Slashdot Mirror
Rambus Wins Case Against Infineon
rednoise writes "Yahoo is running a story about how a Federal Circuit Court in California (I think) has (unbelievably) ruled that RAMBUS did NOT intentionally mislead members of JEDEC when the committee was developing the SDRAM specification. RAMBUS' stock skyrocketed something like 57% on the news. This is very bad news for owners of computers."
You might consider reading the article a little more carefully. It's being sent back for "reconsideration." Nor did the article say that the court found they did NOT perpetrate a fraud... it just said that it wasn't proven that they did.
So, it's neither good news nor bad news for anyone but Rambus, since they're not dead in the water. Infeneon will have to keep going.
rambus was a memeber of the people deciding the SDRAM standard and they didn't disclose to the group that they were patenting one of the standards they helped push for (which is now the standard) or something
The Court of Appeals for the Federal Circuit is based in Washington D.C. They hear appeals from all Patent cases, and almost all of the judges are former patent attorneys. Because Patent cases are so rarely litigated, many district court judges make poor decisions on the law (especially in constructing claims) when they actually have to handle a case. The court overturning the appeal does not necessarily mean that Rambus wins the patent case. The court ruled that, in construing the claims, the judge defined 5 terms wrong. It is still possible to find that Infineon does not infringe the patents. The court also overturned the fraud charges because they felt that the JEDEC did not uniformly enforce their patent policy. This decision shows why it is important that these standards groups be very strict in composing and enforcing their patent policies so crap like this doesn't happen.
Come play Heroes of Might and Magic Mini online.
In a nutshell, Rambus thinks that they own (at least part of) the SDRAM specs. When their own RAM flopped due to being too expensive and not offering enough performance, they resorted to suing just about every RAM manufacturer around for using "their" SDRAM spec. Mind you, Rambus did aid JEDEC in defining the spec, but, as I understand it, lied about the availability of some patents that Rambus had so they could be incorporated into the "open" SDRAM spec.
As soon as Rambus's real business began to suffer, they turned around and said "Actually, no, we changed our mind and now you DO have to pay us huge sums of money to use this spec that has now become an industry standard".
So, it's bad for computer users because Rambus is going to levy huge "royalty" fees against other RAM manufacturers, and that cost is gonna get passed right back to us, the customers. Say goodbye to the $12 128MB chip...
Arseholes to 'em, I says.
End of lesson. You may press the button.
That was published by Reuters. Yahoo doesn't make news; they just rebroadcast feeds.
For more information, click here.
Umm iirc this case isn't because Rambus invented the iron-steel-copper interconnecting rails and patented them, it's because they invented them, got everyone to agree to make them the standard, retool their factories to make them AND THEN patented them, pretty much screwing them.
It's a different case, and here's why.
At the time that the SDRAM spec was created, great steps were taken to make sure that the spec did not contain patented material, or that the company with the patent did not plan on pursuing royalties. Essentially, Rambus was asked if they had any patents relating to the SDRAM spec. Rambus said, basically, "Don't worry about it."
So JEDEC happily went on to create the SDRAM spec in the confidence that all companies involved had shown good faith.
Later, once SDRAM was everywhere, Rambus tried to make everyone switch to their vastly more expensive Rambus RDRAM. When no one wanted to pay for it, it looked like curtains for Rambus, right? WRONG!
Rambus turned around and said "Oh, by the way, we DO have patents on stuff in the SDRAM spec, and now we're going to collect insane amounts of money because it's now the industry standard."
The issue is not, and has never been, whether or not Rambus's patents are valid. The issue is that they acted in bad faith, violated several signed agreeements, and are putting large portions of the IT industry at risk to preserve their own sorry asses.
So yes.. it's a different case.
End of lesson. You may press the button.
Explain how. Rambus does not physically produce RAM, nor do they sell RAM. They simply license the technology they came up with to other companies, who are responsible for producing the chips.
This is very bad news for owners of computers
:)
Hehe...but not if you own some Rambus stock
Anyways, this might lead to lots of other pending lawsuits and cases against Rambus getting settled. The district court judge is taking a lot of flack now for how this case was handled - jury persuassion, etc. Expect to hear more about this in the coming days.
The EETimes article on this story has (besides a lot more meat) a bit of an insight:
"Having framed the duty of disclosure in the above terms, the court concluded that Rambus did not breach its duty as to the SDRAM standard because none of the claims in its patents and pending patent applications reads on that standard," [legal analyst] Balto added.
HIV Crosses Species Barrier... into Muppets
Actually, the SDRAM royalties it has sought are thought to be closer to 0.75%. It was on DDR-SDRAM, which Rambus considered to be a would-be competitor to its own Direct Rambus RDRAM, that Rambus sought a royalty of 3.5%. See this story from Electronic Buyers News in 2001 concerning statements made in the trial.
Sorry, not a lawyer, but keep in mind RAMBUS hasn't "won" anything. As one of the previous posters explained, the case was sent back because the earlier judge misdefined 5 terms.
To answer your question, I think the legal process will solve this. (IOW, the threat of a lawsuit would prevent this kind of deception)
I imagine they missed a decimal point. Should've been 5.40$ I am guessing.
> but the effect of Rambus royalties is insignificant at best.
Ah yes, the "it only costs us a couple of bucks, why worry" argument. So, when DRM laws get passed, and all the DACs on your sound card have to support DRM and they start costing $7 per unit on the sound card instead of a couple of cents, don't worry? And the same in the CD/DVD drive.. And then the video card makers find out about one of the partners in the AGP spec have a patent on something or other and video cards cost $15 more for "royalties", and then the network cards, and so on and so on, and then all together that brand new PC costs an extra $100 as a result of legislation and royalties, we just should suck it up because it's only a bit here and a bit there?
>I read a while ago about the cost of RDRAM. The cost premium has not so much to do with Rambus royalties as it does low yields during fabrication.
So that's the buyer's fault that they have an inherently less efficient production process? And almost nobody was paying Rambus royalties, because almost nobody besides a couple of Rambus' best friends was/is making RDRAM anyway, since DDR SDRAM is cheaper and faster.
>The story mentioned that in the total cost of a 128 or 256 MB RIMM, some 3-5 dollars was the royalty.
>This was also when memory prices were significantly higher per meg.
The price of a DIMM on a per meg basis will not have any effect on the royalies that go to RAMBUS as a result. If a DIMM cost $25 before and $30 now, when its price would have dipped to $10, it'll still cost $15.
> Maybe $3-5 pisses you off a whole lot, but in exchange for great memory design (RDRAM is damn good for P4) I'd say its a small price to pay.
>I didn't here you complaining that Philips owns the CD standard.
Maybe that's because Philips didn't sit on a council to decide what the next audio system to replace cassette tapes was going to be, secretly patent a whole pile of technology around CDs, then push everyone to adopt CDs, wait a few years for CDs to become firmly entrenched in the marketplace, bring out their own SuperCD(TM) format, watch it tank, and then turn to Sony and all the others who used the "jointly developed" CD standard for all of their devices and go "Hey guys, we've got this patent we never told you about on the CD format to do with X. As a result you all owe us $5 a player. Large bills preferred. Thanks."
No. Instead Philips developed the standard themselves with Sony's help and both licensed it to everyone, up front, with clear and open terms . And, Philips is rather adamant about defending the standard as well. All of these "copy protected" audio discs aren't allowed to use the "Compact Disc" logo, you'll notice. Philips won't let them.
It's so bad because you forgot the five letters that follow DDR. And those are S, D, R, A, and M. Double-Data Rate Synchronous DRAM. It's nothing more than faster SDRAM.
The decision was by the U.S. Court of Appeals for the Federal Circuit in Washington, D.C. It is a special federal appeals court that hears patent cases and special federal issues such as takings, veterans affairs, and federal contract disputes.
The outcome is surprising on its facts, as the documentation shows that Infineon had actual intent to deceive other members of the memory industry.
However, the Federal Circuit now has a twenty year history of ignoring the law and creating its own jurisprudence in favor of strong patent rights. In the process, economic efficiency, fairness, and consumer protection have been lost.
Although the court is limited to patents, it frequently rewrites antitrust law to exclude liability for patent misuse. It frequently revises claim constructions on an ad-hoc, nonsensical basis that is impossible to predict, often ignoring the patent itself, and almost always in favor of the patent holder.
It has made invalidity harder to prove, requiring printed documentation of all elements of a claimed invention, and has limited obviousness to a very narrow set of circumstances where documents actually state that a patented combination is in the prior art. This is somewhat of an oversimplification, but it is unfortunately almnost accurate.
It has expanded the realm of patents to include business methods (harming free enterprise by making the mere right to enter an industry subject to patent rights), genomes (which, while they already exist, are always useful for growing hair--by using this claim they can cover all uses for the genome), and algorithms.
It frequently ignores federal procedural law, and has been reversed a few times by the supreme court on these grounds in recent years.
While many on slashdot frequently cite the "Microsoft Patents 1 and 0" story from the Onion, the Federal Circuit has actually gone that far in real life. In a patent case brought by Excel Corp., the Federal Circuit affirmed the validity of a patent over a one bit flag used to determine whether two parties on a phone were using the same phone provider.
Yes, Excel has a patent on a one bit flag (ones and zeros) when applied to phone networks!
There is much more to say about this, but that is for another day.
In addition SDRAM has ~10ns latency, RDRAM adds 2-20ns (do I have my units right? it is ns yes?) because it is a serial infrastructure, meaning some chips on a RIMM are farther from the memory controller than other chips on a RIMM. Every chip on a DIMM is the same distance from a memory controller; In theory this makes boards more expensive and raises system prices, but the licensing fees (and higher expense of manufacturing RIMMs over DIMMs, though that is secondary) raise the prices considerably.
The belief that RDRAM is a superior technology to SDRAM is at best a matter of opinion and at worst an absurd myth. It COULD be a more cost-effective technology, and for systems which involve a minimum number of RAM chips like game consoles, it is effective because you are paying a minimal amount for licensing in exchange for the company playing up the fact that you're using it in their console even though that has NOTHING WHATSOEVER to do with its suitability in a PC, coupled with the fact that you are using very little of it making it a small portion of your total cost... RDRAM makes great sense when you're not using very much of it (down in the ~16 MB range, where you only need two to four chips) but when you start putting it in PCs it has little to offer over DDR SDRAM; When it has an additional licensing cost to pay over an already inflated price for more expensive modules, it has none.
The benchmark you cite shows better performance for DDR SDRAM on intel's solutions than on the athlon system, which leads me to believe that it is possibly cpu-dependent.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
You! Yes, you.
... what is that you might ask?
Time to learn about Docterine of Lasches
Well, its illegal to sit on a patent that you own, and purposely delay patent litigation for your own financial gain. If you didn't know about the infringement, then its fine. Not every company can act as a world-wide watchdog for patent infringers. But if you knew somebody was using your patent, and you held off on enforcing it for your own financial gain, there are precendents that make this illegal.
Link and learn
"Old man yells at systemd"
er, I'll learn to link, first .. but herer it is
"Old man yells at systemd"
U.S. Court of Appeals for the Federal Circuit, Rambus v. Infineon Technologies [corrected]. If the site asks for a password, hit "Cancel" and the document (a 60-page MS Word file) should appear.
Patents are only public when they are finally approved. During the interim between filing and approval, they are not visible. These so-called "submarine patents" can be kept invisible for years, if you know the appropriate paperwork tricks. That whole aspect of the patent process is under close scrutiny and it is already much harder to pull off than it was a few years ago.
But the point is that RAMBUS filed for a patent, then pushed their idea as part of the SDRAM standard so that when their patent was finally granted, everyone would owe them money.
There is no amount of diligence or research that could have unearthed those patents during the approval process, short of torturing the RAMBUS representatives.
RAMBUS had pending patents on this technology. The standards commiteee required disclosure of issued patents, but did not require disclosure of pending patents.
I can't believe that the above post was marked insightfull. RAMBUS did extend pending patents, so nobody was able to read them BECAUSE THEY WERE NOT WRITTEN. Additionally, what do you think a standards body is? A place to fight against each other or a place to cooperate for the benefit of all companies? A common hardware standard protects agaist a battle of rivaling incompatible standards. It removes the threat of buying the wrong flavour of technology from the consumers and saves the company the losses of investing in the same. Think of GSM, it was developed this way. The things RAMBUS did are hurting the trust between companies. It may lead to expensive parallel developent, rivaling, more expensive products, irritated consumers, money burned and development slowed down.
Actually, i think in the long run this breach of trust hurts the economy as hard as the things WORLCOM and ENRON did.
That would be because the FSB on the Athlon system is only 266(133 x2) [333 (166 x 2) on the new Athlons]whereas the FSB on the P4 is 533(133 x 4). The P4 simply has more available bandwidth 4.2 GB/s versus 2.1 Gb/s [2.6 for the newer Athlons]. Read here for how RAM bandwidth is calculated http://www.ocsystem.com/calmemban.html also http://www.upgradingandrepairingpcs.com/articles/u pgrade12_02_02.asp .
... Governments are instituted among Men, deriving their just Powers from the Consent of the Governed...
It's not at all unusual for an appellate court to reverse a jury's verdict. However, the standards for doing so are very high; it's not sufficient that the appellate court strongly disagree with the verdict, but rather that no reasonable jury could reach the verdict based on the evidence before it.
In a sense, it is a procedural review, because a jury's finding that can not reasonably be based on the evidence generally means that the finding is on a question that should not have been submitted to the jury in the first place.
"Suits at common law" does indeed refer to civil suits, but only those "at law."
The more significant exception is the word "preserved": you are entitled to a jury trial only in those cases for which the common law provided one at the time the amendment was adopted. This makes the analysis of "at common law" extraneous, because only the law courts provided for trials by jury: the equity and ecclesiastical courts provided trials only by (their equivalents of) judges. (Equity courts heard, among other things, cases involving remedies other than money damages; Delaware still has a separate "chancery" court. The ecclesiastical courts heard divorce actions, and other stuff no longer relevant.)
The seminal recent case interpreting the amendment is, in fact, from a few years ago on patent law. The trouble was, of course, that patent law in its present form did not exist in 1790, but the Supreme Court found the issue of infringement close enough to cases handled in law courts then. Thus, either the plaintiff or defendant may demand a jury trial.
I Can't Believe It's A Law Firm, LLP does not necessarily endorse the contents of this message.