over a sudden you see it started to go sides, smoke brakes from under the hood and this “rival” car slows down considerably, and, say, pools over. What would you do in this situation? Interesting and somewhat challenging question, isn’t it?

Well, in the general society, as to me, I would expect anyone, including you and myself to pool over, back up and ask the driver whether (s)he is alright and needs any help. But then again, it’s just me. You would probably agree that there are very many people who, in this situation, would just go not even glancing aside and neither even thinking of doing what I just mentioned. And… well, you really couldn’t blame them for this. If you think carefully and be honest enough to yourself, could you really say: “what a bastard, didn’t stop!” You can’t, correct? Why? Because in the end of a day, this is their right. There are not obligated to do this, so you can’t judge them and, further, blame them for this.

Propagating this situation on business world, makes your point simply not applicable as such. There weren’t any scavengers because it’s simply not the matter of case. Period. Seriously, what do you expect, Informix started to go down… BTW, could you remind me why exactly? Oh, I know – dirty gimmicks of its ex-CEO, isn’t it? And I very well know about the team of 12, including the head of the team, who leaved to Oracle. But here again – they were OO and those advanced data types, including special, developers. Well, that’s definitely some funky stuff and definitely did add some competitive advantage to Oracle, but… not a huge one for the obvious reasons – development involving and utilizing OO and advanced data types/structures still makes up very minor part of today’s systems because the industries where these technologies are applicable at are rather few: vehicle control of whatever kind in logistics/transportation, industrial and other civil engineering construction (bridges, pipes, power lines, etc.), aerospace, etc. What I’m saying here – while there are some, rather few places which utilize Oracle’s features which it might have initially gained from “scavenging Informix” (I’m glad you didn’t use the word “raiding” – thank you!), it’s a tiny portion of entire Oracle’s market, the rest of which is relational. Just another reason for “scavenging” not to be applicable here at all. It simply is not to the point.

(to be continued)

distas said...

And last, talking about support, just compare how many dbas a company needs to manage an Oracle infrastructure against an Informix infrastructure... Or compare how many years an Informix instance stays up against how many months ( days) for others...
I would love to compare, but… There is no Informix out there, and the only one I know is supported by IBM team, so… And I don’t believe that supporting Oracle after 8i, which is three generations ago version, going back to mid-90s, very beginning of 2Ks (11g most current, 10g, 9i, 8i) takes any more resource than supporting Informix. And regarding the instance uptime, you must have confused it with SQL Server, may be, and too, a 2K version of it. Quite on contrary, claiming Informix to be all that automated and stuff makes me think it rather lacks many features other databases have. Simply because RDBMS is a very complex system to start with. Any highly complex system can, and will eventually, have many things which might go wrong, or, otherwise, will require certain tweaking. I’ll stress this: tweaking, not fixing. Therefore, if I hear that there is some “miracle DB” which just “does it all by itself”, to me that means only and only that it’s simply far less sophisticated than others, which (OMG!!!) require some maintenance.

Make your own mind based on your experience with Informix, not on what you have heard or not heard about it. Just try IDS 11.70 and really compare it.
Well, thank you kindly for the offer, but I think I’m going to pass on that.

Regards,

Stan
Neither Oracle, nor Microsoft employee

distas said...

Hi Fernando,

I very well see where you go about Oracle’s competition on a hardware market – you are somewhat right yet completely wrong in the end of a day. How is that?

Oracle has worked hard since its very formation back to 1977 to make the product available on as whole variety of platforms as possible. I don’t think they’ve been struggling so hard for almost three decades in order to start killing the portability now. As a matter of fact, that’s exactly Oracle’s portability, the fact that it works on literally any platform, may be except MAC (yet I heard about cases when people did it, just for fun, and it worked!), is one of its biggest competitive advantages. Probably the second right after Oracle RDBMS itself. Why would Oracle plan to kill a chicken which lays the golden eggs?? In fact, why would any company do that? As a huge mistake – yes, that’s a possible hypothetical scenario, but I’m still pretty sure Oracle is not taking a suicide route.

Our positions regarding the matter are completely adverse. While you believe Oracle has discontinued Itanuim support in order to kill it and thus pushing 140K users towards buying Oracle’s hardware, what I think is Oracle just made a proactive move which, probably, it should have done quite a number of years ago because for one – Itanuim was a looser technology from its very beginning, for two – it is going down! Not because some company like Oracle is hunting it down (I, honestly, just can’t stop grinning here) but because Itanuim’s weak nature. Yes, I understand, it works, it has many thousands, let it be even many hundreds of thousands devoted followers and yet… It’s the doomed, week technology which is over soon.

It’s very easy to exaggerate and speculate about Oracle ending Itanuim’s support presenting this completely logical and common sense move as some ominous doing, but there is nothing to that except, like I said, completely logical, expectable and, actually, good couple of years late move. Besides, Oracle is not forcing anyone to buy SPARCs from it, go with any other non-Itanuim platform which Oracle doesn’t own – please, be my guest. And yes, I don’t think Oracle will be after any other successful-from-the-beginning architecture because that’s exactly what might weaken them to almost death, compared with how they are doing now. Something very much similar to what had happed to Informix years ago. Like I already said, I’m pretty confident Oracle will go on this suicide mission.

distas said...

As for our, apparently, very different experience with Oracle vs IBM support… Well, this is what it is. I could, of course, go on the web and start browsing “vendor rates reports” and what not, but that won’t tell me much. Oracle is quite a complicated, if not to say a very complicated system, which requires lots and lots of time, effort, brain and patience to be put towards mastering it. But that’s exactly why it’s such an ingenious product, which gives you literally limitless options and opportunities to whatever you want with it with very minimum workarounds, if any at all. I mean the limit is there, but that’s strictly your knowledge of the system which is the limit, not the system itself. Therefore, you really have to know what you’re doing when calling Oracle Support. Yes, I mean it – YOU HAVE TO KNOW ORACLE (DB) WELL ENOUGH BEFORE DEALING WITH ORACLE SUPPORT. And if you don’t, but do expect Oracle Support to do all the job for you – it’s not gonna happen. And most probably, that’s exactly the reason Oracle Support might be “down there” in whatever independent user reviews. But to me it’s completely irrelevant. Because I come from the other, opposite side. Guys in Oracle Support not once and not twice actually helped me not to just solve the problem, but gave me advices and hints which made things work even better than I would even expect.

So, reviews and ratings in this particular situations are barely of any good. Consequently, that wasn’t the case with my very bad IBM Support experience when it came to Cognos – it wasn’t because I didn’t know Cognos well enough and couldn’t understand what was going on when being told things. Instead, it’s a hugely hierarchical, stall, highly bureaucratical system of running things IBM has. It took me a week or so of time and some ridiculous number of phone calls, with some multiple “service request numbers” to just get some lousy key to access that part of Cognos “IBM-now” knowledge portal. Which, by the way, was much, much better organized when Cognos belonged to itself. Have you seen “Lock, Stock and Two Smoking Barrels”, the first one? In the last minutes of the movie, there is a scene where US customs officer asks the guy, I guess, a standard question: “Anything to declare?” And the guy goes: “Yes, never go to England!” The same here, just not England and not “go to”, but never deal with IBM, if only you are put in the position being totally required to do so. But if you have a choice… So, here you go. And I really do not care about any of those reviews, simply because they can’t make anything different FOR ME, can they?

distas said...

Regarding the matter of various RDBMSs’ popularity. Oracle’s popularity talks for itself. Unlike SQL Server, which has become very well spread too for all over these years, especially since 2005 version came out, Oracle became what it is today for completely different reasons. I would even say those reasons were opposite.

SQL Server has ever been “cheap and easy”, plus much more aggressive marketing, which is rather obvious – where would Microsoft be without its ever competitive not so much the products’ quality but prices and marketing… As for the “cheap”, it’s obvious, and “easy” links to anything from relatively easy to use (developers can manage this RDBMS with just a minimum training), training time span is much, much lower than Oracle would require to get to the same level of expertise, GUIs are great, which adds to it because the management likes pictures and all other visual cool stuff, which, once again, Microsoft is unbeatable with, and to whatever else you could put under that “easy” wing. Well, the weakest point of SQL Server, as for now, is that while it can do pretty much the same as Oracle does on the large scale in the majority of situations, there might arise a particular situation when SS hits the wall and there you have to start dancing with workarounds, which, in turn, more often than not is doable but is a complete mess the same time.

As it goes about Oracle – it’s completely different. Like I mentioned earlier, it’s complex, requires lots of time, effort, patience and cognition until it starts unfolding itself. But when this happens… You can’t get enough of its awesomeness. And I talk about Oracle DB only, as really the only Oracle’s stronghold, but the best one out there. Oracle GUI sucks (but when you’re not used to it from the very beginning, you don’t care at all about it), Oracle Middleware kind of does too, but the DB(!) really takes it as far as you only can think it can go. And that’s exactly what makes Oracle unbeatable, in despite of all it’s rather small and inconsiderable disadvantages.

To conclude all this. Oracle is not after killing it’s support of Itanium as a measure of making Oracle/Itanium customers to buy hardware from Oracle. To me it sounds like some completely made-up conspiracy. You can search for a black cat in a dark room for whatever time you’d like, but the trick is – there isn’t any black cat and the room isn’t dark. Moreover, Oracle is not the only one, and not even the first one to announce de-support of its product for Itanium platform – Microsoft and Red Hat have done so earlier. What it means in the both Microsoft/Red Hat and Oracle cases – they all just have stopped developing software for Itanium – as simple as that. Do you think Microsoft and Red Hat have done so trying to… do something (I’m sorry, can’t really think what it might be) with Itanium users too? Or, on your opinion, in their case it is quite alright, as soon as they don’t market hardware, but if Oracle does such a thing – that’s for sure “to make Itanium customers buy servers from Oracle” simply because Oracle has it. Don’t you think it’s quite a wicked logic? Kind of double-standard too. Of course, it’s very easy to blame it all on Oracle, just because you can, the “right reason” is there, isn’t it, right?

distas said...

What I say, once again, there is no any guarantee that any now-Oracle/Itanium user would go for SPARC, instead of going for any other platform, which doesn’t belong to oracle. How exactly in this situation Oracle “pushes” customers to buying its Suns? Oracle consultants might be trying to talk you in for doing this but that’s a common practice for any business – talk you in to buy from them what they want you to buy for their benefit, not what you want to buy in your best interests. But everyone thinks for himself and makes the decision. It’s that simple.

Say, you need to change the server now, yes – because it’s being discontinued from support. (In the end, this happened multiple times throughout the history of computing, Oracle is not the first, neither the last one to do this.) Oracle suggests you buy SPARC, you don’t like the price, tech specs or a combination of whatever; it doesn’t matter what. You go with HP, IBM, Acer, whatever else is there – who stops you from doing this? Not Oracle, at least.

Further, you really think Oracle will keep discontinuing support of other architectures. I say – should this be true, Oracle will lose so much from it, it might even kill it, in a sense of massive migration towards other RDBMSs. I don’t believe Oracle’s management has “suiciders” to do this. Hence, expecting Oracle to go that route… simply goes against any common sense. Answering you question: “how many times are you ready to be forced to do it?” (change the box) – None after it’s done with Itanium.

As for “And what do you expect if Oracle hardware becomes just about the only platform to run its software?” Like I said somewhat twice by now – it’s not going to happen. Actually, this very assumption makes me laugh. Comparing of Oracle to Mainframe story doesn’t have anything in common – Mainframe was the only way computing had been done before PC and PC-servers where introduced. Afterwards, Mainframe simply remained as a completely different product on a market of its own, kind of thing. It’s not that Mainframe “shrank down” from multiple platforms to a single one, is it? Thus, comparing Mainframe history with some possible direction Oracle might go… Once again –not even somewhat relevant.

And for the last. Informix is in there – cool, I’m really glad. Honestly, no irony. I know that Informix is a good RDBMS, it has it’s market, has it’s applications and let it be. All I’m saying, and that’s exactly what brings it back to the very argument, I can’t think why would anyone migrate from Oracle on Itanium platform to Informix on Itanium just to save on platform and be sure Oracle won’t do “oops, I did it again” thing. To me, it would be some very, very wrong decision. Probably, the worst out of all the possible ones.

Anonymous said...

@distas

I agree with you regarding Itanium. Let's not forget that a decade ago HP did worse things to the Alpha userbase than Oracle is doing to Itanium now.
HP wrote a public letter reassuring it's customers about Alpha's future but discontinued it shortly afterwards.

Fernando Nunes said...

@distas

I can't understand your arguments... You say that Oracle should do this because "think is Oracle just made a proactive move which, probably, it should have done quite a number of years ago because for one – Itanuim was a looser technology from its very beginning, for two – it is going down!"

The future of Itanium should be up to Intel and HP. It has a large user base, so why don't support them? After the hardware suppliers announce the end of a platform (like HP did with Alpha) I consider it normal for software suppliers to discontinue support.. Unless of course you're loosing money on it, but as you agreed, the code base is the same. So in fact, Oracle is being "proactive" by killing a platform before the owners do.