If you want an URL to read the story, go here. The same story can be found at Cnet.com and in a Rambus press release.
All you need to know is that Hitachi conceded defeat in the lawsuit Rambus launched against it and will pay Rambus royalties on SDRAM and DDR also.
Now Hitachi likely looks upon this as a way to get rid of litigation so it can sell its memory business to NEC, but it greatly increases the likelihood that the other memory manufacturers will follow suit.
However, it’s not the death of DDR. It’s not good news, but it’s far from fatal.
What next?
The other memory companies have two choices, cave in now and start paying royalties on all the SDRAM, DDR on top of the RDRAM they make, or not cave in and face the certainty of being sued by Rambus.
Will they? Like the old Sherlock Holmes tale, this will probably be a case of the dog not barking. If you read about the memory companies proclaiming defiance, then there will be a fight. If there is silence, then they’re probably negotiating with Rambus right now.
Is this bad for DDR?
Yes, but there’s a difference between bad and fatal.
Rambus is seeking and getting a higher royalty for DDR than for RDRAM. That’s a decent discouragement for making DDR and an encouragement to make RDRAM instead.
However, if you have a server, you don’t want RDRAM. There, you have BIG latency problems. For them, the issue isn’t DDR vs. RDRAM; it’s DDR vs. current SDRAM. The issue now becomes: will memory manufacturers continue to make DDR for servers no matter what happens on the desktop, or will the server folks be satisfied in the near term with current SDRAM?
Also keep in mind that if everyone caves in, Rambus doesn’t care about DDR anymore. It will get paid no matter what.
Does this hurt AMD?
If the memory manufacturers start rolling over and playing dead, maybe.
If memory and manufacturers commit to DDR for servers, AMD is safe. If they don’t, AMD has trouble. While AMD has a Rambus license (don’t know if Via does), it would take time for them to come up with RDRAM motherboards if it proves necessary.
The most likely outcome will be AMD starting off with DDR, while putting their RDRAM contingencies into gear. It’s not helpful, but it’s no killer.
Does this help Intel?
Yes. If the memory guys roll over, it will be much easier to persuade them to start shifting production over to RDRAM.
Does this mean I have to buy RDRAM now?
No. What it means is that the likelihood you’ll buy RDRAM for a new system six or twelve months from now has risen. However, DDR will probably survive, and it’s likely we’ll see a memory split:
RDRAM for Intel; DDR for AMD.
Be the first to comment