Rambus Allowed to Continue Patent Dispute Case

ZuperDee writes "According to an article at Forbes, Rambus has just won a major victory against Hynix semiconductor. They have also signed a $75 million licensing deal with AMD." The victory? Well, come March they get to go to trial against Hynix.

Another yawner

[phavens]

Rambus lost big a long time ago. I always assumed if they sued enough companies they may win one. :\

Re:RAMBUS - Another company people love to hate.

[Stormwatch]

Keep in mind that, unlike SCO, Rambus may actually have a case. [wikipedia.org]

AMD

[PetriBORG]

While I guess I can get behind Rambus's right to sue, I'm not convinced that the whole submarine patent thing can be excused. Of course this forbes article is really light on the history of the cases other then to mention, "Rambus designs and licenses methods for moving data into, out of and between semiconductors."

I'm more interested/worried in the whole AMD part, I do not want to see AMD mobo's running with Rambus's insanely expensive memory on it.

Any more info on the deal with AMD?

[taskforce]

Does anyone have any more info on the 75MUSD deal with AMD? This is blind, uninformed speculation, but isn't RAMBUS's latest thing XDR memory? The stuff that runs at 3.2Ghz with a muliplier equal to the CPU's in the PS3?

If AMD were to net this stuff for Athlon64s it would explain why they've been holding out on DDR2 for so long and would also prompt me to run out and buy Quad Opertons with XDR very quickly; memory bandwidth seems to be the greatest hurdle on the otherwise extremely broadly equipped Athlon64 line; I can see how it would make a lot of sense to pair it up with HyperTransport.

AMD + Rambus Multicore

[adam31]

75$ mil deal and AMD gets access to "the good stuff". There's a reason XDR is going into the Cell processor, and it's because 25.6 GByte/s is the right bandwidth to feed 9 cores @ 3.2 Ghz. But it's way, way more than you need for a dinky 1- or 2-core processor (for those you're better off spending money on the super low-latency SRAM instead).

So does this mean that AMD is jumping on the many-multicore design bandwagon? They must have something up their sleeve...

Re:AMD

[sl3xd]

(I'm just an observer, but that's what slashdot is for...)

IIRC, the big beef with Rambus cost isn't the royalty; as you point out, it's a rather small amount.

The big deal for price with RDRAM was that the stuff was more significantly expensive to manufacture; this meant that while RDRAM didn't cost much more (to a consumer) than DDR memory, it was far less profitable to the memory manufacturers.

How does this play into DDR and SDRAM? If Rambus's patents are held as valid, Rambus gets to set terms for how DDR and SDRAM is licensed.

Don't underestimate the effect this can have. For the sake of playing the 'worst case' scenerio (which may or may not be the end result). To obtain a license for DDR and SDR patents, the memory company (ie. Samsung, Hynix, Micron, Infineon, etc.) must also agree to:
* License XDR (for another 3.5% fee).
* Agree to halt all R&D into new RAM technologies. (ie. give up all ability to develop their own technologies, essentially making it impossible to be self-sufficient again.)
* Agree to license every new whiz-bang buzzword Rambus develops, until the patent expires.
        * All of which in turn requires:
                * Licensing of all new Rambus IP
                * The licensee must halt (or continue to not participate in) R&D into new RAM technologies.
* Increase the price of all competing (non-RDRAM, non-XDR) memory by 300%. (Or even halt all production of competing products entirely.)

The thing to remember about a patent: it is a government granted monopoly. Price-fixing, gouging, and anti-trust do not apply. The patent holder gets to set all terms on the manufacture, and sale of the technology the patent covers, reguardless of how arbitrary, unreasonable, and selective those terms may be. It would be perfectly within their rights to allow any maker but brand 'X' to license their IP.

The broadness of what a patent holder can demand of its licensees is the reason why Hynix, Micron, Infineon, and others have been fighting these patents; the patents literally give Rambus the ability to force any (and all) of them out of business entirely. It's espescially galling to these companies as they feel that Rambus illegally took technologies that were discussed (and developed, depending on whom you talk to) at JEDEC, and patented work and ideas originating from JEDEC (and not Rambus). Hynix, Micron, and Infineon feel much like their IP was stolen, and some thug came pounding on their door for 'protection money.'

Whether that is actually the case may never be known; Rambus destroyed many of the documents that are involved (the article itself states that the Judge ruled that it was OK for Rambus to do this, because the documents were not intentionally destroyed to avoid incrimination in a lawsuit.) I'm quite certain similar things can be said of the various RAM makers; there are no white knights in this tale. The courts will continue to plod through the various cases, and eventually decisions will be made as to whether the patents are valid (and legal) or not.

Regardless of the outcome, a great many people feel that Rambus's patents were obtained in a dishonest way, and feel Rambus is going to be similarly dishonest in how the patents are licenced should they be deemed enforcable.

I've got to say, though, it's looking like the patents will have expired before the litigation ends... the issue is far from resolution, and the patents are about halfway through their lifetimes.