You may have added to the confusing surrounding virtualisation. From the same document it points out that Solaris Containers are treated as hard partitions for the purposes of Oracle licensing.
Oracle has quietly jacked up software prices on IBM's Power6 iron by 33 per cent, after removing a multicore scaling factor that was in effect to reduce prices. The change was spotted by an intrepid reader at El Reg, and can be confirmed with a visit to Oracle's website. After coming under pressure in late 2004, for pricing …
Which is exactly why we have been moving to Power6. We can drastically reduce our software stack costs by spending a relatively small amount on hardware. Oracle's transportable tablespaces have made it very easy to move all of our databases to IBM Power. The increase from .75 to 1 license per core will not matter to use since we have so many licenses sitting around by moving to Power. The virtualization we have recently completed basically means we never expect to buy another new license.
What's funny is Power should have never been given .75 to begin with. The .75 was because of Sun's move from SPARCIII to SPARCIV and the decrease in performance per core. The same is now true for SPARC64VI to quad core SPARC64VII. The performance per core goes down as shown by Sun's MValues.
Net of the story is play the Oracle pricing game and implement the highest performance cores that allow virtualization which gives you the highest utilization rates....i.e. Power6
Oracle is notorious for low-balling you on licensing to get you entrenched and then sticking to you one way or the other on down the road. For this reason, we are now redesigning much of our software to minimize its reliance on the Oracle rear-end. Very soon, we will use it only the RDBMS and only for storage. We're moving all our business logic out of Oracle. They priced themselves almost out of our shop and we've used their products for 10 years.
As we've ramped up business, we've spent $8M on new hardware to handle it. Oracle's response was to slap us with $10M in new licensing costs and threatening us with more. It's ridiculous and it's going to come back to haunt them. A lot of people are sick of Oracle's bullshit and Larry Ellison's greed. Oracle is opening a door of opporunity for IBM to steal enterprise customers and migrate them to DB2 and for MS to steal their low-end and mid-range x86 customers for SQL Server. Oracle's good stuff, thut it isn't that good.
The Oracle Processor Core Factor has nothing to do with discounts. It has to do with how many licenses are required for a particular processor.
The quad-core SPARC64 is the SPARC64-VII, not the SPARC64-VII+. SPARC64-VII is rated at 0.75 per core (three Oracle licenses per quad-core chip). For an IBM POWER6, two licenses per chip are required (two cores, 1.0 licenses per core).
Originally, Oracle's rule was based on the fact early dual-core RISC chips were about 1.5 times as powerful as the single-core chips which preceded them (hence a 0.75 multiplier), and the first dual-core x86 chips had a much slower clock rate and had a performance on par with the fastest single-core x86 chips (hence a 0.5 multiplier).
The UltraSPARC T1 0.25 multiplier was basically a gift from Oracle to promote the T1 platform. A 0.5 multiplier was more accurate based on performance. Likewise, the 0.5 multiplier for Itanium was a gift to HP. Why Oracle has not rescinded it like the UltraSPARC T is beyond me. Oracle has tripled the licenses required on the UltraSPARC T with three iterations of the chip, and Oracle has only increased IBM POWER licenses by 33% over the course of three iterations, and maintains Itanium at a significant license discount compared to SPARC64 and POWER.
For some reason, Oracle only requires 0.75 licenses on the SPARC64-V single-core processor, the only single-core processor which gets that reduced license.
It doesn't have to make sense. It's Larry.
That is a fact. The power of the hardware has nothing to do with it. Oracle would prefer that Itanium wins. If not Itanium, then SPARC. The last thing they want to do is help IBM win more of the business, as they will do everything they can to move it over to DB2.
Oracle, on the other hand, is every ones enemy.
Oracle takes forever to update their pricing and because it is not "RISC" it fell into the "AMD and Intel" processors. Oracle was like woops when Montecito came out, but because of the poor performance per core they did nothing about it.
HP customers BEWARE, the new pricing specifically says "Itanium 91XX or earlier" is .5.
Tukwila 92XX which was supposed to come out last year will have four cores and will require either three or maybe even four licenses per socket. a.k.a. chip.
I think Itanium is dead anyways once HP puts Xeon in the rpxxxx line. The real windjammer will be the death of HP-UX.
There's a few misconceptions going on here, firstly Sun's RISC chips are a lot slower than IBMs, but use less power and have more threads. IBM have been crowing about how much better performing their CPUs are and how much less you need to spend on Oracle licences if you go for visualised Power CPUs.
So far as the T Series goes, well actually it's only the low speed T1s that have a .25 multipler. The higher speed ones have .75 and all T2s also have .75. ,75 for any T1 is a bit unfair .5 would be more like it, and .75 is a bit over the top for T2s, but really not that much.
the other thing about T series, and in fact most software that runs on multicore processors, is that the software doesn't really make much use of the ability to do multi-threading. in fact some search/index tasks crawl on multi-core processors, because they are serial processes bound to a single thread.
lastly, a lot of people pay far too much for their Oracle licences because they don't look at what they are buying. Sure enterprise edition is somewhere north of £20k, but you get a hell of a lot for that. Try the standard edition if you don't need everything in the Oracle stack, it costs a lot less, and your limitation is the number of CPU's in a server or cluster, not the number of Oracle systems you have, you can have as many clustered pairs as you like as long as no single cluster or server breaks the CPU limits.
Having said that, yes Oracle is overpriced for what it is, but so long as Oracle keep growing revenue, they don't care. I use Oracle because I need high availability and security that just isn't available anywhere else, but when I don't need it hello MySQL.
Bring back real programming, that's what I say, all this relational database malarkey, letting the query optimiser decide on the best access for you, drag this, drop that, its not real programming I tell you, bring back punch cards, now there was a reliable medium, no damm computer operators playing Frisbee in the computer room with the tape case covers and hitting the reset button on the tape drives while they are in use. No damm DBA interfering with your application design. I remember back in 1972 when all we had is sequential files and everything was coded in assembler and a sophisticated programming was a 6-way file merge and blah blah blah blah……..
Not many jobs in Open Source databases? You're looking in the wrong places!
If you already know how to use Oracle, learn one of the others (they pretty much all do the same stuff anyway) and then work on devising migration strategies. This is a lucrative service you can sell to businesses currently using Oracle, if you make the right pitch.
Obviously this opportunity will run out sooner or later, as the number of Oracle customers is finite to begin with and you're cutting it down as you go; but you just have find something else in the meantime with which to keep yourself going when that job runs its course.
Licence prices didn't increase when CPU speeds went from 1 GHZ to 2 GHz to 3 GHz (unless they were class-based).
Now CPU evolution dictates that frequencies stagnate, and core count increases. How is that different from a licencing viewpoint?
It is actually worse from a user view point in many cases (not for OLTP work obviously) as only multi-threaded loads benefit from a multi-core CPU, whereas sequential loads can use all the extra grunt from a frequency increase. I'd take 10 GHz CPU any day rather than a quad-core 2.5 GHz, where most cores idle away.
So we get conned twice: small performance gain for most apps (applies mostly to desktops, but not only), increased licence costs (applies mostly to servers, but not only).
Greed at play, again...
Article says, "The Oracle price hike on Power6 chips seems unfair given that the quad-core Sparc64 VII processors used in Sun and Fujitsu machines have the same 0.75 scaling factor"
The Power6 processor gains most of it's horsepower from an increase in clock rate (POWER6: 2.2GHz to 5.0Ghz, while SPARC64 processors gain their throughput through an increase in number of cores (SPARC64: 2 to 4 cores), while the CoolThreads SPARC T processor gain their throughput through a massive increase in threads per core (T: 32 threads to 64 threads.)
Article says, "To be fair, Oracle should run a database benchmark test on each processor and come up with a literal scaling factor based on possible clock speeds of all processors and make the scale all relative to the performance of one machine that it picks as the gold standard."
I agree with this sentiment. Just comparing POWER and the SPARC processors...
To be FAIR, the scaling factor for POWER6 should have been 1.50 instead of 0.75.
To be FAIR, the scaling factor for SPARC64 should have been 0.50 instead of 0.75.
To be FAIR, the scaling factor for SPARC-T should have been 0.50 instead of 0.75.
Even with the chart above, POWER would still have an advantage of scaling factor of 0.50 per core over SPARC64 due to IBM's incredibly high clock rate!!!
Most applications that require a third-party database will use a major commercial vendor, like: Oracle Database, IBM DB2, or Microsoft SQL.
Oracle's continuing punishment of SUN and "giving the farm" away to IBM has always seemed odd to me, considering that IBM is a direct commercial competitor, while SUN's MySql is not a direct commercial competitor. The migration of Oracle RDBMS to IBM DB2 is something that IBM's professional services is something that they would LOVE to do, while there is no equivalent professional services group in SUN to move Oracle RDBMS to MySQL.
Oracle needs to "get over it" and stop "cutting it's nose" on SPARC "to spite it's face" - it is just feeding the competition!
Article says, "It will be interesting to see what Oracle does when Sun and Fujitsu roll out the Sparc64 VII+ quad-core chips, which they are expected to do soon"
I don't know why it would prove interesting. It is not like the clock speed will double, as POWER5 to POWER6.
Article says, "Intel gets its quad-core "Tukwila" Itaniums out the door in June or July. The Tukwila chips should certainly get their scaling factor removed, but given that HP and Intel do not have a database software business, I would venture that Tukwilas might sneak by with a 0.75 scaling factor."
There is Microsoft SQL Server on those platforms. The scaling factor may remain, to just be competitive with Microsoft SQL Server. If the clock rate doubles, removing the scaling factor removal may be a reasonable thing, but I doubt the clock rate would double.
Article says, "various other Sparc, and other chips have a 100 VUP rating. ...it is hard to imagine a 16-core "Rock" UltraSparc-RK chip not being in the same range when it comes out sometime in the fall."
I highly suspect that a 16 core RK chip will be 16x faster than a SPARC64 V, VI, or VII core. This being the case, IBM cranking up the VUP rating would be completely "off the chain".
HP systems will not ship with Tukwila chips until November. That is from HP directly as they try to pretend Itanium is socket based pricing and customers will only have to pay for one license when Itanium goes to four cores. It will be a rude surprise when customers have to pay for 4*.75= 3 licenses per chip.
Anonymous says, "Try the standard edition if you don't need everything in the Oracle stack, it costs a lot less, and your limitation is the number of CPU's in a server or cluster, not the number of Oracle systems you have, you can have as many clustered pairs as you like as long as no single cluster or server breaks the CPU limits."
The licensing on this is very restrictive.
"When licensing Oracle programs with Standard Edition One or Standard Edition in the product name, a processor is counted equivalent to a socket; however, in the case of multi-chip modules, each chip in the multi-chip module is counted as one occupied socket."
Most Intel and POWER processors are MCM's... making each core a socket under their licensing terms.
Biting the hand that feeds IT © 1998–2019