`
` UNITED STATES PATENT AND TRADEMARK OFFICE
` BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`INTEL CORPORATION,
` Petitioner,
`vs. IPR20018-00234
`ALACRITECH,
` Patent Owner.
`------------------------------/
`
` DEPOSITION OF ROBERT HORST, Ph.D.
` Redwood Shores, California
` Tuesday, August 28, 2018
`
`Reported by:
`JANIS JENNINGS, CSR, CCRR
`Job No. 146788
`
`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
`
`INTEL EX. 1440.001
`
`
`
`Page 2
`
` DEPOSITION OF ROBERT HORST, Ph.D., taken on
`behalf of the Patent Owner, at Weil, Gotshal & Manges,
`201 Redwood Shores Parkway, Redwood Shores, California,
`beginning at 9:45 a.m. on Tuesday, August 28, 2018,
`before Janis Jennings, Certified Shorthand Reporter
`No. 3942, CLR, CCRR.
`
`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
`
`INTEL EX. 1440.002
`
`
`
`Page 3
`
`APPEARANCES:
`
` For the Patent Owner ALACRITECH:
` QUINN EMANUEL URQUHART & SULLIVAN
` 50 California Street
` San Francisco, California 94111
` BY: SEAN LI, Esq.
`
` For the Petitioner, INTEL CORPORATION:
` WEIL, GOTSHAL & MANGES
` 2001 M Street, NW
` Washington, D.C. 20036
` BY: SUTTON ANSLEY, ESQ.
`
` For the Petitioner, CAVIUM:
` DUANE MORRIS
` 2475 Hanover Street
` Palo Alto, California 94304
` BY: KARINEH KHACHATOURIAN, ESQ.
` DAVID XUE, ESQ.
`
`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
`
`INTEL EX. 1440.003
`
`
`
` I N D E X
`
`Page 4
`
`WITNESS EXAMINATION
`ROBERT HORST, Ph.D.
`
`BY MR. LI 6
`BY MR. ANSLEY 52
`
`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
`
`INTEL EX. 1440.004
`
`
`
` E X H I B I T S
`
`Page 5
`
`NUMBER DESCRIPTION PAGE
`Exhibit 1 Declaration of Robert Horst, Ph.D. in 13
` Support of Petition for Inter Partes
` Review of U.S. Patent No. 8,850,948
`
`Exhibit 2 Third Edition Computer Networks; 18
` Intel Ex. 1006.001 - 2006.594
`
`Exhibit 3 Article "A Reduced Operation Protocol 31
` Engine for a multiple-layer bypass
` Architecture;
` Intel Ex. 1015.001 - 1015.016
`
`Exhibit 4 TCP/IP Illustrated Volume; 46
` INTEL Ex. 1013.001 (excerpts)
`
`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
`
`INTEL EX. 1440.005
`
`
`
` REDWOOD SHORES, CALIFORNIA;
` TUESDAY, AUGUST 28, 2018; 9:45 A.M.
`
`Page 6
`
` MR. LI: This is Sean Li from Quinn Emanuel
`representing Alacritech.
` MR. ANSLEY: Oh, sorry. Sutton Ansley of
`Weil Gotshal & Manges here on behalf of Intel and
`the witness.
` MS. KHACHATOURIAN: Karineh Khachatourian
`and David Xue from Duane Morris on behalf of Cavium.
`
` ROBERT HORST, Ph.D.,
` the witness herein, was sworn and testified
` as follows
`
` EXAMINATION
`BY MR. LI:
` Q. Good morning, Dr. Horst.
` A. Good morning.
` Q. Nice to see you again.
` Could you state your name and address for
`the record.
` A. Robert Horst, 1182 Glen Avenue in San Jose.
` Q. Before I do our deposition, let me go
`through the ground rules so we're on the same page.
`
`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
`
`INTEL EX. 1440.006
`
`
`
`Page 7
` In this deposition I'm going to ask a lot of
`questions and the court reporter is going to record
`all the questions and answers. If we talk over each
`other, it will be hard for the court reporter to
`record the answers so I would suggest we do not talk
`over each other.
` Do you understand?
` A. Yes.
` Q. And the court reporter cannot transcribe
`non-audible responses; such as nods or, you know,
`gestures. Can you avoid doing that?
` A. Yes.
` Q. So during this deposition we can take a
`break every hour or so. Let me know if you need a
`break. But if there's a pending question, we have
`to finish the question before I give a break.
` Do you understand?
` A. Yes.
` Q. Are you taking any medication that could
`impede your ability to tell the truth?
` A. No.
` Q. Have you been deposed before?
` A. Yes.
` Q. How many times?
` A. Over a dozen. I don't know how many total.
`
`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
`
`INTEL EX. 1440.007
`
`
`
`Page 8
`
`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
`
` Q. How many times did you work with Intel?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: I'm not sure what you're
`asking. How many different cases or how many
`different...
`BY MR. LI:
` Q. Right. Let's start with how many different
`cases you worked for Intel.
` A. I believe there have been several, but I'm
`not sure how many total.
` Q. Okay. When was your first time you worked
`for Intel?
` A. I don't recall.
` Q. Was that in the 1990s?
` A. Are you asking about litigation on behalf of
`Intel or are you asking actually working for Intel?
` Q. I'm asking litigation.
` A. Okay. There have been some cases where
`Intel was a party and I was primarily retained by
`some others, so it's hard for me to remember which
`ones Intel was part of.
` Q. Did you work on any cases that -- you know,
`with Cavium before this case?
` A. Before this case, I've not worked with
`Cavium.
`
`INTEL EX. 1440.008
`
`
`
`Page 9
`
`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
`
` Q. What is your hourly rate?
` A. $550 per hour.
` Q. How much time have you spent on this case so
`far?
` A. I haven't added up the hours so I don't know
`for sure.
` Q. Has it been over 300 hours? Is it fair to
`say over 300 hours?
` A. I'm not sure what you mean by "this case."
`Again, there's been different proceedings here, so
`what are you referring to?
` Q. So let me ask you this: Do you have an
`engagement letter with Intel?
` A. I have an engagement letter with Weil on
`behalf of Intel.
` Q. Do you have an engagement letter with
`Cavium?
` A. I have an engagement letter with
`Duane Morris on behalf of Cavium, yeah.
` Q. Do you bill them separately --
` A. Yes.
` Q. -- for this petition? Specifically I'm
`talking about IPR2018-00234 -- -236?
` DEPOSITION REPORTER: It's 234.
`/ / /
`
`INTEL EX. 1440.009
`
`
`
`Page 10
`
`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
`
`BY MR. LI:
` Q. 234. For this petition working on the '948
`patent, how long did you bill to Intel?
` A. I don't divide my hours separately by which
`patents I'm working on so it's hard for me to know
`the total number for this one proceeding.
` Q. Okay. How often do you send a bill to
`Intel?
` A. Once a month.
` Q. When was the last time you sent a bill to
`Intel?
` A. It was in early August.
` Q. And how much was that?
` A. I don't remember the exact number, but it
`was only a few hours for last month.
` Q. What petition is that for?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: I don't remember which things
`I covered during the last month. I'm involved with
`a lot of different IPRs.
`BY MR. LI:
` Q. Did you prepare for this IPR -- I mean, did
`you prepare for this deposition?
` A. Yes.
` Q. When did you prepare?
`
`INTEL EX. 1440.010
`
`
`
`Page 11
` A. Over the last few days I read materials and
`met at the law firm.
` Q. How much time did you spend on the
`preparation?
` A. Yesterday I was here for a full day, and
`before that I spent several days with just a few
`hours per day preparing.
` Q. Did you meet with counsel on this case?
` A. Yes.
` Q. Who did you meet?
` A. I met with two of the attorneys from Weil
`and two of the attorneys from -- representing Cavium
`in this.
` Q. Did you meet with them together or
`separately?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: At different points of the day
`the attorneys came and went, so sometimes there were
`multiple and sometimes only one or two.
`BY MR. LI:
` Q. Is there any -- is there a certain point you
`were meeting with counsel for both Intel and Cavium?
` A. There were times when they were all in the
`room, yes.
` Q. Did you speak to any non-lawyers to prepare
`
`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
`
`INTEL EX. 1440.011
`
`
`
`Page 12
`
`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
`
`for this deposition?
` A. No.
` Q. Did you talk with Dr. Lin --
` A. No.
` Q. -- for this deposition?
` A. Not for this deposition.
` Q. Did you talk with Dr. Lin ever in this case?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: I was -- early on in this
`case. Dr. Lin and I were both at the law firm at
`the same time so I spoke with him briefly at that
`time.
`BY MR. LI:
` Q. When was that?
` A. I don't remember exactly when that was.
`It was early in the preparation of some of the IPRs.
` Q. Did you speak with any current or former
`Intel employee?
` MR. ANSLEY: Objection to form.
` THE WITNESS: No.
` MR. LI: I'm going to introduce the first
`exhibit, Exhibit No. 1.
` (Exhibit 1 marked for identification.)
`BY MR. LI:
` Q. Do you recognize this exhibit?
`
`INTEL EX. 1440.012
`
`
`
`Page 13
` A. Yes. This is my declaration in the '948.
` Q. Do you see on the lower right corner:
`"Intel Exhibit 1003"? Do you see that?
` A. Yes.
` Q. Do you know which petition this is related
`to?
` A. This is related to a petition on the '948.
`I don't recall the exact number of that petition.
` Q. Can you go to page 78 of this document.
` A. Yes, I'm there.
` Q. And do you see your signature there?
` A. Yes.
` Q. Is that your signature?
` A. Yes.
` Q. Do you see the date?
` A. Yes. April 18th, 2017.
` Q. When did you prepare this declaration?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: I prepared it in the time
`before this signature so it would have been in early
`2017.
`BY MR. LI:
` Q. Do you know when this declaration was filed
`in this case?
` A. I believe in this case the same declaration
`
`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
`
`INTEL EX. 1440.013
`
`
`
`Page 14
`
`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
`
`was filed in a later IPR.
` Q. It was filed after April 18th, 2017. Is
`that fair to say?
` MR. ANSLEY: Objection.
` THE WITNESS: I'm sorry. I couldn't hear
`you.
`BY MR. LI:
` Q. Is it fair to say this was filed after
`April 18th, 2017?
` A. I'm still not sure what you're referring to.
`What is filed?
` Q. Okay. Let me ask you this: Did you add new
`stuff to this declaration after April 18th, 2017?
` A. No. This declaration is -- the substance of
`this declaration has not changed.
` Q. Let's go to the page A-14 of this
`declaration. If you look at the top of this page,
`there's:
` "[1.4], checking, by the network interface,
` whether the packets have certain exception
` conditions, including checking whether the
` packets are IP fragmented, and checking
` whether the packets have a FIN flag set,
` and checking whether the packets are out
` of order."
`
`INTEL EX. 1440.014
`
`
`
`Page 15
`
`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
`
` Do you see that?
` A. Yes.
` Q. What does it mean, "IP fragmented"?
` MR. ANSLEY: Objection to form.
` Go ahead.
` THE WITNESS: "IP fragmented" means a packet
`which has been sent as separate IP fragments. In
`other words, the full TCP packet hasn't been sent
`at one time, but it's sent in smaller pieces in
`separate IP datagrams.
`BY MR. LI:
` Q. How do you tell if something is IP
`fragmented or not?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: The IP fragmented packets do
`not contain a full TCP packet. So, in other words,
`the TCP checksum would not be able to be computed
`over that IP fragment, so that would be a packet
`that's fragmented because it doesn't have a complete
`TCP packet.
`BY MR. LI:
` Q. How do you determine whether a packet is
`IP fragmented?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: The packet is not fragmented
`
`INTEL EX. 1440.015
`
`
`
`Page 16
`
`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
`
`if the TPDU is a full one; so if there's a full
`transaction TCP protocol data unit.
`BY MR. LI:
` Q. How do you determine if a TPDU is a full
`one?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: A full TPDU has the format of
`a TCP segment which has the -- all the parts of the
`TCP, including the -- everything from the source to
`the data to the checksum.
`BY MR. LI:
` Q. What is the difference between a full TCP
`TPDU and not a full TCP TPDU?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: One that's not full doesn't
`have the complete -- all the parts of the protocol
`data unit.
`BY MR. LI:
` Q. Which part doesn't it have?
` A. It doesn't have all the data.
` Q. How about headers?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: In TCP, it's unusual for TCP
`to be IP fragmented, and the fragments are just all
`of the pieces that would make up a normal frame. So
`
`INTEL EX. 1440.016
`
`
`
`Page 17
`the first part -- the first fragment would have the
`first part of at least the header of the TCP packet
`and then the subsequent IP datagrams would have the
`rest of it.
`BY MR. LI:
` Q. Is there anything in the TCP header that can
`tell you if the packet is IP fragmented or not?
` A. Can you repeat that? I didn't understand.
` Q. Is there anything in the TCP header that can
`tell you if the TCP packet is IP fragmented or not?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: By just looking at the header
`you can't tell that because it could be fragmented
`after the header.
`BY MR. LI:
` Q. Then how do you tell if a packet is
`IP fragmented or not?
` A. I'm not sure I understand the question.
` Q. Okay. So if you read the claim elements as:
` "Checking, by the network interface,
` whether the packets have certain
` exception conditions, including checking
` whether the packets are IP fragmented..."
` Do you see that?
` A. Yes.
`
`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
`
`INTEL EX. 1440.017
`
`
`
`Page 18
` Q. Where is this checking whether the packets
`are IP fragmented that's disclosed in the prior art?
` A. If there's an entire TCP segment, and TCP
`segments can be validated to know that they're
`complete. So if it's complete, it's not fragmented;
`and if it's incomplete, it's fragmented.
` Q. And where is that in the prior art cited in
`your declaration?
` A. I cite it on page 99 of my report where it
`says:
` "Neither side is trying to close the
` connection. The TPDU is a full one,
` e.g., no IP fragmentation," and then it
` goes on: "No special flags," et cetera."
` Q. Right. So my question is: Where in the
`prior references? I mean, what prior references did
`you cite to anticipate or make this patent obvious?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: That's cited in Tanenbaum 96.
` MR. LI: I'm going to introduce the next
`exhibit, Exhibit No. 2.
` (Exhibit 2 marked for identification.)
` MR. ANSLEY: Thank you.
` MS. KHACHATOURIAN: Thank you.
` MR. ANSLEY: This is just an excerpt 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
`
`INTEL EX. 1440.018
`
`
`
`Page 19
`
`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
`
`1006? Is it 1006?
` MR. LI: Let me ask the witness first.
` MR. ANSLEY: Okay.
`BY MR. LI:
` Q. Do you recognize this document?
` A. This is at least part of Tanenbaum 96, the
`reference I was speaking of.
` Q. Okay. If we look back to your declaration,
`A-14, you cite to Tanenbaum 96 page No. 565, 566,
`and 567, as well as 585, 545, 542. Is that right?
` A. That's what it says in my report, yes.
` Q. Okay. Can you find in this reference on the
`cited pages where it discloses checking whether the
`packets are IP fragmented?
` A. On page 56- -- sorry, 585, the second full
`paragraph, it says:
` "The TPDU is then checked to see if it is
` a normal one: The state is ESTABLISHED,
` neither side is trying to close the
` connection, the TPDU is a full one, no
` special flags are set, and the sequence
` number is the one expected."
` Q. Is there any other places that discloses
`this checking whether packets are IP fragmented?
` MR. ANSLEY: Objection. Form.
`
`INTEL EX. 1440.019
`
`
`
`Page 20
`
`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
`
` THE WITNESS: I haven't read the full
`thousand-page reference since you asked me the
`question --
`BY MR. LI:
` Q. No. Right.
` A. -- so I can't tell you if there are other
`places.
` Q. But let me ask you this: In the pages cited
`in your report, do you have any other disclosure in
`these pages cited in your report disclosing checking
`whether the packets are IP fragmented or not?
` A. Starting on page 545 it talks about the
`checksum field and that it checksums the data -- I'm
`sorry -- it checksums the header, the data, and the
`conceptual pseudo header as shown in Figure 6.25.
`And it goes on to say on the next page:
` "As a consequence, when the receiver
` performs the calculation on the entire
` segment, including the Checksum field,
` the result should be 0."
` So that checksum field is the way that the
`segment is validated and that checksum will not be
`correct if it's been IP fragmented.
` Q. Is this checksum IP checksum or is it
`TCP checksum?
`
`INTEL EX. 1440.020
`
`
`
`Page 21
` A. That's a TCP checksum that's being referred
`to in this paragraph.
` Q. So if a packet is IP fragmented, the TCP
`checksum would not be correct. Is that what you
`said?
` A. Yes.
` Q. Is that always the case?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: If it's been fragmented and
`the entire TCP segment isn't there, then the TCP
`checksum can't be computed correctly.
`BY MR. LI:
` Q. Then how do you transmit an IP fragmented
`packet if the TCP checksum can't be correct?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: The TCP checksum is not
`correct until the IP fragments have been put
`together to build the entire TCP segment.
`BY MR. LI:
` Q. So let's explain more about IP fragmented.
`Why does there need to be IP fragmented in the first
`place?
` A. In TCP there doesn't really need to be
`IP fragments because a large transfer is broken up
`into separate TCP segments, and each TCP segment
`
`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
`
`INTEL EX. 1440.021
`
`
`
`Page 22
`is a full IP datagram. It's a single IP datagram.
`IP fragmentation is used for other protocols that
`don't do segmentation at the network -- at the
`network layer.
` Q. And how do you do an IP fragmentation from
`the sending side?
` A. The sending side will build up packets with
`less than the full data to be transferred, so it
`will break up the data into smaller packets and send
`them individually.
` Q. Okay. So when there is a large TCP packet
`to be fragmented, would the TCP header be copied to
`each fragmented IP packet or it would only in one IP
`packet?
` A. The entire TCP segment just looks like
`data to IP, so it just looks like one long string of
`bytes. So the fragments -- the IP fragments divide
`up that complete TCP segment into however many
`pieces it needs to be divided into.
` Q. And if we look back to the claim of 1.4
`of A-14, is there anything else on the pages you
`cited talking about checking whether packets are
`IP fragmented or not?
` A. You're asking if there's any more on this
`page than the part that I read?
`
`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
`
`INTEL EX. 1440.022
`
`
`
`Page 23
`
`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
`
` Q. Uh-huh.
` A. The part that I read is the part that talks
`about the TCP fragments -- I mean the IP fragments.
` Q. Let's look at the Tanenbaum reference on
`page Bates No. 047.
` Do you see this Figure 1-16?
` A. Yes.
` Q. What does this figure show?
` A. This figure shows the OSI reference model
`and shows levels 1 through 7 of the OSI model.
` Q. So what is the OSI model?
` A. The OSI model, it was a standard way of
`describing the layering in networks.
` Q. And so do you see on the left side and right
`side there are both seven layers?
` A. Yes.
` Q. And on the bottom it says: "Physical."
`What does that mean?
` A. The physical layer is the layer having to
`do with the signals that get sent over the wires,
`the electrical or optical signals.
` Q. And what does the physical layer do?
` A. The physical layer communicates between the
`endpoints through electrical or optical means.
` Q. Can you give an example of a physical layer.
`
`INTEL EX. 1440.023
`
`
`
`Page 24
`
`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
`
` A. For instance, in Ethernet, 100Base-T is
`100 megabit Ethernet which is signaled with
`differential pairs over wiring; like CAT5 wiring.
` Q. Okay. And do you see on the right side of
`the figure, do you see next to "Physical" it says:
`"Bit"?
` A. Yes.
` Q. What does that mean?
` A. That's the unit that's transmitted at
`a time. In the physical layers sometimes it's
`a single bit and with later protocols they use
`multilevel signaling and communicate multiple
`bits at a single time.
` Q. What is multilevel signaling?
` A. Normally in logic there are two different
`voltages. For instance, 0 volts may communicate a
`logical 0, and 5 volts may communicate a logical 1.
`But it's possible to have, say, four different
`voltages that is 0, 2, 4, and 6 volts, for instance,
`and with 4 voltages you can communicate 2 bits
`instead of a single bit.
` And that's the way some of the later
`versions of Ethernet work.
` Q. Okay. And right above "Physical" you see
`"data link."
`
`INTEL EX. 1440.024
`
`
`
`Page 25
`
`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
`
` Do you see that?
` A. Yes.
` Q. What does "data link" mean?
` A. The next page describes the data link layer
`on page 30.
` "The main task of the data link layer
` is to take the raw transmission facility
` and transform it into a line that appears
` free of undetected transmission errors to
` the network layer."
` And goes on to describe how it does that.
` Q. Can you give me an example of data link
`layer protocol.
` A. The one I'm most familiar with is the data
`link layer in Server Net, which is a network that I
`worked on, that used a coding of bits to communicate
`packets of data across the network with error
`detection as part of that protocol.
` Q. Do you have other examples of data link
`layer protocol?
` A. Well, every network has a data link layer so
`there are lots of other examples.
` Q. Can you give me one.
` A. Fibre Channel uses a sequence of 8B, 10B
`encoded groups of signals that are then grouped into
`
`INTEL EX. 1440.025
`
`
`
`Page 26
`
`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
`
`packets that are sent across the network.
` Q. And then do you see next to "Data link"
`there's "Frame"?
` A. Yes.
` Q. What does "frame" mean?
` MR. ANSLEY: Objection to form.
` THE WITNESS: "Frame" is kind of another
`word that sometimes -- a synonym with "packet." It
`just means a group of bits that are sent together.
`BY MR. LI:
` Q. And do you see the network -- the word
`"Network" above "Data link"?
` A. Yes.
` Q. And what does "network" mean?
` A. The network layer is commonly a layer like
`the IP layer, which is the IP packets that are sent
`across the network.
` Q. Is IP protocol a network layer protocol?
` A. Yes.
` Q. And do you know any other example --
`examples of network layer protocol?
` A. There are other layers -- layered protocols,
`like Fibre Channel, for instance, that have a
`network layer.
` Q. What was the name of the network layer for
`
`INTEL EX. 1440.026
`
`
`
`Page 27
`
`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
`
`Fibre Channel?
` A. Fibre Channel's layers are in -- are
`numbered. I can't recall which number is the
`network layer.
` Q. Have you heard of a term called "ATM"?
` A. Yes.
` Q. What does "ATM" mean to you?
` A. "ATM" is asynchronous transfer mode, and
`that's a network layer.
` Q. Do you see "Transport" above "Network"?
` A. Yes.
` Q. What does "transport" mean?
` A. The transport layer is another layer
`which may, for instance, be used to send reliable
`communications across the network.
` Q. What's the difference between a transport
`layer and a network layer?
` MR. ANSLEY: Objection to form.
` THE WITNESS: The transport layer is
`encapsulated inside the network, and it offers
`other capabilities depending on what the particular
`transport layer you're talking about.
`BY MR. LI:
` Q. So what other capabilities does the
`transport layer have in addition to the IP or
`
`INTEL EX. 1440.027
`
`
`
`Page 28
`
`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
`
`network layer?
` A. For instance, TCP offers reliable and
`out-of-order transmit of packets so that the packets
`can be lost and still recovered based on TCP.
` Q. How does TCP do that?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: It's a pretty long discussion
`to talk about this. I'm not sure what part of it
`you're asking about.
`BY MR. LI:
` Q. Can you summarize the answer; like how does
`TCP offer its reliable and out-of-order transmit of
`packets?
` MR. ANSLEY: Objection to form.
` THE WITNESS: There are sequence numbers
`associated with the packets, and if the receiving
`end gets a packet that's not in order, it knows it
`because the sequence numbers are not sequential,
`and then it can buffer the packets or ask for the
`packets to be retransmitted.
`BY MR. LI:
` Q. And do you see "Session" above "Transport"?
` A. Yes.
` Q. What does "session" mean there?
` A. The session layer can be another layer on
`
`INTEL EX. 1440.028
`
`
`
`Page 29
`top of TCP that offers more services and has its own
`header and packet format.
` Q. And what is the "Presentation" above
`"Session"?
` A. And that's another usually optional layer
`that's on top of the session layer; again, with its
`own protocol.
` Q. And what is the "Application" above
`"Presentation"?
` A. And the application is the actual work
`that's trying to be performed.
` Q. What do you mean by "actual work"?
` A. If there's a program that's running
`that's accessing the network, the application is the
`program that's using the network to get some mission
`accomplished.
` Q. Can we turn to page with the Bates No. 052.
` Do you see Figure 1-17?
` A. Yes.
` Q. What does this figure show?
` A. It shows an OSI model for sending -- between
`a sending process and a receiving process.
` Q. Do you see there's a data link layer there?
` A. Yes.
` Q. And in the data link layer there are "DH"
`
`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
`
`INTEL EX. 1440.029
`
`
`
`Page 30
`
`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
`
`and the "DT." What do they mean?
` MR. ANSLEY: Objection to form.
` THE WITNESS: That's the data link layer
`header and the data link layer trailer.
`BY MR. LI:
` Q. What does a "trailer" mean?
` A. "Trailer" are some bits that follow the
`data, typically used for a checksum or a CRC.
` Q. Would IP layer have a trailer?
` A. I'd have to look at the format to tell you
`exactly what the format is.
` Q. So if you go to the page with Bates No. 054.
` And do you see Figure 1.18?
` A. Yes.
` Q. And what does that figure show?
` A. This shows a rough correspondence between
`the OSI layering and the TCP/IP layering.
` Q. Do you see the line which is not present in
`the model?
` A. Yes.
` Q. What does that mean?
` A. This -- that shows the presentation and
`session layers of the OSI model are not always
`present in the TCP/IP, although the same functions
`that are in the session and presentation layer can
`
`INTEL EX. 1440.030
`
`
`
`Page 31
`be still present in TCP/IP as protocols between the
`transport layer and the application.
` Q. Why are they not always present in TCP/IP?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: It's possible to write an
`application that directly uses the transport layer
`that doesn't have an intermediate layer, so they
`don't always have to be present.
` MR. LI: I'm going to introduce the next
`exhibit which is Exhibit No. 3.
` (Exhibit 3 marked for identification.)
` MS. KHACHATOURIAN: Thank you.
`BY MR. LI:
` Q. Do you recognize this document?
` A. Yes.
` Q. What is this document?
` A. This document is a paper by Thia and
`Woodside titled "A Reduced Operation Protocol Engine
`(ROPE) for a multiple-layer bypass architecture."
`And this is one of the primary prior art references
`related to the '948 declaration.
` Q. Okay. If you look at page 1 of this
`document, in the first section, "Introduction," the
`first sentence of the second paragraph, it says:
` "The key problems associated with
`
`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
`
`INTEL EX. 1440.031
`
`
`
`Page 32
`
`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
`
` offboard processing include:"
` Do you see that sentence?
` A. Yes.
` Q. What means "offboard processing"?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: Offboard processing is the way
`they're describing what's in the previous sentence,
`which is it talks about "offloading all or part of
`the protocol functions to an adapter."
`BY MR. LI:
` Q. Okay. And what are the problems they're
`describing here?
` A. They have five paragraphs that describe
`these different issues. Do you want me to read
`those paragraphs?
` Q. So let's talk about the first paragraph
`then. I'll start with:
` "Partitioning the functionality between
` the host and the adapter is difficult and
` may easily lead to complex additional
` protocol between the two parts, which may
` cancel out or offset the potential gain
` from offloading."
` Do you see that?
` A. Yes.
`
`INTEL EX. 1440.032
`
`
`
`Page 33
`
`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
`
` Q. Do you agree with that?
` MR. ANSLEY: Objection. Form.
` THE WITNESS: In any electronic design
`you have to be careful of partitioning, so they're
`pointing out that however you're doing the
`partitioning, you have to be careful to avoid
`problems.
`BY MR. LI:
` Q. What means "partitioning" here?
` A. The host processor has the application
`that's trying to communicate and it has to do that
`communication via the adapter, so some of the
`function can be done in the host processor and some
`of the function can be done in the adapter. So
`this paragraph is just talking about making the
`determination of what should be done in the host
`and what should be done in the adapter.
` Q. If you look at the second paragraph it says:
` "Non-protocol-specific processing is a
` large part of the total load...Examples
` include interrupt handling, context
` switching and data copying at layer
` boundaries in deeply layered protocol
` stacks."
` What does this paragraph mean?
`
`INTEL EX. 1440.033
`
`
`
`Page 34
` A. This paragraph is talking about the need to
`reduce the number of interrupts and contact switches
`and data copies in order to make the implementation
`efficient.
` Q. And why is that a problem of offboard
`processing?
` A. It's