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."
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.
Amazing that Slashdot still can't master the technology of the "spellcheck", which I had in WordStar in 1987.
There doesn't appear to be anything about this patent that is Ethernet specific. The claims appear general enough to apply to a modern implementation of virtually any network technology.
A quick scan seems to indicate that virtually any network adapter that directly accesses transmit descriptors in host memory or writes packets into ring buffers in host memory (i.e. does DMA in any practical way on a packet by packet basis) violates the patent.
I believe that covers about every state of the art network adapter in existence. I am somewhat curious about whether there is prior art in the way IBM mainframes handle I/O. Anyone know enough to comment?
If that is the case then there is much prior art. For example, the UK's JANET had its origins around 1970 and was pretty mature around 1980. Is this another case of the US issuing a patent to somebody for something alaready in common use elsewhere?