Slashdot Mirror
Apple, Others Hit With Lawsuit On Ethernet Patents
bth nods an AppleInsider story on a patent troll who has gotten hold of fundamental Ethernet patents and is wielding them broadly. Three guesses which US Appeals Court the lawsuit was filed in. "A Texas company has targeted a number of technology companies, including Apple, in a new lawsuit regarding a handful of computer networking patents issued in the 1990s. ... 3Com Corporation was granted four patents from 1994 to 1998 pertaining to network adapters. Two deal with the automatic initiation of data transmission, and one addresses 'host indication optimization.' ... The company's Web site states that U.S. Ethernet Innovations was founded 'to continue 3Com Corporation's successful licensing program related to a portfolio of foundational patents in Ethernet technology.' A press release from the company states that it is the 'owner of the fundamental Ethernet technology developed and sold by 3Com Corporation in the 1990s,' suggesting it purchased the patents. ... In addition to Apple, the lawsuit names Acer, ASUS, Dell, Fujitsu, Gateway, Hewlett Packard, Sony, and Toshiba as defendants."
Seems to me that asking for a trial by jury may very well backfire on them.
Token ring, here we come
Three guesses which US Appeals Court the lawsuit was filed in.
None! You cannot originate a patent infringement suit in a United States Court of Appeals, any more than you can file in the Supreme Court. Instead, patent litigation must start at a United States District Court. The losing party may appeal to the United States Court of Appeals for the Federal Circuit.
Everyone uses the _internet_, now. Who cares about ethernet?!
Now I suppose it's back to those pain-in-the-ass coaxial cables.
How difficult will it be to find a judge and jury whose sole access to the Internet is through dial-up, and whose workplace involves no computer networking whatsoever?
I seems to me that by waiting until this late in the game, to the point which nearly the entire world's Internet and telecommunications infrastructure is based on Ethernet in on incarnation or another is just plain sleezy.
The fact that 3Com, once a reputable company of top notch networking technology appears to be trying make money by exercising their patent pool through a 3rd party to raise much needed money. This is sad.
There was a point when the 3Com 3c509 and 3c905 ethernet adapters dominated the Ethernet world. In fact, while their cards were more expensive and more complex than nearly any other on the market, they were likely to be found in nearly every PC that was built of quality parts (meaning machines that chose ASUS motherboards over some fly by night).
The integration of Ethernet logic within chipsets pretty much destroyed the 3Com business model, after all, 3Com made more off the adapters than anything else. Today however nearly every high end motherboard I encounter implements a Marvell Ethernet PHY. Intel is selling tons of Ethernet PHY's to embedded vendors that implement their designs on FPGAs (meaning most high end rack based devices). 3Com is nowhere to be seen.
I have been patiently waiting for 3Com to come back and start taking the high end workstation and the server market seriously. I've been waiting for them to make great products again. Instead, they keep shoveling out lower and lower interest items. The trust the world once associated with their quality is decreasing so rapidly, soon people will see them as no better than Linksys or D-Link.
If I were 3Com, I'd seriously consider taking a project like Vyatta or the likes, start developing it into a high end system capable of managing switching (Layer-2 to Layer-4) and then build every product starting with their cheapest routers based on it. I'd start producing new silicon with high end features like TCP and UDP offloading and trying to get into the mass market PHY business. Most importantly, I'd start selling trust.
The problem is, by exercising these patents which people knew about but trusted 3Com to never exercise since it would just force all the other vendors on the market to lash back at them with their huge patent pools, they're destroying the last remaining bit of trust which was for a long time the only the 3Com had left.
Rest In Peace 3Com, it's unfortunate that everything I ever loved about you is gone. Another great innovator has died. Now you can sleep eternally in a grave dug next to SCOs
I wonder why the plaintiff is not suing some obvious companies. Cisco would be an obvious candidate -- and they have deep pockets. But all of the current defendants don't actually make Ethernet equipment. They buy Ethernet chipsets from companies like Intel, Broadcom, AMD, Marvell, VIA, and NVIDIA, among many others. Why isn't the plaintiff suing them?
Legitimacy of the patents aside, I wonder why an Ethernet technology suit would be leveled against companies that do little more than assemble circuit boards.
Because those companies have money and make some indirect use of the technology. They'd probably sue Coke if they could find a networked vending machine.
Amazing that Slashdot still can't master the technology of the "spellcheck", which I had in WordStar in 1987.
Maybe he was not terminated, but instead became himself an Ethernet terminator.
<sarcasm>It patents the idea of putting a memory buffer on the network card. Who would have thought of that?</sarcasm>
I'm starting to think that one of the biggest problems with patents is being able to sell them and hold them with out making products based on them. Well, the US patent system has a lot more problems like obvious things being patented and being able to patent business processes.
But apart from the US,some of the biggest problems come up in companies buying patents and being patent trolls. Patents were supposed to protect the inventor, selling the patent isn't protecting the inventor anymore. Also unless a company merges with another company I think all patents that the company owns should dissolve with the company and be unpatentable.
If you have a patent, I think you should be required to have products out using that patent, or at least working on making products with that patent. Too many companies patent something that they heard someone else speak about, they have no plans to use the patent, just profit off of someone else's work by beating them to the patent office, or just plain having the money to buy the patent where the person doesn't. I guess that still wouldn't quite solve the problem. There would need to be some process where you proved that you created the idea.
*start rant* Now as for the US Patent system, there is an official report that calls it too much of a joke for us to merge our patent system with their's. Our company started looking at US Patents and as far as we can tell, you can't write a line of code with out violating a patent. It is so silly that "If...else" is patented. "ifelse" is patented. "Begin...End" is patented. I think you can get away with "{...}" blocks, but not much else.
Microsoft, Apple, Google, Amazon what's the difference? All steal money from devs and control with walled gardens.
It varies by type, but the standard appears to be 20 years. This patent was filed in 1990.
US patents used to last 17 years from date GRANTED, and were kept secret until granted. The company requesting the patent could deliberately delay the process of getting it granted for years, so that it remained secret and was valid until later (when the market for the covered "invention" is expetcted to be bigger). With the current system, there is less secrecy and patents are valid for 20 years from date of filing. I checked one of the 4 patents in the article. It was filed in 1992 and granted in 1994, so it should be valid until 1994+17=2011. Still quite a while to go without ethernet.
From 1994 to 1997 I worked for a company that worked with 3Com writing device drivers for the 3Com adapters that seem to be based on these DMA patents. My vague recollection is that we spent at least 2 years working with at least 3 different versions of 3Com's EISA and PCI ethernet adapters and as far as I know the full DMA mode never worked 100% correctly on any of them. The different versions of their controllers had different bugs. Some would lockup, some would drop interrupts, or randomly stop DMA mode, or corrupt buffers, etc. I think in most cases the workaround suggestion from 3Com was to periodically poll the adapter status and if you detect it's wedged force a full reset and completely reinitialize the adapter. Of course doing stuff like that totally destroys system performance so we simply ignored the DMA capabilities and did PIO in the interrupt handlers to transfer packets to/from the adapters 1-2K buffers (one packet per interrupt). I might be wrong but I think it was at least 1998 and their 4th generation PCI-ethernet adapter before 3Com had a PCI-ethernet design that did DMA to/from ring buffers correctly and by that time the rest of the world was all shipping products based on $10 dnet-clones from 6 or 7 companies in Taiwan (and 3Com's design had morphed into a something that looked a lot like dnet-clone but with a lot of extra, useless features).
I was excited recently (OK, I'm a language geek) to learn that there's a "real" word for patent trolling: champerty (http://wordsmith.org/words/champerty.html)
Is that little rule - I before E except after C - still actually being taught?
Besides 'weird', here's a collection of other words that do not fit that rule:
e-ism (atheism), e-ing (being), e-ish (blueish), etc.
interrupted: albeit
sounds like 'A': airfreight, aweigh, beige, bobsleigh, heinous (heehee, he said heinous), heir
sounds like 'I': abseil, feisty
silent: forfeit
Even if you take the modified form of "I Before E except after C when it sounds like 'E'"...
caffeine, casein, codeine
Even if you modify it further "I before E except after C when it sounds like 'E' and isn't some chemical compound"...
Ashleigh, Keith, O'Neill (two "L"s, the one with a sense of humor)
So let's make it "I before E except after C when it sounds like 'E' and isn't a name like that of a person, object, or substance.."
ceiling, conceit, conceive, deceit, deceive, obeisance
I haven't even paged through my dictionary file beyond the letter 'O' yet and already I think you can tell that the "I before E" rule is an abomination. Not in the least due to all of the exceptions made above where you'd have to come up with new rules, or just teach the kids the correct spelling of the word 'as is'. I sure hope "I before E" is not still being taught.
Oh hell, I just noticed there's a Wikipedia topic for it - of course there is. *mutter* - and at least the Brits clued up; "In June 2009, the British government advised primary school teachers to stop teaching the rule."
http://en.wikipedia.org/wiki/I_before_E_except_after_C
Naaa, the biggest problems are pretty much the following:
1)There are no effective checks to stop people patenting obvious things or things that are not patentable. Combined with a complete lack of penalties (at least in practice ) for abusing the system and the expense involved with defending yourself against a lawsuit this allows patent trolls to cause a great deal of harm to companies and individuals who have done nothing wrong.
2)There is in practice a complete lack of punishment for deliberately filing invalid patents and patent claims.
3)Patent law is to a very large extent not based on any form of independent analysis of its consequences but rather the work of lobbying by special interest groups. In otehr words, patent law is designed to be profitable, not just.
4)Because patent law allows very vague and broad interpretations of patent claims, and because you can be found to be infringing even if you had never heard of a plaintiff's patent, it is in practice impossible to market ANY product without infringing on SOME patent.
5)Because large plaintiffs can essentially force smaller companies to settle by simply dragging a case along, the outcome of a lawsuit is often determined not by who is in the right, but who has the most money to spend on legal battles ( this is a more general problem with the US legal system ).
Or simply put: Due to intense lobbying by patent holders and existing monopolies the system is more or less designed to allow plaintiffs to abuse the system for purposes different from the original motivation of promoting arts and sciences. There is little justice or balance in the system and patent lawsuits mostly boil down to who has the deepest pockets rather than who is in the right.
First - someone please notify Al Gore that 3Com is claiming patent rights over the internet HE invented. I'm sure a movie will be forthcoming.
Second - and slightly more seriously - what about RFC 826 (http://www.faqs.org/rfcs/rfc826.html) regarding ethernet? Oddly enough 3Com isn't a part of that document and it clearly says "...This protocol was originally designed for the DEC/Intel/Xerox
10Mbit Ethernet..." and was published in 1982. Here are the other ethernet related RFCs (http://www.faqs.org/rfcs/np.html#ETH).
Now there IS a patent for the original ethernet standard issued in 1977 to Dr. Metcalfe, et al on behalf of Xerox (do we hear echos of PARC?) - http://www.google.com/patents?vid=USPAT4063220