Slashdot Mirror
Rambus Gets Toshiba To Sign Patent Concession
Modular writes " This press release Toshiba Signs Patent License Agreement With Rambus For SDRAM & DDR SDRAM Memory and Controllers surprised me. I didn't know that Rambus held patents for SDRAM and DDR SDRAM. It seems like they hold cards on both sides of the SDRAM vs RDRAM issue. Toshiba already had an agreement to produce RDRAM. Did they sign this to strengthen Rambus claims to SDRAM technology? Are these patents pending and contested by other memory manufacturers?" Check out the tech report for an analysis of this - it doesn't look good, I fear.
...At least not quite yet.
Just because Toshiba folded doesn't mean that everyone is going to. RAMBUS's claims of controlling DDR SDRAM are totally bogus. It is and always has been an open standard which many companies have invested a lot of money it. They are not going to just sit by and watch RAMBUS take it over for the sole purpose of destroying DDR SDRAM and forcing everyone to buy RAMBUS. This is definately going to be disputed by a lot of players in the memory market, and I think that any court would have to realize that RAMBUS doesn't have a case.
So why did Toshiba fold? If it had been Micron or somebody similar, I would have to say that things would be looking quite bad. But Toshiba has never been a big DDR-SDRAM supporter in the first place. It is possible that they cut some deal with RAMBUS because RAMBUS is desperately trying to establish some kind of justification for a future campaign against other memory makers. This could serve as a precedent, or at least some kind of evidence, for legal action. However, if that's all RAMBUS has to go on it's not going to be enough
I'm not saying that this is a good thing. Not by any means. However, I don't think RAMBUS will be able to stop DDR-SDRAM from becoming popular by doing this. At worst they might be able to delay it for a while and increase prices a bit. Considering how incredibly non-competitive RAMBUS is and how many memory manufacturers are simply sick of Intel and RAMBUS trying to force them to use what they consider to be a poor technology, I don't think anyone is going to give up on DDR-SDRAM anytime soon.
I am not an idiot. Please use my name to email me.
"That's right, I'm quoting myself."
-Upsilon
I for one welcome this shifty, greedy maneuver. If RAMBUS does manage to corner the market on volatile RAM technologies, then the computer industry will thrash around violently looking for a way out. And it will be about time.
We've become sloppy and trained to the notion that memory should be divided into segments varying by speed, size, volatility, and cost. We all spend months in college or in the field learning about the subtle differences between L1 caching, L2 caching, main memory, hard drive memory, ROM, and the trillion variations of RAM. We don't see the forest for the trees: this model of data management is the single most crucial hindrance to the advancement of computer science in our entire industry.
I for one would love to see a technology like the magram become viable through hard research and buckets of funding. Can you imagine the virtues of a system that could boast cheap, fast, large, and non-volatile memory in one consolidated chunk?
Imagine how intelligent an OS design such as the orthogonally persistent EROS operating system would become if the distinction between disk and memory were eliminated at low cost.
So, while I fear the short term repercussions of what it would mean for a company as shoddy as RAMBUS to gain broad control over the hardware market, in the long term such a development might just shake us out of our doldrums. Which can only be a good thing.
-konstant
Yes! We are all individuals! I'm not!
-konstant
Yes! We are all individuals! I'm not!
Over the past 6-12 months, we've seen a lot of articles on the 'patent wars' between Japanese companies (and when you consider the RamBus partners, this is pretty much a treaty in these ongoing wars)
To understand the significance (or lack thereof) in RamBus having patents on both sides of the SDRAM/RDRAM divide, you have to understand the outlook of the Japanese patent system (both its examiners and users)
In Japan, companies tend to patent much smaller details and modifications than are considered patentable in, say, the US. If you patent an "iridium-based' lightbulb, you can expect other companies to patent 'iridium based' bulbs with different bases, shapes, colors, multiple filaments, higher output, lower voltage, etc. "Iridium filament bulb with bayonet attachent for use in battery-powered flashlights" or "Iridium filament stage lights" would not be considered patentable in most of Europe or the US, but I've seen exactly this sort of patent on special bulbs in Japan.
Before WWII, this allowed japanese manufacturers to efectively bypass international patents in the domestic Japan market: "Maybe GE has the patent on that bulb, but we have the patent on green ones!" This would hold up in a Japanese court -- not that many companies tried to enforce their rights in Japanese courts back then. it wasn't a major market, and "Made in Japan" meant cheap and disposable.
In time, The Japanese system led to a strangling network of interlocking patents. Any big company could push you to license your patents, else they'd threaten to enforce the myriad tiny patents they or their corporate allies ('Kureitsu' relations is 'schoolyard politics') held to paralyze you or block your use of your own patent.
Result, lots of sharing and intertwined relationships, which is, of course, how the Japanese preferred to do business, anyway.
Ask anyone who tried to open a Japanese market (or business presence) in the 70's/80's. The culture shock was like a 2x4 to the head. It felt like socially condoned fraud, theft, monopolism, -- and a lot of words I can't print here. Comparisons to the Mafia were common (behind closed doors)
This is simply business-as-usual in Japanese business. It's not corruption or even scheming, it's their culture. "The nail that protrudes must be hammered down" is one of their most famous proverbs, and many of their well-accepted (by society in general) applications of it would be megabuck civil rights violations here.
It's like a Soviet and an American arguing economics in the 50's -- the premises are simply so different that you can't make a single logical step without considering them
If you can go to bed, knowing you did a valuable thing today, you're very lucky. If you can't... it's not bedtime
I'm not surprised that, of all the memory manufacturers out there, Toshiba settled first. Back in Nov '99, they settled a $2.1 Billion dollar class-action lawsuit over a minor floppy drive problem.
I recall that some analysts explained that Toshiba settled because they were, afraid of the potential of a runaway lawsuit that could have produced far greater losses (which boggles the imagination). The same lawyers (who walked away with an obscene pile in this case) have also gone after other companies (Such as Compaq) that had similar floppy drive problems, but in these cases the accused companies have decided to fight it out.
This all dates back to Spring of 1990, when Rambus filed their original patent. (TTBOMK) This patent was abandoned and continued in 1992. Eventually the 1992 patent issued, but not before several more continuations and other legalese things like that were also filed.
In the early 1990's, the SDRAM standards were being hammered out in JEDEC. Rambus was a silent participant at those early meetings. Eventually, Rambus quite attending JEDEC meetings, somewhere around 1992 or 1993, IIRC. Many of the salient aspects of the SDRAM had been settled by this time, though not hammered into their final form.
The 1990 patent was abandoned, and went without note. The 1992 continuation issued without note, as well, since it rather specifically defined an early form of the Rambus architecture.
The continuations of the 1992 patent are the things causing all of the current fury. They reference the teachings of the original 1990 patent application, and extract new claims related to the current SDR and DDR SDRAM designs.
It's interesting to note that with continuations, office actions, and the like, you can extend patent protection well beyond the intended term. Your protection begins the day you file, and extends 17 years after the date your patent is granted. Recent patent reforms have fixed this somewhat, so your protection is 20 years after filing or 17 years after issue. I don't know if this reform addresses the issue of a string of continuations being used to extend patent life.
The living have better things to do than to continue hating the dead.
A business process consisting of utilitzing the lameness of the USPTO staff to create increased revenue without doing a whole lot of work.
Prior Art
Just about every patent ever filed. But who cares.
Detailed Description
We've become sloppy and trained to the notion that memory should be divided into segments varying by speed, size, volatility, and cost. We all spend months in college or in the field learning about the subtle differences between L1 caching, L2 caching, main memory, hard drive memory, ROM, and the trillion variations of RAM. We don't see the forest for the trees: this model of data management is the single most crucial hindrance to the advancement of computer science in our entire industry.
The problem is that this multilevel approach to memory exists for a very good reason - all serious candidates for implementing RAM have an inverse relationship between speed and cost.
Sure, you could put a gig of SRAM in your machine, but you'd have to spend a pretty penny to do it. This isn't because of some evil manufacturer jacking up prices - it's because SRAM is many times less dense than DRAM, and you're paying for the silicon area that you use.
There's also an inverse relationship between size and speed. This is why a chip's L2 SRAM is slower than its L1 SRAM. The larger RAM array has longer traces and a number of other features that increase settling time and in other ways make the response time longer.
Similarly, even if you did decide to fill your motherboard with SRAM or something comparably fast, you wouldn't be able to access it at full speed. Motherboard traces are long, which increases signal propagation latency, and noise injection from the environment is worse, which decreases the rate at which you can transfer data even if you can tolerate high latency.
These signalling problems aren't going to go away, and won't change even with different RAM types. At best, you can learn to live with them (RAMBus uses a clever scheme to reduce noise, allowing them a faster throughput while keeping the lousy latency of external memory).
As for using disk as memory or vice versa - this just isn't practical. Disk is a thousand times slower than memory and a hundred times cheaper. Take out most of your RAM and watch your machine swap to see what using this as system memory is like. Check out the price tags for solid state drives to see what the cost of using memory as a drive is like (or just ask your local store how much 30 gigs of DIMMs would cost).
In conclusion, I think that your suggestion is not compatible with existing or proposed storage technologies.
I don't think most companies will just cave in like toshiba. I can't imagine the computer industry will simply let rambus essentially take over. rambus has basically made themselves corporate and public enemy number 1. That's not a good position to be in. Companies like micron and others will do everything they can to stop this, and money talk in the courts.
In addition, since when was rambus able to get patents on an open standard? Didn't IEEE or some other organization come up with the specs? Or did rambus get patents on specific technologies while the spec only covers the description perhaps? Does anyone know? This has me more pissed off than anything else in the computer world for a long time...
Intel, magnificently I might add, is playing this beautifully. They have the option to buy 10% of Rambus (in essence, they own 10% of Rambus) and they are the ones pulling the strings in this drama. They know that the Justice Dept would put the smack down on them had Rambus initiated this agreement if Intel had exercised its option to purchase the Rambus stock.
Intel is very slyly doing all they can to ensure Rambus memory at the expense of consumer dollars and AMD's viability. I have always said that Intel was much more evil than Microsoft. They just weren't as stupid and arrogant as Microsoft.
-- You see, there would be these conclusions that you could jump to
but what would stop a company from doing the same thing to SDRAM or whatever else RAMBUS is liscencing that tons of IBM clones did to IBM?
What most people are unaware of is that while IBM appeared to be getting "slaughtered" by clone makers, IBM was reaping billions of dollars in license fees for their patented technologies. Most of this was intentionally kept very hush-hush, the license included an NDA. You could not even admit that you had spoken with IBM regarding their patents.
Yes, the ISA bus came from an Intel periodical. Yes, the CGA graphics adapter came from a Motorola periodical. Yes, the trick of using the keyboard controller as an MMU was also public knowledge, but there were *Many* aspects of the original PC, XT, and AT, that were patented IBM technology.
People get the impression that IBM somehow got screwed. They made out like bandits, and enjoyed every minute of it.
This is just like television, only you can see much further.
on sending data to/from a memory chip with both edges of the clock. At the time, it was called "toggle mode", and the patent predates Rambus by 3 or 4 years.
Rambus has patented two slightly different aspects of this: First, they do it with two clocks. (differential? which DDR SDRAM uses.) Second, they have a slightly different receiver arrangement. They use two complete receiver/latch circuits, where the IBM patent used a receiver/buffer and two latches.
Rambus also has a patent on adding a DLL to the mix.
Far more serious than these is their patent on the "CAS Latency Register" used in both SDR and DDR SDRAMS. This is a technique used YEARS before on other chips, but this is the first time it has been placed on a memory chip.
The history of this whole mess is badly checkered, and beyond this I really can't comment. Besides, I don't want to burn the phosphors off my tube and blister the keycaps by delivering my true feelings on this issue.
The living have better things to do than to continue hating the dead.
I've been hoping that someone would bring this up so that I could rehash the discussion we had on RDRAM back when this whole latency story broke. Below you will find a number of links to other places. Suspiciously, out of the holy trinity of hardware review sites (Tom's Hardware, Anandtech, and Sharky Extreme), the ONLY one that speaks up in favor of RDRAM and doesn't talk about its latency problems is Anandtech. Hmm...
From Sharky Extreme on this page:
From Tom's Hardware: This page tells what the theoretical bus bandwidth is for SDRAM, DDR SDRAM, and RDRAM. I quote from the following page:
(Emphasis is mine.) The next paragraph reads:
Finally, An article from Real World Tech explains just what the timings are like, why they occur, and why they mean that DDR SDRAM is going to be faster for the forseeable future. A very instructive paragraph on the general problems with RDRAM follows:
So let's see, RAMBUS memory has higher latency, less bandwidth, consumes more power and therefore dissipates more heat, and it's more expensive. It basically sucks compared to DDR SDRAM in every way... Where's the plus side?
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"