Slashdot Mirror
Rambus Loses $4B Antitrust Case
UnknowingFool writes "In a vote of 9-3, a jury found that Micron and Hynix did not collude to manipulate DRAM prices in a violation of California anti-trust law against Rambus. The jury also ruled that the Idaho based Micron and the South Korea based Hynix did not interfere with Rambus' relationship with Intel. On the first point, Rambus argued the two chip makers conspired to keep Rambus RDRAM prices high while artificially keeping their SDRAM prices low. Micron and Hynix countered that high RDRAM prices were due to technical problems of the design. On the second point, an Intel manager testified that Rambus contract stipulations soured the relationship. The clause that Rambus insisted and would not waive was that to use Rambus RDRAM, Intel had to agree to give Rambus the ability to block Intel processors if Rambus felt Intel was not promoting RDRAM sufficiently. Rambus initiated the suit and the $4B was how much Rambus calculated it lost in profits."
The leech that used the courts.
No hard feelings, eh?
A feeling of having made the same mistake before: Deja Foobar
Microsoft comes to mind
Muchas Gracias, Señor Edward Snowden !
They lost four dollars and a capital B it seems.
it has been well-established that Rambus ran off to patent a developing industry standard in RDRAM, stealing what was to be a public standard. they can rot in hell with 640K, they have it coming. got all the morals of Darl McBride and his little troll company.
if this is supposed to be a new economy, how come they still want my old fashioned money?
I wish Karma would come back and wipe out all the trolls! I will settle for it one at a time. Perhaps the MAFIAA's will be next. I made it a point never to buy anything with RDRAM in it after those lawsuits RAMBUS filed. Maybe if we are lucky they will go RAMBUST.
And there was much rejoicing.
nuff said
all language nazi's will burne in heil!
an Intel manager testified that Rambus contract stipulations soured the relationship. The clause that Rambus insisted and would not waive was that to use Rambus RDRAM, Intel had to agree to give Rambus the ability to block Intel processors if Rambus felt Intel was not promoting RDRAM sufficiently.
Wow. I'd have told them to F off too...
"You want to know how to help your kids? Leave them the fuck alone." -George Carlin
But I saved $4 billion by not using RDRAM in my home PC!
Sorry, but gray text on gray background is making my eyes bleed.
Intel had to agree to give Rambus the ability to block shipments of Intel processors if Rambus felt Intel was not promoting RDRAM sufficiently
Summary gets confusing when you leave words out.
Nothing used today came from Rambus. Anything they had they which was different from what JEDEC was already planning for has withered and died. They didn't know how to design RAMs. They were a just patent troll company from Los Gatos whose only real purpose was to hijack the open standards efforts of the RAM industry. Nothing in modern RAMs bears ANY resemblance to an RDRAM, unless you count the obvious use of differential signaling that had been contemplated by JEDEC for a decade before Rambus existed. Oh, wait, they both used transistor memory cells as well. Surprised Rambus didn't manage to get the gullible patent office to give them a patent for that as well.
...
Imagine a world without Slashdot, for surely, without this stuff as fodder you'd be able to hear a pin drop in here.
Patent litigation: A doctrine of Mutually Assured Destruction... in which everyone seems willing to push the button
this place is quite silent already. There are no voices of reason and the rest don't matter anyway.
Bzzzt, wrong!
All Rambus did was patent the SDRAM specification that was already being worked on. They attended the SDRAM meetings then patented everything they heard. They were nothing more than leeches attempting to game the system for profit.
So yeah, maybe SDRAM uses stuff that is patented or owned by Rambus but Rambus stole it all from SDRAM in the first place.
Again, all of that was in pre-existing products or draft JEDEC standards. None of it was patented until Rambus came along and took credit for other people's ideas.
They helped to advanced the industry by patenting once-public ideas ? I cannot agree with that statement on any level.
If anyone should be accused of articially maintaining high RAM prices, it's Rambus. Their trolling and subsequent royalty racket has cost the world far more than 4 billion, not to mention the costly and frustrating period where Intel boards exclusively supported Rambus. That move alone set the SDRAM industry back a few years.
Rambus is the perfect example of how NOT to run a tech company. Leave IP theft to the Chinese, at least they don't patent the stuff they steal.
-Billco, Fnarg.com
Wrong. Rambus submitted much of the SDRAM specifications to the group, they didn't steal it from the group. They failed to disclose they they were seeking a patent on the technologies, and as because of that failure to disclose, have lost much of the revenue they might have made had they properly disclosed it.
make imaginary.friends COUNT=100 VISIBLE=false
Based upon court rulings, that's false. Rambus mistake was failure to disclose that the information they were submitting to JEDEC was patent encumbered and then trying to collect patent royalties, not "taking credit for other people's ideas".
make imaginary.friends COUNT=100 VISIBLE=false
... by patenting once-public ideas ?
That's false. Rambus did develop and submit the technology to JEDEC, but they failed to disclose that it was patent encumbered and then refused to agree to JEDEC FRAND terms and tried to license it separately. They lost those cases in court.
I agree with everything else you said.
make imaginary.friends COUNT=100 VISIBLE=false
I was just starting in the IT industry when Rambus came out. Part of my quiz as a low-level tech for my first IT job was to ID some pieces of hardware. I had been building PCs using AMD & SDRAM for quite a while, due to how cheaper they were vs. Intel. I was handed a Rambus chip, and my exact response was:
"It's memory, and it has a heatsink. Either it's high-performance, or very inefficient"
Got the job, and I soon learned it turned out it was the latter. I saw how much Rambus sucked firsthand, and was glad when regular DDR RAM started to take its place. Farewell, it wasn't nice knowing you.
Imagine a world, where people weren't in fact sheep and could think for themselves for a moment.
Guess that depends on which court ruling you refer to. The article you mention describes repeated appeals with different outcomes in each case. Neither the courts nor the regulatory bodies have shown themselves capable of grasping or even caring about the technical history of the RAM business. Rather than wikipedia, you might want to check out old issues of EETimes or read postings from the comp.sys.ibm.pc.hardware newsgroup in the 1999-2001 timeframe.
I had to buy an entire second hand server just to upgrade the server at my old work. it was cheaper to buy the whole server than an extra 1GB ram. We then had to throw out the extra 2 processors and motherboard
... and here I was about to mention that the "upgrade path" available a few years ago for one of my customer's system when he wanted to make his Dell/Intel/Rambus machine better was a choice between buying $600 worth of Rambus, or $600 worth of new machine (dual-core instead of single-core, 10x the hard drive space, twice the DDR of the "upgraded" Rambus, with a DVD burner instead of a cdrom, and a fairly nice graphics card)... guess which he purchased?
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
>Leave IP theft to the Chinese, at least they don't patent the stuff they steal.
Now you gone and done it, they'll start pattenting stuff in the US and own yet another big part of the US market.
by TheSpoom (715771) Uncaring Linux user here. I have nothing to add to this but please continue. *munches popcorn*
It had all been done before, but Rambus added the idea of doing it -- for RAM!
A protocol bus -- for RAM!
Terminated, controlled-impedance transmission lines -- for RAM!
Differential (actually threshold-based pseudo-differential) signaling -- for RAM!
Source synchronous clocking -- for RAM!
Delay-locked loops -- for RAM!
Bidirectional signalling using (current-mode drivers) -- for RAM!
When I worked there I patented the principle of linearity when it is used in time-domain simulations -- for RAM! (One of the less valuable but most crazy ideas.)
All that silliness aside, I don't think Rambus' 2.5% royalty justifies the price premium for RDRAM. However, the fancy wafer-scale packaging, the impossibility of production testing at 800Mbps in the year 2000 and the horrific heat generated by those 28-ohm current-mode drivers was enough to kill the technology.
If they had disclosed it, the standard would probably have been different. Standard making groups generally don't want to base their standard on a technology owned by a single company.
Contribute to civilization: ari.aynrand.org/donate
I remember looking at the Rambus approach back in the nineties, and it didn't look practical. High latency, high power consumption, high cost, high complexity (which implies lower yield, which drives cost) and my most favorite of all, a single vendor. I didn't see why anyone would buy their stuff. Yes, they had the fastest throughput for awhile, but throughput isn't everything.
Does this mean they're finished as a company? Does anyone still buy Rambus devices?
Oliver's law of assumed responsibility: If you're seen fixing it, you will be blamed for breaking it.
But they didn't properly disclose it, because if they had, everyone else would have told them to get bent. They wouldn't have made shit for money if they would have been up front about it from the beginning. RAMBUS will be extremely lucky if following this fiasco, they aren't sued for everything they made (plus damages) for their funny business involving SDRAM. Whether they contributed to the SDRAM spec or not is pretty much irrelevant. Nobody likes in 'indian-giver'.
Rambus had also accounts with Sun Microsystems to integrate Rambus technology into their memory infrastructures. Also, when HP was making it's revision to their holding of Compaq's DEC Alpha architecture they achieved godly bandwidth that excelled over architectures of Intel and AMD. HP never released any RECENT systems using Rambus. It's like nothing makes the light of day when it comes to Alpha (ever since Intel Itanium was forced on everyone). What we are seeing from Rambus today is their critique over COMPETING monopolists. Sadly in this patent SYSTEM it is being used to assert violence on 3rd-world countries to corner a market. I would blame Intel for losing all the diversity in computer architectures. This past 15 years has been the destruction and liquidation of all in-house chip fabrications and competing lines of computer systems at the hands of what appears to be agents of other corporations becoming CEO double-agents. It's truly sickening. Homebrew computing has never been the same.
Yes, because nobody understands the minute details of modern computer technology like judges...
So I google searched RDRAM vs. SDRAM and found the first page basically entirely pro-RDRAM pages, many of them old, saying the stuff was basically gloriously great with the only exception being its high cost. You ever wonder if a company would hire a contractor to google bomb some info if they are in a major court case and want to try to influence jurors who may be doing some after-school research? It took a few pages before I found a realistic retrospective page. I contrast, the same realistic retrospective page was visible on the first page of bing.com
Not saying it DID happen, but saying it could have happened and probably does happen.
Bottom line was:
Chip companies weren't making money on it
RDRAM took up a larger amount of the die so you got less parts per wafer
RDRAM was more complicated so there were more things that could go wrong in the fab process and thus lower yields
RDRAM was inherently marked up due to royalties
RDRAM was not widely adopted beyond intel motherboards
RDRAM made little difference to the average user
Companies aren't working for free and aren't going to sell a product cheaper and at a loss when they have a more lucrative alternative
> On the subject of productivity in the past ten years, Rambus has produced the memory controller for the Cell processor, developed the first quad-data-rate memory, and other notable things.
And they were the first to design the memory bus as a proper terminated bus. DDR3 still does not do that - it relies on the connections between all the memory chips (even across modules) being very short. I think they really have something there. But their business attitude seems to be pretty evil, so nobody wants to touch it.
Rule number one in business: it is never a good idea to sue your customers, even if you are right.
Ahh yes, now that you mention it, you're exactly right. I got fuzzy on the timelines after all these years.
-Billco, Fnarg.com