UNITED STATES PATENT AND TRADEMARK OFFICE
`_________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_________________
`
`INTEL CORPORATION, CAVIUM, LLC, DELL INC., and
`WISTRON CORP.
`PETITIONER,
`
`VS.
`
`ALACRITECH, INC.
`PATENT OWNER.
`_________________
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`Patent 7237036
`_________________
`
`Record of Oral Hearing
`Held: September 13, 2018
`_________________
`
`
`
`Before STEPHEN C. SIU, DANIEL N. FISHMAN, and CHARLES J.
`BOUDREAU, Administrative Patents Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`_________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_________________
`
`INTEL CORPORATION, CAVIUM, LLC, DELL INC., and
`WISTRON CORP.
`PETITIONER,
`
`VS.
`
`ALACRITECH, INC.
`PATENT OWNER.
`_________________
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`Patent 7237036
`_________________
`
`Record of Oral Hearing
`Held: September 13, 2018
`_________________
`
`
`
`Before STEPHEN C. SIU, DANIEL N. FISHMAN, and CHARLES J.
`BOUDREAU, Administrative Patents Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`
`APPEARANCES:
`
`ON BEHALF OF PETITIONER INTEL CORPORATION:
` GARLAND T. STEPHENS, ESQUIRE
` ANNE M. CAPPELLA, ESQUIRE
` MELISSA HOTZE, ESQUIRE
` WEIL GOSTHAL & MANGES
` 700 Louisiana
` Houston, Texas 77002
`
` CHRIS KYRIACOU, ESQUIRE
` INTEL CORPORATION
`
`ON BEHALF OF PETITIONER WISTRON CORP.:
` ERIK HALVERSON, ESQUIRE
` BENJAMIN WEED, ESQUIRE
` K&L GATES
`
`ON BEHALF OF PETITIONER CAVIUM, LLC:
` DAVID XUE, ESQUIRE
`
` KARINEH KHACHATOURIAN, ESQUIRE
` RIMÔN LAW
`
`
`
`ALBERT HARNOIS, ESQUIRE
`CAVIUM, LLC.
`
`
`ON BEHALF OF PETITIONER DELL INC.:
` CHRIS DOUGLAS, ESQUIRE
` BRADY COX, ESQUIRE
` ALSTON & BIRD
`
` MARK VOGEL, ESQUIRE
` DELL INC.
`
`
`
`
`2
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`ON BEHALF OF PATENT OWNER:
` JIM GLASS, ESQUIRE
` BRIAN MACK, ESQUIRE
` SEAN LI, ESQUIRE
`
`ANTONIO SISTOS, ESQUIRE
` QUINN EMANUEL URQUHART & SULLIVAN
` 50 California Street, 22nd Floor
` San Francisco, California 94111
`
`
`
`
`
`
`The above-entitled matter came on for hearing on September 13,
`2018, commencing at 9:30 a.m., at the U.S. Patent and Trademark Office,
`Silicon Valley Regional Office, 26 South 4th Street, San Jose, California
`95112.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`3
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`
`P R O C E E D I N G S
`
` USHER NGUYEN: Calendar No. 179, Appeal Nos.
`IPR2017-01391, IPR2017-01392, IPR2017-01393, IPR2017-01405,
`IPR2017-01406, IPR2017-01409, IPR2017-01410.
` JUDGE BOUDREAU: Good morning.
` MR. MACK: Good morning, Your Honor.
` JUDGE BOUDREAU: I'm Judge Boudreau. We are joined by
`videoconference by Judges Fishman and Siu.
` Could we have counsel for each of the parties come up
`and enter an appearance and state your names for the record.
` MR. STEPHENS: Good morning, Your Honor. Garland
`Stephens of Weil, Gotshal and Manges representing Intel
`Corporation. With me today are my colleagues, Anne Cappella and
`Melissa Hotze, and in-house counsel for Intel, Chris Kyriacou.
` Oh, sorry.
` JUDGE FISHMAN: Excuse me; this is Judge Fishman. I
`can't hear anyone.
` MR. STEPHENS: Yes, I -- is that better? No?
` JUDGE FISHMAN: Yes.
` MR. STEPHENS: Okay. Can you hear me now?
` JUDGE FISHMAN: Yes.
` MR. STEPHENS: Okay. It's Garland Stephens of Weil,
`Gotshal & Manges representing Intel Corporation. With me today
`are my colleagues, Anne Cappella and Melissa Hotze, also from Weil
`Gotshal, and in-house counsel for Intel, Chris Kyriacou.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`4
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
` Also appearing for the other Petitioners who joined
`these IPRs are Eric Halverson of K&L Gates for Wistron; Ben Weed
`of K&L Gates for Wistron; Albert Harnois of Cavium, Inc.,
`in-house counsel; David Xue of Duane Morris for Cavium;
`Christopher Douglas of Alston & Bird for Dell; Brady Cox of Alston
`& Bird for Dell; Mark Vogel, in-house counsel for Dell; and
`Karineh Khachatourian, also representing Cavium.
` Thank you, Your Honors.
` JUDGE BOUDREAU: Thank you, Mr. Stephens. And will you
`be presenting the argument on behalf of --
` MR. STEPHENS: Yes, sir, I will. Thank you.
` JUDGE BOUDREAU: And for Patent Owner?
` MR. GLASS: Good morning, Your Honors. I'm Jim Glass
`for Patent Owner. With me today is Brian Mack, Joe Paunovich,
`Sean Li. I'm also going to say is, from Patent Owner is
`Larry Boucher and Arthur Trueger. I will not be
`presenting today; we have split up our argument with the three of
`our attorneys, Brian Mack, Joe Paunovich, and Antonio Sistos will
`be presenting. Actually, I'm sorry; I didn't introduce Antonio
`Sistos. I apologize.
` JUDGE BOUDREAU: Thank you.
` And before we begin, we have just a few housekeeping
`matters. Because of limitations on our audio technology, as we
`just saw when Mr. Stephens got up to present, Judges Fishman and
`Siu will only be able to hear you when you're actually standing at
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`5
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`the podium speaking into the microphone and the microphone is on,
`so we ask that you just come to the podium whenever you're
`speaking.
` We also remind the parties that this hearing is open to
`the public, and a full transcript of it will become part of the
`record.
` As you know from the oral hearing order that we
`entered on September 5th, Intel and Alacritech will each have 90
`minutes to present their arguments in total for the seven cases.
`Because the Petitioner has the burden of showing patentability of
`the claims at issue, Intel will proceed first, followed by
`Alacritech. Intel and Alacritech each filed motions to exclude
`evidence, and Alacritech also filed contingent motions to amend.
`So the parties may also, if they choose to do so, address their
`respective motions during their principal arguments and the
`opposing party will have the opportunity to respond during their
`response.
` Intel may reserve some, but not more than half, of its
`argument time for rebuttal to Alacritech's presentation, and
`Alacritech, likewise, under our new revisions to the trial
`practice guide, may reserve some of its argument time for
`sur-rebuttal, limited to addressing Intel's rebuttal.
` Although we have allotted 90 minutes to each side based
`on the considerable overlap and the issues among the cases being
`discussed, if we end up running over as a result of our asking a
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`6
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`lot of questions from the bench, we'll take that into
`consideration.
` Beyond that, we'll leave it up to the parties to decide
`how they divide up their respective time among the seven cases,
`although we do note that there's considerable overlap among the
`cases. So we'd just ask that the parties clarify for the record
`which case or particular cases that a given argument relates to if
`it isn't immediately apparent from the discussion.
` Fortunately, it appears from both parties' demonstrative
`exhibits in this case that each of you intend to present in
`essentially in the same order, beginning with the 1391, 1406, and
`1392 cases, and ending with the 1393. I think the only exception
`to that was in the respective ordering of the 1405 versus the 1409
`and 1410 cases.
` So if you just want to continue to present in the order
`that you have presented in the demonstratives, that will probably
`be most efficient for the record, but I'll leave that up to you.
` While we're on the topic of demonstratives, we also note
`that the most recent updates for trial practice guide require that
`demonstrative exhibits be clearly marked as being merely aids to
`oral argument and not as evidence. And we further note that
`neither party has included such an annotation on its demonstrative
`exhibits. So I'm just going to ask for verbal acknowledgment of
`your understanding that the demonstrative exhibits are not
`evidence in this case.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`7
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
` Mr. Stephens?
` MR. STEPHENS: Yes, Your Honor. We acknowledge that.
` JUDGE BOUDREAU: Thank you.
` And Mr. Glass or Mr. Mack?
` MR. GLASS: Yes, Your Honor, we acknowledge.
` JUDGE BOUDREAU: Okay. Thank you.
` Also, with respect to the demonstrative exhibits and as
`also explained in our oral hearing order, we ask that you please
`keep in mind that if you use any slides, whatever is projected on
`the Elmo or on the slide projector here in San Jose won't be
`visible to Judges Siu and Fishman, appearing remotely, or by
`anyone reading the transcript, for that matter.
` We have electronic copies of all the demonstrative
`slides and all the papers in the case, so if you're using the
`screen and you're referring to a specific slide or figure or any
`document for the record, we just ask that you call out the slide
`number, figure number, or exhibit or page number as appropriate so
`that the record is complete both for the transcript and so that the
`judges appearing remotely can follow along.
` Finally, I have been informed -- and you may have seen a
`sign in the lobby -- that there's going to be an earthquake
`preparedness drill in just about 20 minutes, at approximately 10
`a.m. local time. My understanding is that the drill will consist of
`just two public announcements, approximately 30 seconds apart
`from each other. I think the first one will say to duck and cover
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`8
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`underneath the desk. There won't be any need to evacuate as far
`as I know, but I do anticipate that we'll go off the record for
`about a minute or so at the beginning of the first announcement
`and then go back on the record after the second announcement.
` So with that, unless there are any questions from either
`party before we begin, we'll proceed. Mr. Stephens?
` MR. STEPHENS: Thank you, Your Honor. Garland Stephens
`representing Intel Corporation.
` JUDGE BOUDREAU: And just before you begin, how much
`time, if any, do you wish to reserve for rebuttal?
` MR. STEPHENS: Thank you for that reminder. 40 minutes
`we'd like to reserve, Your Honor.
` JUDGE BOUDREAU: 40 minutes.
` MR. STEPHENS: Yes.
` JUDGE BOUDREAU: And we don't have any timer displayed
`in this room, but I'll be keeping track of the time and can give
`you a reminder if you'd like, say, 15 minutes before your
`principal argument.
` MR. STEPHENS: Or 10 would be ideal.
` JUDGE BOUDREAU: Ten minutes.
` MR. STEPHENS: If you don't mind. Thank you, Your
`Honor.
` ARGUMENT
` MR. STEPHENS: I'd like to start with a couple of
`general observations before I dive into our voluminous slides and
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`9
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`the specifics that are there in. And the first of those is that,
`for many of the key points of alleged differentiation over the
`prior art, the Patent Owner relies entirely on the declaration of
`their expert, Dr. Kevin Almeroth. And if you look at many of
`those citations to his declaration, it is verbatim what's in their
`briefing, with no additional citation of evidence whatsoever.
` And I can give you some specific examples. So, for
`example, Dr. Almeroth says every received packet requires an
`interrupt at every protocol layer, and that the Erickson reference
`involves many undisclosed interrupts. He doesn't cite any actual
`evidence for any of those, just his opinion, which mirrors exactly
`what's in the briefing.
` And it's also, in that particular instance and in
`others, it's directly contradicted by what's in the evidence. In
`that particular case -- if we could have slide 143, please --
` JUDGE BOUDREAU: Well, I'm aware of that, Mr. Stephens.
`And I'm also aware that, in some places in Dr. Almeroth's
`declaration, he explicitly said, "As explained by Dr. Almeroth" --
` MR. STEPHENS: Yes, Your Honor.
` JUDGE BOUDREAU: -- such as paragraph 110 of the
`declaration.
` But is that a basis for excluding the declaration or is
`that only a basis for giving limited weight --
` MR. STEPHENS: Well, it's certainly a basis for giving
`it little weight. We did move to exclude the portions that are
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`10
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`identical because we were precluded from asking questions in
`deposition about which came first: The report or his expert
`report. So we did move to exclude that. I didn't propose to
`argue that particularly, and at a minimum, we think that his
`declaration should be given little weight, particularly when it
`cites no evidence whatsoever.
` Another sort of general point, I think, that's worth
`making -- and I know the Board is aware of this -- a persistent
`failing of the Patent Owner is that they attack individual
`references and not the proposed combination. I think we have been
`very clear in our petitions -- we have certainly tried to -- to
`identify particularly the combinations that we're relying on and
`how the references relate, and many of the rebuttals that we have
`seen from the Patent Owner had not addressed the actual
`combination.
` So, with that, I propose to talk at a high level about
`the references themselves and those combinations to ensure that we
`have conveyed as accurately as we can what the combinations are we
`are relying on, starting with Tanenbaum and Erickson. Now, the
`combination of Tanenbaum and Erickson we're relying on for the
`'036, the '072, and the '241 patent, all of the transmit claims of
`the '241 patent, there are some receive claims for which we also
`rely on Alteon, which I'll get to later.
` If we could have slide 10, please. So as we have
`explained in the petition, Erickson is a prior art patent that
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`11
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`describes fast and slow protocol processing on an outboard network
`interface device.
` If we could have the next slide, please. Sorry; let me
`back up here.
` So the slow application uses a traditional protocol
`processing approach, a traditional TCP/IP and UDP/IP protocol
`stack on the host computer, and the fast application sends
`information directly through the I/O device through a commodity
`interface using virtual memory map hardware.
` It's very clear that, when it's talking about the slow
`path, that's normal streams processing. That's also shown in the
`figure there at 308 on the right. And the fast application
`bypasses that traditional TCP or UDP/IP protocol stack processing.
` And it also is very clear that it involves both receive
`and transmit. The preferred embodiment is described in more
`detail for transmit, but it clearly encompasses receipt. That's a
`persistent complaint of the Patent Owner, that it doesn't actually
`disclose receipt but it does, and I'll explain how it does.
` So we're at slide 13. And if it would help the judges
`that are remote, I also have a mapping into the PDF page numbers
`of the PDFs that we have submitted. I don't know if you're
`looking at these electronically. My colleague reminded me that
`the slide numbers don't necessarily match up with the PDF page
`numbers. And I can certainly identify the PDF page number if that
`would help.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`12
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
` JUDGE BOUDREAU: They appear to match up --
` JUDGE FISHMAN: So far, I have seen them match in slide
`number and PDF page number.
` MR. STEPHENS: I think once we get through part 1, that
`will no longer be true. So if they start to diverge, please let
`me know, and I'm happy to announce the PDF page numbers as well.
` JUDGE BOUDREAU: It will be fine just to use the
`demonstrative number.
` MR. STEPHENS: Okay. We'll do that then.
` So slide 13 of our deck shows -- we produced some of the
`text from Erickson who describes how the adapter stores protocol
`scripts and data for moving data. So on the adapter -- the
`adapter memory shown as 512 in the figure -- includes an end point
`table that points to various protocol data also stored in the
`adapter memory. And that is for transferring data that's received
`directly to applications in a host memory and also transferring
`data from host memory directly onto the wire.
` So it's off-loading the protocol processing in both
`directions. And it does so by executing scripts on the adapter.
`It's very clear it's -- this is just one of a number of citations,
`which I think the Board is aware of, that show that there is a
`processor on the adapter that executes these scripts to perform
`the protocol processing that's described.
` The script has also passed an appropriate datagram
`template based on the specific software register, and there are
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`13
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`different scripts for different types of datagrams, and that
`script is pre-negotiated between the host computer and the adapter
`to include almost everything about the datagram that's going to be
`transmitted. This is slide 16; I'll try to remember to announce
`the slide. So far, I have been going in order.
` The adapter uses scripts for various protocols; I think
`I already mentioned this. It specifically says that there is a
`protocol script for TCP/IP. Now, the one that's described in
`detail and actually laid out in pseudocode is for UDP, but it
`clearly discloses that it -- the intention of the invention is to
`have scripts for multiple protocols, explicitly including TCP/IP.
` And as a reference for the interested user, person of
`skill in the art, it points them to Tanenbaum's computer networks.
`Now, this is an earlier version than the one we're relying on.
`We're relying on the 1996 edition because, of course, the priority
`date for the patent is in 1997, and a person of ordinary skill in
`the art would have looked for the latest version of Tanenbaum at
`the time.
` Obviously, network technology advances, and you wouldn't
`want to get an older edition in a newer one was available.
` JUDGE BOUDREAU: Mr. Stephens, maybe you don't know the
`answer to this, but does the 1981 edition of Tanenbaum have all of
`the disclosure in it that you're relying on from the 1996 edition?
` MR. STEPHENS: It does not. It does not. In fact, I
`think most of the key pieces that we're relying on are not present
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`14
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`in that older edition. The older edition was more focused on OSI
`and less on TCP/IP. It did have both, but there was more
`disclosure on OSI and less on TCP/IP. So it is missing, for
`example, the fast TPDU section that we rely on.
` JUDGE BOUDREAU: So, in that case, if a person of
`ordinary skill in the art seeing the reference to TCP/IP and
`Erickson and the reference to Tanenbaum in the Erickson reference,
`what would they have been expected to get from the 1981 edition
`that they incorporate by reference?
` MR. STEPHENS: Well, for example, what the patent points
`you to in part is for the form and structure of TCP sockets and
`packets. And, of course, the -- both editions have some
`description of TCP/IP sockets and packets, and I don't recall
`whether the earlier version has as much detail about the contents
`of, for example, the TCP headers. That's certainly laid out in
`detail in the later version.
` If the later version was a popular college textbook --
`and an important fact is that the Patent Owner's expert used it to
`teach a class before the priority date that the patent ensued. So
`there's an admission by the Patent Owner's expert that he, as a
`college professor, used it to teach an undergraduate class in
`September of 1997.
` So I think there's no genuine dispute (indiscernible) as
`public availability to persons of skill in the art. And if you
`think about how a person reading Erickson in '97 might approach
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`15
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`it, you'd see the reference to Tanenbaum and you might very well
`either look in the library or go on Amazon.com and order the
`current version. You're not so likely, I wouldn't think, to look
`at a version several editions old.
` The Patent Owner's patents themselves also acknowledge
`that Tanenbaum is a college-level textbook. This is slide 20, I
`believe. Let's see here.
` So let's talk in a little more detail about Tanenbaum
`and how it relates to Erickson. So Tanenbaum describes a fast
`path, very much like Erickson's fast protocol processing and also
`very much like Erickson's fast protocol processing, it uses a
`prototype header. In Erickson, it's called a template header, but
`in Tanenbaum, it's called a prototype header. This disclosure is
`not in the 1981 version. This is only in the 1996 version, I
`believe.
` JUDGE BOUDREAU: And this is the disclosure on slide 21?
` MR. STEPHENS: Sorry; thank you. Slide 21. I'll try to
`get better with that.
` And in the particular approach that's described here as
`one we'll see in other references, it was very well-known at the
`time called header prediction. It was something developed by a
`man named Van Jacobson and was actually widely built into TCP/IP
`implementations at the time, including the very popular BSD UNIX
`version, which was developed at Berkeley, here in the Bay Area.
` Tanenbaum points out that while TCP/IP requires a
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`16
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`sequence of special packets -- this is on slide 22 -- to get into
`the established state, which is the state in which TCP/IP and
`other transport protocols transfer data as opposed to connection
`setup and other housekeeping packets, processing is
`straightforward once you're in that established state.
` So the fast path that's described in Tanenbaum is
`specific to the established state. So it talks about how the
`packet headers change very little from packet to packet once
`you're in the established state.
` It also specifically contemplates and acknowledged that
`it was known that the transport entity could be on a network
`adapter card or a programmed co-processor chip. And very much
`like Erickson, Tanenbaum's fast path reuses a prototype header,
`and it spells out the contents of the prototype header and
`explains that the fields are the same between consecutive TPDUs
`when you're in the established state.
` So the reason it's called the fast path and the reason
`it saves so much processing time is only a very small number of
`fields that we see there unshaded in the template header here
`change from packet to packet. The rest are all the same, so you
`don't have to actually process them.
` Slide 25 explains exactly how the combination would be
`made, we believe, by a person of ordinary skill in the art. They
`would see the reference to Tanenbaum for how TCP works. They
`would see the reference in Erickson to the TCP script. They would
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`17
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`see the UDP template header in Erickson and say, Well, let me take
`a look at Tanenbaum and figure out how to adapt this to work with
`the TCP/IP script. I want to write the TCP/IP script that
`Erickson says this invention supports. I'll look at Erickson --
`excuse me, I'll look at Tanenbaum to figure out how to do that.
` And Tanenbaum explains exactly how to do that. You take
`the prototype header that we just looked at, and you put the TCP
`header in place of the UDP header, and the IP header matches up to
`the IP header that's already there, and then, of course, you have
`a MAC header as well in Erickson's template. So it tells you
`exactly the changes that you would need to make to offload the
`fast path that's described in Tanenbaum.
` JUDGE BOUDREAU: Now, there's evidence of record that
`UDP is a connectionless protocol whereas TCP is connection
`oriented. How would you address the argument that this wouldn't
`just be a simple substitution of TCP for UDP?
` MR. STEPHENS: Well, it's true that UDP is
`connectionless and that TCP is connection-oriented, but that's the
`point of the fast path. The fast TPD processing says that that
`portion of the processing is simple. So that's how a person of
`ordinary skill would address it. Rather than trying to offload
`the entire connection setup, for example, before you get to the
`established state, a person of skill in the art would say, well,
`let's do the part that actually engages most of the processing --
`protocol processing. Right? I mean, one of the teachings of
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`18
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`Tanenbaum and the Van Jacobson header prediction algorithm is that
`most connections spend most of their cycles -- processor cycles in
`the established state, because that's the state in which they
`transfer data.
` So the focus would naturally be when you're trying to
`offload and reduce host CPU utilization on offloading the part
`that will do you the most benefit, and that part is, actually,
`straightforward. Let's see. Yeah, that's the part that Tanenbaum
`explicitly teaches as straightforward.
` So the complexity is in the setup. It's not in the
`actual fast path processing, and that's what we think would
`motivate a person of ordinary skill to turn to the fast TPDU
`header prediction section of Tanenbaum -- and I've referenced to
`slide 22 -- in place of the header template disclosure and the UDP
`script that's shown in Erickson.
` And this, Tanenbaum has explicit disclosure on how this
`works in both directions. So it talks about using a template very
`much like Erickson's for transmit, and it talks about using a fast
`path TCP procedure for receiving packets where you look up a
`connection based on the send and receive addresses and ports for
`TCP and IP, and you identify a connection record by doing it,
`determine whether it's eligible for the fast path, and then call
`the fast path procedure and update the connection record.
` And Tanenbaum also teaches that the connection record is
`whether all of the TCP state is stored for that particular
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`19
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`connection.
` And this is the part that I was referring to a moment
`ago: The connection record can be stored in --
` JUDGE FISHMAN: Excuse me, counselor. This is Judge
`Fishman.
` MR. STEPHENS: Yes.
` JUDGE FISHMAN: Is this connection record, is that
`analogous to the context that's recited in (indiscernible)?
` MR. STEPHENS: Exactly, Your Honor. This connection
`record is the context -- we believe it maps on to the context of
`the claims. It's also a very well-known feature --
` JUDGE BOUDREAU: Oh, this might be the earthquake drill.
`We'll just go off the record for a minute.
` (A brief recess was taken.)
` MR. STEPHENS: Shall I proceed, Your Honor?
` JUDGE BOUDREAU: Yes, please.
` I believe, Judge Fishman, were you in the middle of a
`question?
` JUDGE FISHMAN: No, I think I have it. Thank you.
` MR. STEPHENS: Okay. Returning to slide 28, which I
`have now, again, we see the TCP connection record can be stored in
`a hash table, which is accessed using some simple function of the
`two IP addresses and the two port numbers from the TCP header and
`the IP header. And what we see is that's very much analogous to
`the end point table in Erickson.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`20
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
` So Erickson describes in the adapter an end point table
`that has entries that contain various protocol data, and that's in
`the memory 512, in the adapter memory. And it also contains
`defined protocol scripts, which are shown as 516, again, in the
`adapter memory, and it uses those to transfer data from the
`memory of the I/O device adapter to the portions of main memory.
`So we have a clear description in Erickson of storing protocol
`information that corresponds directly to the connection record --
`the TCP connection record in the fast path description of
`Tanenbaum on the adapter and using that protocol information in
`scripts that run on the adapter to transfer data from the I/O
`device that enter directly into host memory.
` So that is the same notion precisely as we see in
`Erickson, but Erickson describes it in more detail for TCP
`specifically.
` JUDGE BOUDREAU: What evidence is there, Mr. Stephens,
`that the scripts run on the adapter as opposed to running on the
`host computer?
` MR. STEPHENS: It says so specifically in a number of
`cases. Here, it shows that they're stored on the adapter, for one
`thing, and then -- back up here -- when I say "here," I mean slide
`29, which is the one I was just talking about. If I can get my
`clicker to work here. There we go.
` And then -- let me see here. If we can go to slide 14,
`please. So it says explicitly that the script is executed on the
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`21
`
`
`
`
`Cases IPR2017-01391, IPR2017-01392, IPR2017-01393,
`IPR2017-01405, IPR2017-01406, IPR2017-01409, IPR2017-01410
`
`I/O device adapter. This is just one of the number of citations
`we have identified in our papers -- this is slide 14 again --
`where it says a script is pr
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
