HP vs. Oracle over Itanium: A look at your options

HP vs. Oracle over Itanium: A look at your options

Summary: The Oracle/HP/Itanium dust up puts customers in the middle. Fortunately, they have many options to consider.

SHARE:

Oracle announced that it was going to drop support for HP's Itanium systems in May 2011, pointing out that Red Hat and Microsoft had made similar announcements in 2009 and 2010 respectively. Oracle also said, "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, of course, and HP disagreed with Oracle's assertion and countered with statements that they were still committed to Itanium. HP has taken its complaint against Oracle to court.

Putting the customers in the middle

While this drama plays itself out in court litigation, customers that have deployed HP's Itanium servers to run Oracle's database software have to consider their options and alternatives and select an appropriate course of action.

Possible directions

While it is unfortunate that Oracle put these customers in the middle of a battle with Oracle on one side and HP/Intel on the other, it would be wise for them to step back and consider all of their alternatives. The options include the following:

  • Do nothing and wait
  • Stay on HP/Itanium platforms and replace the Oracle software
  • Keep the Oracle software and move to a different platform
  • Change out the hardware and the software

Let’s look at each option in turn to consider the ramifications:

Do nothing and wait

Some companies will continue to use their current HP systems with their Oracle database servers until they will no longer function at all. They will make a move only when what they're doing is no longer tenable. In their view, they'll consider a migration only when the cost of maintaining their current solutions is higher than migration. Only then will they consider other options. While this approach is workable for the short term, it would be wise for decision-makers to seek longer-term solutions.

Stay on HP/Itanium platforms and replace Oracle software

Other companies will prefer to stay with their HP/Itanium systems. This leads to options including staying with Oracle's database software until it is no longer supportable or moving to an entirely different database platform.

Customers face the fact that database solutions are among the most complex types of software ever developed. To achieve the level of performance and scalability customers need, database suppliers have had to develop their own memory management, file systems, clustering and other capabilities normally reserved for operating systems. Achieving total compatibility is a nearly insurmountable task. When customers are considering a move to another database, they are looking for an easy transition that will minimize the costs of migration. One approach is seeking out a highly compatible database product. This approach would be particularly desirable if the customer has custom applications that must continue to work after the migration has been completed.

Another is going with a supplier that has excellent migration tools and services which make the process relatively painless. This approach would be appropriate if the customer is making use of packaged software products and has little or no custom applications in use. Both HP and IBM would present themselves as candidates in this scenario.

It is fortunate for these customers that several suppliers offer database software that provides high levels of Oracle compatibility. Both EnterpriseDB, a supplier of an enhanced version of the open source PostgreSQL database, and IBM, the supplier of DB2 database software, have made Oracle compatibility part of their offerings. Just as importantly, both companies’ solutions support Itanium-based HP systems.

Making a move would require that the replacement offer either a high level of compatibility or excellent migration tools and services. This requires workable replacements for the following things:

  • Scripting language
  • Functions
  • Triggers
  • Stored procedures
  • Libraries of standard packages
  • Either the ability to support user-developed packages or a way to convert them transparently

Furthermore, for customers in this category it would be necessary for a replacement database server to process SQL statements in precisely the same way, have the same known bugs and deal with collections of data precisely as the original database does. This, of course, is a tall order. Database compatibility can be difficult to achieve.

Even if a competitor offered 90% compatibility, it still could create challenges for a customer. If a given customer's portfolio of applications used only the features covered by the 90% compatibility, the migration would be straightforward. If the customer’s code made use of the 10% of the database server's features that are incompatible with the potential replacement's features, the migration could be more challenging. To make this proposed transition workable, the supplier must provide ways to reproduce or support those features in other ways.

Moving to another database

Some companies are more interested in the strength of the database platform they’re considering moving to than in how compatible it is with their existing database solution. For these customers, what the new database could do for them is a primary consideration.

EnterpriseDB would suggest that these customers move to their open source, PostgreSQL-based, product. This approach, they would say, offers good price performance and good-enough compatibility. IBM would suggest moving to DB2. It would point out that customers move to DB2 because they desire the features, functions, performance and reliability the company’s database offers. Oracle compatibility was not an overriding issue for these customers even though DB2 has a high level of Oracle compatibility. IBM would also point out that its worldwide partners offer a complete set of programs and tools to make the transition straightforward.

EnterpriseDB can’t match IBM’s staff, the scope of its tools nor its worldwide reach.

Changing out the hardware

This option means replacing systems, system management software, development tools and application software, as well as the database software. It may also require replacement of server-attached storage devices as well.

What Oracle would suggest.

This can certainly be a very costly approach but Oracle is obviously hoping that its customers will chose this path and replace the HP Itanium-based servers with its own Sun servers. However, shifting from HP to Oracle/Sun hardware is not the only option clients can pursue.

What HP would suggest

HP would hope that customers in this category would stay with them and move to one of their X86-based platforms. They would point out that Intel is behind them regardless of whether customers stay with Itanium over move to X86-based systems.

What IBM would suggest

IBM would propose a number of different migration strategies here. Customers could move to IBM's X series (X86), its P series (Power architecture) or even its Z series (mainframe). They would suggest DB2. IBM would suggest that Oracle’s approach means moving from one platform with an uncertain future to another facing a similar future. It would point out that the benefit of its extensive enterprise mainframe experience has made all of its systems more scalable, reliable and manageable that those from other hardware suppliers. Furthermore, IBM offers "Server Makeover" programs to help these customers determine what hardware platform would be their best choice and then helps them make the move.

IBM clearly believes that its offerings outgun anything Sun or HP has to offer. Furthermore, IBM is stepping into this dust up to point out that it offers solutions that outperform HP’s Itanium systems and Sun’s SPARC systems when supporting Oracle’s database software. the company also notes that it has executed thousands of migrations from HP Itanium and Oracle/Sun SPARC-based servers to its own systems using DB2.

What would the ideal migration solution include?

An ideal migration solution would contain the following elements. Each customer would be able to select the options that made the most sense and would address their unique requirements. Hardware and software offerings must have a track record of success and a strong, compatible set of planned future offerings.

  • A selection of servers that range from small to very large
  • A complete stack of software (ranging from single vendors to numerous offerings from third parties) making it possible for customers to pick and choose the products needed for their unique situations
  • Free or low cost services that can help customers decide the best course of action. This could include determining the reduction in ongoing cost of ownership, how to move from one set of applications and tools to another, training for IT administrators and developers, and even complete proof of concept demonstrations.
  • Packaged offerings of hardware and software simplifying the move from current to future platforms. It would also be very helpful if the supplier offered trade-in programs to take the obsolete systems off the hands of the customer.
  • Professional services and tools to make the migration easy, timely and cost effective.

Final analysis

Since every company's IT environment is a bit different than any other, there is no single cut and dried solution. I project that most companies finding themselves caught between Oracle and HP will stay the course and continue using their current systems for a time while they sort out their options. They have to sort through from the following options:

  • Do nothing and wait
  • Stay on HP/Itanium platforms and replace the Oracle software
  • Keep the Oracle software and move to a different platform
  • Change out the hardware

But if businesses decide to move ahead, the key to making a good choice is to stop for a moment and consider everything that would be needed to make the transition work. They need to consider if they have a significant tie to HP’s hardware or Oracle’s database software while they weigh their options. If companies have a significant portfolio of custom applications, then choosing with a supplier that offers some level of database software compatibility is a good start.

If companies have a strong tie to HP’s hardware, it would be good to consider a supplier that continues to support HP’s Itanium systems. It also offers X86 systems as well.

For most customers, however, would be best to select a supplier that has the tools, the services, the partners and systems to create a long-term solution to this problem.

IBM certainly can be seen as a strong supplier of products and services that can address this issue. The company offers a broad portfolio of systems, software and professional services to help in a migration. They also have an extensive, worldwide partner ecosystem can offer help wherever the customer is located.

EnterpriseDB, on the other hand, largely offers a much more limited migration plan — an Oracle compatible database, migration tools for that database and services that could be used to help companies in transition from the Oracle database to EnterpriseDB’s product.

Related:

Topics: Oracle, Data Centers, Data Management, Enterprise Software, Hardware, Hewlett-Packard, IBM, Software, Storage

About

Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. In his spare time, he's also the managing partner of Lux Sonus LLC, an investment firm.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

25 comments
Log in or register to join the discussion
  • Like some of my friends trying to find a "perfect" husband!

    Rots of ruck!
    kd5auq
    • Message has been deleted.

      filhomarques
  • EnterpriseDB--stronger than you think

    EnterpriseDB's ecosystem, together with HP, is a large as IBM's, and remember, IBM's Oracle compatibility comes from EnterpriseDB's technology.
    sean.doherty
    • RE: HP vs. Oracle over Itanium: A look at your options

      @sean.doherty

      IBM is the one who invested in EnterpriseDB, not HP... so counting it in HP's stack is non-sensical. Also, IBM has a far greater market share of the database market that Enterprise DB.
      scotth_z
      • RE: HP vs. Oracle over Itanium: A look at your options

        @scotth_z Agree, EnterpriseDB is not in the same class as Oracle or DB2 from a feature/function comparison or a performance comparison. It is not a fit for the workloads that would have been running on Oracle-HP-UX.
        AdamS12
  • RE: HP vs. Oracle over Itanium: A look at your options

    Switch to Informix performs better than Oracle anyway.
    mrlinux
  • Too bad HP fell for the Itanium

    Titanic. Their PA processors were a good product. Itanium turned out to be a huge waste of time and money, which would have been better spent continuing their PA line. Oops!
    GoPower
  • RE: HP vs. Oracle over Itanium: A look at your options

    Beware Oracle - I am a huge supporter of Oracle products, but this should be a warning to all current and potential customers. Unless you are on Sun hardware, you run the risk of Larry deciding you SHOULD be on Sun hardware.
    tbuccelli
    • RE: HP vs. Oracle over Itanium: A look at your options

      @tbuccelli Yup. Many Sun customers have figured that out and are leaving the platform at unprecedented speeds. Take a look at marketshare numbers in places like Gartner. Amazing thing. Sun is nothing more at this point than a supplier of Exadata and Eaxlogic components. Wow. All those Billions Oracle spent was to get a hardware source for Exadata. Things that make you go hmmmm
      mikies
  • Degree of cross-platform capability

    One thing that an organization should already have, or create, is a detailed analysis of their dependence on vendor specific, proprietary capabilities.

    Can 90% of the organization's database requirements be met by any ANSI:2008 conforming product? What percent can?

    The argument goes that a vendor specific proprietary extension provides a competitive advantage in speed performance. While there are capabilities that can be difficult to recreate in another product, many times this can be reduced to the amount of hardware resources provided. Adding hardware is generally much less expensive than changing software.
    pwatson
  • HP's biggest issue...

    ... dropping the Alpha for Itanium. Itanium was a poor design to begin with, slower than the Alpha processors it was designed to replace at first.

    But what do you expect from HP, the company that is one bad decision after another.
    Champ_Kind
  • RE: HP vs. Oracle over Itanium: A look at your options

    Decent analysis here... going with HP all the way. Goofy Oracle can keep their hot-headed decisions for someone else to get tricked into.
    MarauderX
    • RE: HP vs. Oracle over Itanium: A look at your options

      @MarauderX Exactly...there is one more thing that is not mentioned...if one were to migrate to DB2 from Oracle, your cost to own goes way down. IBM knows that pricing is the main method to attract Oracle customers and they are using pricing big time to get their marketshare up. Question is, how many customers will see reduction of maintenance payments say by a very large percentage as enough incentive to make the switch. Some IT departments forget that they are there to support the business and may actually stick with Oracle.
      mikies
      • RE: HP vs. Oracle over Itanium: A look at your options

        @mikies,
        I agree that some IT departments may work against the best interest of the companies they work for. But what Oracle is forcing is a major cost add... not only are they suggesting a processor switch.. but there is a huge cost in migrating and the bean counters may force the move to a more trustworthy vendor... ie... IBM, simply because Oracle screwed them.

        One thing to think about with Oracle, is that they claim that Intel doesn't have a roadmap for the future of Itanium... the same goes for Oracle with Ultrasparc and their dwindling market share.
        scotth_z
      • RE: HP vs. Oracle over Itanium: A look at your options

        @mikies Moving to DB2 on HP-UX *is* an option, but if you're going to migrate, why not really lower costs by moving to Postgres Plus? PostgreSQL is a renowned, mature, stable database core, and EnterpriseDB's Postgres Plus Advanced Server *is* the Oracle compatibility technology IBM licensed, and Postgres Plus costs less than half the cost of DB2! With support from HP *and* EnterpriseDB, you're covered.
        sean.doherty
    • RE: HP vs. Oracle over Itanium: A look at your options

      @MarauderX HP and EnterpriseDB can help customers move from Oracle with relative ease. Our Oracle Migration Program provides a detailed assessment, and we can perform the migration for you inexpensively--that's the beauty of the Oracle compatibility technology!
      sean.doherty
  • RE: HP vs. Oracle over Itanium: A look at your options

    Good analysis after a really long time. Any enterprise system has a lot of custom database code which is really difficult to phase out. This surely gives Oracle advantage. Personally I think moving to x86 rather than Sparc should be better choice.
    padmarag@...
  • Go with SQL Server and Wintel

    All your problems will be over

    http://www.microsoft.com/sqlserver/en/us/product-info/benchmarks.aspx
    hubivedder
  • Beware of IBM AIX/POWER

    "...IBM clearly believes that its offerings outgun anything Sun or HP has to offer. Furthermore, IBM is stepping into this dust up to point out that it offers solutions that outperform HP?s Itanium systems and Sun?s SPARC systems when supporting Oracle?s database software...."

    Now this is silly. Oracle SPARC has the world TPC-C record. IBM's TPC-C record is 10 million tmpc. Oracle has 30 million tpmc. HP has 4 million tmpc. Just google on TPC-C and see. There is no way that IBM can reach these TPC-C numbers. No way. The fastest TPC-C database solution on the planet is given by Oracle/SPARC/Solaris. Oracle is the only vendor that can give you this performance (typically Wall street firms).
    .

    .

    IBM has said that IBM AIX will be discontinued in some point in the future. AIX will be replaced with Linux in the long term:
    http://news.cnet.com/2100-1001-982512.html

    Speculation from me:
    --------------------------
    AIX only runs on IBM POWER servers. If AIX is killed, there is no reason to sell POWER servers. Hence, POWER will be killed too.

    IBM POWER6 servers were several times faster than x86 servers, and costed 5-10x more.

    IBM POWER7 servers are only 10% faster than Intel Westmere-EX cpu, and costs 3x more:
    http://www.anandtech.com/show/4285/westmereex-intels-flagship-benchmarked
    Next year, Intel Ivy Bridge will arrive, and they will be 40% faster than today's x86 cpus.

    Future IBM POWER8 servers will probably be slower than Intel x86 cpus, and be even cheaper.

    We all know that IBM only does high margin business. When POWER cpus will be slower than x86, then IBM needs to cut the price. Then POWER will be low margin business. IBM dont do low margin business. This means POWER will be killed off. POWER servers runs AIX. Without POWER, there will be no AIX. That is the time, when IBM will switch to x86/Linux.

    BTW, Solaris runs on x86 and on SPARC. Solaris is open sourced (check Illumos and OpenIndiana). Solaris can not be touched by Oracle.
    Orvar
    • RE: HP vs. Oracle over Itanium: A look at your options

      @Orvar
      You obviously have no clue about what is going on in IBM's universe. Sufficient to say, IBM has done certain moves to retain Power's money making ability. Unlike Ultrasparc or Itanium which have been in revenue decline for several years, Power is actually increasing in revenue stream over the last year as has their mainframe revenue.

      They recognized over 10 years ago, that the UNIX/x64 market would decline... and their strategy has been tailored to be the last HARDWARE survivor in the x64 space, just like they're the last survivor in the mainframe space. Being the last survivor, means they will continue to make a profit.

      Linux is the one taking over the space that was once dominated by the HP/UX and Solaris's of the world. The manufacturers have used OS's to sell hardware...and OS's ARE A COST CENTER, not a profit center. This strategy has run it's course in the x64 market. IBM knew this to be true and has been pushing Linux for more than a decade.

      The question is, is not whether AIX will go to the x86 market, but whether Solaris can compete with Linux and Windows Server in the x86 market... and how much money Oracle will invest in what will ultimately be a marginal OS that can run on anyone's x86 system, (but why would anyone run it when Windows and Linux have the advantages in R&D, developers and workforce, over the long run).

      The other problem Oracle faces, is PR... This is the 3rd major offering that they are killing off...in the last year and a half. Many companies have invested millions in their IT systems... and are now facing possible multi-million dollar migration projects. Because unlike IBM, who will sometimes help their customers migrate to other technology... Oracle has stuck it to their customers, not just the migration costs, but the risk to THEIR business.... This is generating anger in the IT community.

      BTW, Solaris has been running on x86 for several years now... How are Oracle's hardware revenue numbers been? Now compare that to AIX/Power's UNIX market share...over the last decade.
      scotth_z