throbber
trials@uspto.gov
`
`
`571-272-7822
`
`IPR2015-00158, Paper No. 34
`IPR2015-00159, Paper No. 45
`IPR2015-00163, Paper No. 51
`March 4, 2016
`
`
`
`
`
`RECORD OF ORAL HEARING
`UNITED STATES PATENT AND TRADEMARK OFFICE
`- - - - - -
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`- - - - - -
`APPLE, INC, HTC CORPORATION, HTC AMERICA, INC., SAMSUNG
`ELECTRONICS CO. LTD., SAMSUNG ELECTRONICS AMERICA,
`INC., AMAZON.COM, INC., SONY CORP., SONY ELECTRONICS
`INC., SONY MOBILE COMMUNICATIONS AB, SONY MOBILE
`COMMUNICATIONS (USA) INC., LG ELECTRONICS, INC., LG
`ELECTRONICS USA, INC., and LG ELECTRONICS MOBILECOMM
`USA, INC.,
`Petitioners
`vs.
`MEMORY INTEGRITY, LLC
`Patent Owner
`- - - - - -
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`Technology Center 2100
`- - - - - -
`
`Oral Hearing Held: Monday, February 8, 2016
`
`
`Before: JENNIFER S. BISK, NEIL T. POWELL, KERRY
`BEGLEY, Administrative Patent Judges
`
`The above-entitled matter came on for hearing on Monday,
`February 8, 2016 at the U.S. Patent and Trademark Office, 600 Dulany
`Street, Alexandria, Virginia in Courtroom A, at 10:00 a.m.
`
`

`
`APPEARANCES:
`ON BEHALF OF THE PETITIONERS
`
`APPLE, SAMSUNG, HTC, AND AMAZON:
`
`
`W. KARL RENNER, ESQ.
`
`
`ROBERTO J. DEVOTO, ESQ.
`
`
`DAVID L. HOLT, ESQ.
`
`
`Fish & Richardson P.C.
`
`
`1425 K Street, N.W.
`
`
`11th Floor
`
`
`Washington, D.C. 20005-3500
`
`
`202-783-5070
`
`
`renner@fr.com
`
`ON BEHALF OF PETITIONER SONY:
`
`
`
`
`
`
`
`
`WALTER E. HANLEY, JR., ESQ.
`Kenyon & Kenyon LLP
`One Broadway
`New York, New York 10004-1007
`212-425-7200
`whanley@kenyon.com
`
`2
`
`
`
`
`
`
`
`
`
`
`
`

`
`APPEARANCES (Continued):
`ON BEHALF OF THE PETITIONER LG:
`
`
`JAY GUILIANO, ESQ.
`
`
`RYAN M. MURPHY, ESQ.
`
`
`Novak Druce Connolly Bove & Quigg LLP
`
`
`1875 Eye Street, N.W.
`
`
`Eleventh Floor
`
`
`Washington, D.C. 20006
`
`
`202-333-7111
`
`
`jayguiliano@novakdruce.com
`
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`
`
`
`
`
`
`
`JONATHAN D. BAKER, ESQ.
`MICHAEL SAUNDERS, ESQ.
`Farney Daniels, P.C.
`
`411 Borel Avenue, Suite 350
`San Mateo, CA 94402
`
`424-268-5210
`jbaker@farneydaniels.com
`
`3
`
`
`
`
`
`
`
`
`
`
`
`
`

`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`P R O C E E D I N G S
`
`(10:00 a.m.)
`JUDGE POWELL: Please be seated. Good
`morning. This is a consolidated hearing for the following
`three cases: IPR2015-00158, IPR2015- 00159, and
`IPR2015- 00163. IPR2015-01353 has been joined with
`IPR2015- 00163. And IPR2015-01376 has been joined with
`IPR2015- 00159.
`Each case involves U.S. Patent Number 7,296,121
`B2. In the hearing room with me I have Judge Bisk and Judge
`Begley.
`
`Can counsel please state your names for the
`record? We will start with Patent Owner, I suppose.
`MR. BAKER: Thank you, Your Honor, Jonathan
`Baker on behalf of Patent Owner, Memory Integrity.
`MR. SAUNDERS: Michael Saunders of Farney
`Daniels on behalf of Patent Owner, Memory Integrity LLC.
`MR. RENNER: Karl Renner of Fish & Richardson
`on behalf of Apple, Samsung, HTC, and Amazon.
`MR. DEVOTO: Rob Devoto on behalf of Apple,
`Samsung, HTC, and Amazon.
`MR. HANLEY: Walter Hanley of Kenyon &
`Kenyon for Petitioner Sony.
`JUDGE POWELL: Okay. With that, I will note
`that, consistent with our hearing order, that the Petitioners
`
`
`
`4
`
`

`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`will have 120 minutes of argument time. And the Patent
`Owner will also have 120 minutes of argument time.
`This morning we will hear arguments related to the
`original claims of the patent. After that, we will take a break
`for lunch. And then after lunch we will hear arguments
`regarding Patent Owner's motions to amend.
`So we will start with Petitioners presenting their
`case regarding the original claims of the patent. Petitioners
`may reserve time. After the Petitioners' arguments regarding
`the original claims of the patent, the Patent Owner will
`respond. And Petitioners then may use any remaining time to
`respond to Patent Owner's presentation.
`One other housekeeping item, during your
`presentations make sure to identify each demonstrative exhibit
`clearly and specifically. For example, you can refer to it by
`slide or screen number.
`With that, do we have any questions before we
`start? As I said, we will start with Petitioners.
`I guess I do have one question for Petitioners.
`MR. RENNER: Yes, Your Honor.
`JUDGE POWELL: Would you like to reserve
`
`time?
`
`MR. RENNER: Your Honor, it is an interesting
`proceeding. It is unusual.
`JUDGE POWELL: Right.
`
`
`
`5
`
`

`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`MR. RENNER: So, Your Honor, we're thinking of
`it as having two sessions, but there will be some fluidity, of
`course, between the two.
`JUDGE POWELL: Right.
`MR. RENNER: We're thinking that we would like
`to reserve roughly 30 minutes for the motion to amend
`section.
`
`JUDGE POWELL: Okay.
`MR. RENNER: And roughly 30 minutes for our
`redirect in the original claims section.
`JUDGE POWELL: Okay. So you are thinking 60
`minutes on the case-in-chief?
`MR. RENNER: That's right.
`JUDGE POWELL: Okay. Thank you.
`MR. RENNER: If I may please, counsel would
`like to approach the bench and give you copies of the slides,
`if you would like to have them in paper.
`JUDGE POWELL: Certainly. That would be
`great. Whenever you are ready then.
`MR. RENNER: May it please the Board, my name
`is Karl Renner. I'm here with Rob Devoto and David Holt,
`you just mentioned. And we're here on behalf of, as I
`mentioned, Apple, Samsung, HTC, and Amazon.
`Client representatives are here from the
`companies, from Apple and Samsung, in attendance as well.
`
`
`
`6
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`Your Honor, at the outset we wanted to -- slide 2,
`please -- look at the structure of the hearing organizationally.
`We will focus on the Pong and the Koster reference in our
`case-in-chief against the original claims.
`You will see the slides have in them sections -- a
`section on claim construction as well. Our intent is to feather
`that into the conversation as it relates to our coverage in each
`of Pong and Koster.
`Today, just as a prefatory matter, we note that
`most of the arguments you are going to hear, they are
`arguments you have already heard. They were arguments that
`were presented in the Patent Owner preliminary response,
`particularly as it relates to "states" as we will get into, so it
`may be that you are hearing things that sound familiar to you.
`That will encourage that you have already heard most of those
`arguments. And we will remind Your Honors of exactly how
`that went when you did.
`Next slide, please.
`Briefly the '121 patent, we thought it would make
`sense to just to have literally two slides, one to the claim and
`one to the structure, to orient this morning. Main memory is
`shared by multiple processors in the '121 patent.
`And those processors, they maintain cache for the
`purpose of speeding the requests and writes to memory 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
`
`
`
`7
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`might otherwise occur slowly in the memory. Not atypically,
`along with speed, the cost is reliability.
`And when you have multiple copies of a piece of
`data that are being maintained in these caches, you have to
`keep track of where those copies are and exactly what the
`state of them is.
`The solution lies in cache coherence protocols.
`And these aren't new. And they are not truly the subject of
`today's conversation, other than to say this is the backdrop
`against which we are talking.
`And in those cache coherence protocols, what we
`see is there were shared versions of buses and protocols that
`were used where processors would each themselves maintain a
`directory, with a heavy load they would have to monitor the
`bus line, the shared bus line, that is, and identify all
`transactions to keep informed their directory, so that they can
`later direct requests to specific cache versions of the content
`that was otherwise in memory.
`This is, again, fairly inefficient and not scalable.
`So what we have before us is a patent and a variety of prior
`art that instead relates to a directory-based protocols. And a
`directory-based protocol, what we have is a cache coherence
`controller that maintains a map, if you will, of where the
`cached copies are in the shared memory system.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`8
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`So that when a request comes in, be it a read or a
`write, you can direct that to the appropriate cached versions
`and not require every processor throughout the system to
`maintain its own structure.
`It is more scalable. And, most importantly, it
`requires there be a directory of some form and also some
`protocol logic.
`Next slide, please.
`Claim 1 we thought representatively we would talk
`about just for a moment and show you that in red the heart of
`the claim is the probe filtering unit. And at the heart of the
`claim, it receives and selectively transmits things called
`probes. The word probe is used frequently. Really think of a
`memory transaction. This is the read or the write.
`In green we see how that filtering is going to
`occur, how is it that the probe filtering unit is going to
`transmit probes to only select ones of the cache memory.
`Well, it is going to, in green we see, rely on states that are
`associated with the cache memory.
`In blue we see the context for this being done.
`And this is that the processors are themselves connected,
`somewhere between them is an interconnect network of
`point-to-point architecture.
`Next slide, please.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`So without further adieu, getting into Pong, what
`we see is four issues that were briefed on the record; that is,
`states, probes, accumulating, and programmed were on the
`record said to distinctively distinguish the claims.
`We're going to talk today about states,
`accumulating, and programming; leaving probes to the written
`record, as we think that that suffices.
`So in slide 23 as we're talking about here, we have
`the overview in the first instance, and just a slide or two on
`Pong to orient ourselves. And then we will get into each of
`the states, probes, and -- states, accumulating, and
`programming.
`Slide 24, please.
`Here we have clips from the cogent summary that
`was given to us in the Institution decision of what Pong
`represents. At a high level, Pong is, as the first clip
`indicates, it is a multi-processor system with cache coherence;
`similar to what we have talked about.
`The second clip tells us that the processors in the
`Pong system, they are interconnected by, indeed,
`point-to-point links in a memory controller. And the third
`clip tells that there is a directory that is maintained in Pong.
`And we will see that that directory has, as you can see here, a
`presence bit vector of the type we're talking about, a map that
`allows us to know where those cached copies reside.
`
`
`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`Next slide, please. This is slide 25.
`In slide 25 we look at a commonly referenced
`figure that is annotated from within Pong to show how the
`memory transactions navigate within the Pong architecture.
`Here in specific on the left upper side we have got in blue
`processor P0, and we have got numbers 1 through 7 inserted
`into the drawing -- this is from the record, of course -- that
`track what the Pong reference described as that memory
`transaction went through.
`And Dr. Horst used these numbers to describe
`exactly what was happening at each one of these stages. So
`his declaration can be consulted for a more detailed review,
`but for now I will just leave it at arrow 1 takes the memory
`transaction from the processor, that blue box in the upper left
`processor 0, where we initiate or we issue a request, and down
`to the green box, which is the memory controller, step 2
`enumerated, to the directory filter in purple in item 400 in the
`middle of that memory controller, that is arrow 3.
`And at arrow 4 we see that it escapes the memory
`directory filter and goes to the interconnect line. It is called
`here a switch, 410. That is the parallel line, the horizontal
`line below item 4.
`And from there it is being directed to that
`processor P1 and the processor P2. How? The presence bit
`vector that we talked about. Clearly in this example you can
`
`
`
`11
`
`

`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`see that it has been identified that each of these two
`processors has a cached copy of that, which is in the solicited
`memory.
`
`So when the transaction comes in we direct the
`traffic to them. Perhaps most notable here is there is no line
`to P3. That tells us about the selectivity that the memory
`controller, and importantly the directory filter, exacts. That
`transaction goes only to P1 and P2.
`Next slide, please.
`If we go to slide 26, in fact, we see the return
`path. Here we illustrate with just respect to one of the
`message transactions, the probes that come through, but we
`see what processor 1 does with it. And we again use this
`figure, which comes out of Dr. Horst's declaration, to show
`that the path goes from that processor P1 down to the
`interconnect switch that is shown at 422 at item 10 and
`returning up to a buffer that holds multiple pieces of data that
`is between arrow 10 and 11, so that the ultimate responses can
`all reach the processor that made the memory transaction in
`the first instance.
`Now, mind you, as we will later discuss in greater
`detail, Pong, it doesn't contemplate intelligence at the
`recipient, that P1 processor, to make intelligent decisions
`about what to do with a received request or a probe. Instead,
`if it receives a probe and it has a data copy, it replies.
`
`
`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`So we will talk about the fact that when multiple
`of these requests go out, which they do, Dr. Horst tells us
`almost always they do, that almost always we're going to have
`responses, multiple of them coming back from those
`processors in the field.
`Next slide, please.
`So this is slide 27. And slides 27 through 29,
`these really provide a mapping that goes between the claimed
`language and the drawing structure we have here, very briefly
`just pointing out the processing nodes in blue are mapped to
`the blue text of the claim, that the cache memory in yellow is
`mapped to the cache memory in the processing nodes.
`Slide 28, please.
`The memory controller here is representative of
`probe filtering unit. In fact, it has got a directory at its center
`in purple which identifies with the caching -- sorry -- the use
`of states to identify where to select those memory transactions
`among the various different cached copies.
`Next slide, please.
`In slide 29 we, again, consult the enumerated
`version of this figure with 1 through 7, arrows 1 through 7
`tracking to show you that the received probe, that is the
`received probe that comes into that directory filter or the
`PFU, probed filtering unit, that it is used to send out a select
`number of probes to select of the processing nodes.
`
`
`
`13
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`Next slide, please.
`Slide 30 we just use to remind Your Honors that
`you have already come to this conclusion. You have already
`seen this mapping. And, in fact, you have applied it in your
`Institution decision, so this is for your reference is all.
`Now we turn to Pong and actually "states." Next
`slide, please, 31.
`Again, to remind you, we will hit states before
`accumulating and programmed. Slide number 32, please.
`In slide 32 we see that there are two different
`issues that relate to states. One is a construction issue. The
`other is a prior art application issue.
`The first of them is that Memory Integrity would
`like you to revisit the construction that you made in the
`Institution decision. It wants you to narrow "states" from the
`plain meaning of the word "states" to something that
`contextually requires a very specific set of states.
`We will talk about why we don't believe that that's
`appropriate. The next is if Memory Integrity were to achieve
`such a narrowing, they argue that Pong is distinguished by
`that kind of interpretation.
`Again, when we look at Pong, we see quite
`differently that Pong does, indeed, apply even to the narrowed
`version of "states."
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`14
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`Now, mind you, in saying that, there is no dispute.
`Pong applies, undisputedly, to the current construction. The
`only dispute about Pong's application comes about if you
`change your mind on the construction.
`With that, let's look at that construction, please.
`
`Slide 7.
`
`As indicated earlier, our slides are, again,
`organized so that the construction lies differently than where
`we have Pong. So we're going back to slide 7 and looking at
`this construction.
`And you will see in the context of the claim that
`the probe filtering information is representative of what
`states, particular kinds of states, states associated with a very
`large -- it is a very broad term -- with cache memories,
`associated with cache memories.
`Okay. So what did we think this means? Well, we
`think this means, as indicated by the table that is here on slide
`7, that it is not a bounded term. "States" is a very broad term.
`"States" includes things like the presence of the data that's on
`the cache memories.
`JUDGE BISK: Can I ask a question? I am a little
`confused about -- it seems to me that there is two different
`parts of the claim construction we're talking about. I am a
`little confused about how they are related.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`15
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`First of all, I am not quite clear what it means to
`be a cache-coherency state. And since in your petition, you
`seem to say that it doesn't need to be a cache-coherency state,
`so I am wondering if you know what a cache-coherency state
`is.
`
`Because, you know, there are those two examples
`in the patent, the 4 and the 5, the MOESI, and then I can't
`remember the other one.
`MR. RENNER: MESI.
`JUDGE BISK: But then there is the second part
`where presence is not a state, which seems kind of like an
`extra limitation.
`So those are bothering me. I am not sure if they
`are the same limitation or separate. If they are the same, I
`guess I need to know more about what cache-coherency state
`actually means.
`But the thing that is really confusing me is in the
`cache- coherency states, at least the one with the four in the
`patent, lists one of the four states as being an invalid state.
`And could we talk about what is the difference between
`presence that we're talking about all over the place here, you
`talked about it in your petition, and the Patent Owner has been
`talking about it a lot, what is the difference between that
`invalid state that clearly even under Patent Owner's definition
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`would be one of the states and what is the difference between
`that and present or not present?
`MR. RENNER: You ask excellent questions. I
`will try to take them in order or at least address them all.
`Please let me know, obviously, if I don't.
`JUDGE BISK: Okay.
`MR. RENNER: Cache-coherency states, that term
`or that phrase, we're using that phrase because that phrase has
`been used. Candidly we recognize the existence of states like
`the MOESI states or the MESI states, whether those are
`themselves cache-coherency states, they probably are, but the
`full range of cache-coherency states, I think the patent itself
`is quite explicit that there could be a variety of states and
`never defines anything as cache-coherency states.
`So it is our position that that's not actually itself a
`term that needs to have boundary on it, but far more important
`than that or maybe consistent with that, those states that are --
`even if you were to recognize cache coherence as a special
`and narrowed class of states, we will see -- in fact, let's go to
`the next slide. And it will help to point this out. Two slides.
`There it is. Thank you.
`We will see those kinds of states as they have been
`applied in the briefing, they only apply to memory lines
`themselves. And the claim is to the cache coherence memory
`or the cache memory.
`
`
`
`17
`
`

`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`So even if you were to recognize a boundary
`around something called a cache coherent state and say that it
`is only the states of MESI or MOESI, that is modified, shared,
`exclusive, and invalid, for instance, you wouldn't apply that to
`the claim that has been articulated because the claim
`articulated doesn't apply to just the data that is specific to a
`memory line.
`JUDGE BISK: Okay. This brings up actually a
`totally separate question I had, which it was interesting I saw
`in your brief that you pointed out that it uses the word
`"cache" instead of memory line there.
`And to me that made me confused about what the
`claim even meant. And every -- all the prior art and the
`patent, it all seems to be discussing a memory line. So I don't
`really even know what -- what does presence mean when you
`talk about the whole cache? Are we talking about the whole
`cache there or what are we talking about there?
`MR. RENNER: Your Honor, I think, again, you
`are hitting right where you need -- right where we plan to hit.
`And if you look, for instance, at their claim number 3, I
`believe it is -- David, do you have that handy? We can see
`that there is actually recognition of claiming that speaks to
`the distinction between the two.
`It doesn't define that distinction, but there is
`acknowledgment of the need to distinguish between something
`
`
`
`18
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`that is called a memory line and a cache memory in the claims
`themselves.
`And here in claim 1, to your point, if you look at
`claim 3, if you look at the back side of this claim you have
`"memory lines stored in the selected cache memories."
`When the Patent Owner wanted to talk about the
`state of the memory lines, they used the words memory lines.
`Right? Just when they wanted to talk about memory lines,
`they used the words memory lines.
`But in claim 1, what we have is states that are
`associated with -- very broad -- something different, cache
`memories. Now, I am not here to tell you that cache
`memories don't store memory lines, but I am here to tell you
`that there is a distinction between claiming that the states -- it
`is a state of the memory line that would be one way of
`thinking about it, but converse with or contrary to that, what
`we have is a state relating to a cache memory and it is
`associated with it. It is quite broad, relative to.
`I point this out because the intrinsic record that
`has been relied upon by Patent Owner speaks to the state, the
`memory line states. That's not what is claimed.
`And even there, there is no definition as to what
`those states are. In fact, you have got a specification that is
`relatively unbounded as it relates to that term.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`19
`
`

`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`So the definition they are asking for us all to
`observe as it relates to the word "states" is enormously narrow
`relevant to the intrinsic record and the claims themselves that
`they have presented, we believe. And it is for exactly the
`reasons you are asking, Judge Bisk.
`We think that there is an incongruity as between
`the claim and what -- part of the specification, perhaps they
`now want to focus on in trying to distinguish this prior art.
`JUDGE BISK: So could you now talk about the
`difference between the invalid state and presence?
`MR. RENNER: Yes, Your Honor. And we
`planned to do that. Let me fast forward so that we can.
`As for invalid, if you look perhaps at slide number
`12 -- thank you -- in slide 12, what we see is -- bear with me
`for a moment -- because the definitions will help us to get
`there, as will the extrinsic evidence.
`The definition here that was offered at the time of
`Institution, the Microsoft Computer Dictionary, it tells us that
`the word "state" and "status" are comparable, equivalent, if
`you will, similarly seen by those of skill in the art.
`Now, you can go left, but I would rather you go
`right because it is faster. If we go right, we can see how the
`word has been used by references that are relevant to cache
`coherence.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`Chaiken is such a reference. And it tells us that
`the status is something that includes presence or absence. So
`what we have is the word status or state, we believe, is
`encompassing, it is broader than just validity or invalidity. It
`includes presence.
`Now, we see this again if you look at the next
`slide, please.
`JUDGE BISK: So you are saying then that state
`includes invalidity and presence.
`MR. RENNER: Correct.
`JUDGE BISK: But when you say that, you are
`implying there is a difference between invalidity and
`presence? I am not sure what that difference is.
`MR. RENNER: Right. And we will talk about
`that. I wanted to first establish that -- thank you for that --
`but I wanted to first establish that the claim we believe is
`broad enough to encompass both, but there are two different
`things.
`
`JUDGE BISK: Okay.
`MR. RENNER: And we will talk about how they
`collapse because of the way the teaching of the reference is.
`Next, please.
`We see three different extrinsic references
`provided by Patent Owner, actually, that call out that the not
`present state is indeed -- is indeed a state.
`
`
`
`21
`
`

`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`JUDGE BEGLEY: Is presence a state, though?
`MR. RENNER: Is presence a state? We believe
`that it is, yes, Your Honor.
`JUDGE BEGLEY: Where is the evidence in the
`record that presence is a state as opposed to not present?
`MR. RENNER: We think that this bears out,
`Number 1, between the contrary correlation between the two.
`We don't see in the patent any definition that would exclude
`the opposite of something. We just have extrinsic references
`talk about the negative version, invalid, but the word "state,"
`as we have mentioned, is an unbounded term in the claim.
`And it is not defined otherwise in the specification.
`So thank you, Your Honor, for the question. We
`believe that that would mean that the word "state" is broad
`enough, absent some modification to sweep in each of invalid
`or valid. But to pick up on this, it is true that invalid is the
`more commonly discussed state.
`And invalid certainly is something that is
`established by the references here, including the Sorin book.
`You will notice that in Patent Owner's materials, often Sorin
`is relied upon for the proposition, in part, that presence isn't
`part of the state. But, indeed, it is.
`Now I have not answered Judge Bisk's question,
`which is why are we talking about two different things,
`presence and state? It is a good question.
`
`
`
`22
`
`

`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`It is because of the way that the Pong reference is
`taught. So if we look at slide number 34, please, what we see
`is the presence bid vector I had mentioned to you. It has got
`ones and zeros. That's a vector, right?
`And that presence bid vector, a map, for any given
`memory location, memory location, it is a map of where the
`instances in cache are, a map of all the different caches in the
`network supposedly. And where there is a one, we know there
`is presence. When there is a zero, we know there is not
`presence. Okay? That's nice. And that establishes presence.
`What it also establishes, though, based on the
`quote here from Pong at paragraph 69, Pong applies something
`called a write update protocol. It is important to understand
`what that means because it implicates what presence ends up
`indicating, okay?
`When you have a write update protocol, as soon as
`there is a write that is made in your scheme, we consult in a
`write update protocol the presence bit vector. And we find all
`the copies of the data that are out in the field.
`You can tell from the name that write update
`protocol, you go and update them. So you find the map and
`then you go and update them. And what does that mean to us?
`That means that all of the copies of the data of the field are
`always updated. The protocol demands it. I am sure there is
`latency, but that's --
`
`
`
`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
`
`Case Nos. IPR2015-00158, IPR2015-00159 and IPR2015-00163
`Patent No. 7,296,121
`
`
`JUDGE BEGLEY: Where does Pong disclose that
`it is updated?
`MR. RENNER: The write update field. We think
`paragraph 69 tells you that write update is applied by Pong. It
`is one of the two different kinds of protocols that it can apply.
`JUDGE BEGLEY: Sorry, to be more clear, where
`does Pong disclose that the directory clears the presence bit in
`the presence bit vector?
`MR. RENNER: So there is a different kind of
`protocol -- thank you for that question -- called the write
`invalidate protocol. And when the write invalidate protocol is
`put into play, instead of updating all the instances of data, it
`is a different implementation incidentally, but when all the
`instances are consulted and you have a write, if you update
`one of the pieces of data somewhere, one of the process
`updates, it is a piece of memory that the copy the memory has
`in its cache. Or maybe the home node where the cache
`actually -- the prior memory resides is updated.
`In that case you have incongruency. You have a
`difference between. You can consult one of two different
`kinds of protocols.
`JUDGE BISK: Sorry, I have two questions on this.
`One is in some of the, like in the patent they call this dir

This document is available on Docket Alarm but you must sign up to view it.


Or .

Accessing this document will incur an additional charge of $.

After purchase, you can access this document again without charge.

Accept $ Charge
throbber

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.

throbber

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.

Become a Member

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

We are redirecting you
to a mobile optimized page.





Document Unreadable or Corrupt

Refresh this Document
Go to the Docket

We are unable to display this document.

Refresh this Document
Go to the Docket