" ...tightly integrating Oracle software with Sun hardware."
Does this mean support for Oracle software on Linux will slowly degrade? Oracle would likely prefer to sell you the entire suite of hardware, OS and software vs. just selling software.
In the past Oracle has been the type of company where they provide a one stop shop for their customers: software for every department/industry, full suite packages, consulting. While the rest of the companies piece together software from different providers.
So its very likely that they would use their own hardware now that its an option.
When I was interviewing with the DB2 team for IBM, they mentioned that the high-end IBM database servers implement special instructions (CPU-level) for accelerating certain DB operations. I don't know any more details than that, but I'd guess that's where Oracle is going to go with SPARC.
I'm not aware of any DB2-specific instructions, but POWER6 and above implement a decimal floating point instruction set, allowing things like regulatoriliy-mandated financial computations to be hardware accelerated instead of computed in software. That's clearly a nod to the same market that is buying DB2.
That might be as simple as having super-low level operations that they hand write in assembly for each supported CPU, and all other layers staying the same.
Maybe. It could just mean that they are going to make Oracle software work better with the SPARC/Solaris stuff. As to be able to sell the software for other platforms as well as give you a reason to buy SPARC Boxes from them in the future.
SUN has variety of technologies especially in Storage(via StorageTek acquisition). Oracle can integrate backup and recovery strategy with StorageTek tapes, ZFS, for example. Also, db failover, High Availability can be achieved leveraging Solaris platform..
Oracle makes up a magic number called a 'core factor' that determines license costs for a given platform. By manipulating that number, Oracle gets to favor one platform over another by making it cheap or expensive to license Oracle products on a particular platform.
Right now my IBM reps are trying to convince me that I can replace a 88 SPARC CPUs on couple E25k's with POWER on IBM hardware. They argue that with POWER, I'd have less than half as many CPU's to license, and the reduction in per-CPU Oracle license and maintenance costs will pay for the IBM servers.
To counter that, all Oracle has to do is change the core factor for licensing SPARC processors, making Oracle licensing costs on SPARC hardware less expensive than POWER or Itanic hardware. That would drive enterprise customers away from HP & IBM and toward SPARC, as the Oracle license cost for a given performance level could be manipulated such that the TCO on Sun is less than competitors.
Oracle already owned the most important component of MySQL: InnoDB. MySQL AB, the independent firm, never owned the core technology upon which their product depended. Oracle has had a death grip on MySQL for years.
Sun also had a competing product at the time of their MySQL AB acquisition: Their Postgresql consulting/support group. Nothing happened to it when they bought MySQL AB. Large companies can sustain heterogeny.
>>Oracle already owned the most important component of MySQL: InnoDB
I don't keep us as well as I should.
Falcon seemed like the upcoming champion (the ideas seemed relevant for modern systems) and I thought it ought to be a good/better substitute for InnoDB.
MySQL's best hope is that Sun throws it overboard before the acquisition completes; there's no way it will survive within Oracle -- it competes directly as a zero-cost alternative to Oracle's core products.
Well, no, not at all. If you are an "enterprise" then you are paying $$$ for the "enterprise" version of MySQL. If you aren't, then you'd never buy Oracle anyway, you don't need it. At the end of the day MySQL is pretty irrelevant to Oracle's plans for Sun.
Oracle would be crazy to give up MySQL. Controlling it, they can ensure it'll never compete with their high-end products, while taking the air out of the entry db market.
I really want to know what Orcale's real plan is. Before the acquisition they have to have figured out some way to reconcile Oracle Linux and Solaris, I just wish I knew what it was. No matter how you slice it, they are competing products in a very similar space. I suppose they could continue to develop both separately and take advantage of having two brands.
Not necessarily - the could have simply worried about their customers on Sparc falling into the hands of IBM/HP salesmen with DB2/SQLServer salesmen in tow.
Not necessarily. It's not like IBM doesn't sell x86 processor systems along with their POWER (and Cell?) processor systems.
Oracle will probably package their own hardware by leveraging Sun's resources. You will probably be able to purchase a Sun-Oracle SPARC system and have a choice of Solaris or Linux (which also runs on SPARC) to run your Oracle database on. Otherwise you can run Oracle on Linux if you aren't buying into their integrated hardware solutions. I don't see how they would have to drop everything without being able to make it work together.
Must we have this flame war again? Or if you insist, can you please at least pack your arguments into one post, complete with competing definitions of "free", and "open", and sneering sideways attacks at the other side's development model or developer personalities?
Has the DOJ approved the merger? Last I heard they hadn't and the value of Sun was disappearing by the day (because the longer it takes the more people leave and take with them valuable knowledge).
... dramatically improve hardware performance by tightly integrating Oracle software with Sun hardware
Eh? The hardware and its performance stands alone. Adding software cannot improve the hardware. Only the software performance may be increased by better hardware utilization.
It is obvious what this means. Oracle up to 10g accessed NFS filesystems via the OS. They were mounted just as normal filesystems would be. Oracle 11g connects directly to the NFS server, completely bypassing the OS filesystem layer.
Note: I am distinguishing between software and firmware, here. My point is, if the overall performance improves after a software update, you have improved the software performance. The hardware performance is what it is.
Edit: If Michael Schumacher can drive my Honda Civic faster than I can, it's not because the car improved. It's because the driver's performance improved. It is capable of 160 HP even with nobody in the driver's seat.
OK, this makes more sense to me. I would have said: "... dramatically improve hardware performance by tightly integrating Sun hardware with Oracle software".
I'm more curious to know what's going on with the mainframe side of the industry. Sun recently acquired StorageTek - I wonder if they plan to keep a lot that to compete with IBM in that regard, too.
I think they're going to use the CPU simulator Sun used to design a processor (the T1) optimized for MySQL (I think it was) to make a new processor optimized for Oracle.
Does this mean support for Oracle software on Linux will slowly degrade? Oracle would likely prefer to sell you the entire suite of hardware, OS and software vs. just selling software.