`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`UNIFIED PATENTS INC.,
`Petitioner,
`
`v.
`
`PLECTRUM LLC,
`Patent Owner.
`____________
`
`Case IPR2017-01430
`Patent 5,978,951
`____________
`
`Record of Oral Hearing
`Held: August 2, 2018
`____________
`
`
`
`
`Before KEN B. BARRETT, MIRIAM L. QUINN, and
`SHEILA F. McSHANE, Administrative Patent Judges.
`
`
`
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`MATTHEW J. LEARY, ESQUIRE
`DAVID L. CAVANAUGH, ESQUIRE
`Wilmer Cutler Pickering Hale & Dorr, LLP
`1875 Pennsylvania Avenue, N.W.
`Washington, D.C. 20006
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`C. RYAN PINCKNEY, ESQUIRE
`Antonelli, Harrington & Thompson, LLP
`4306 Yoakum Boulevard
`Suite 450
`Houston, Texas 77006
`
`
`
`
`The above-entitled matter came on for hearing on Thursday, August 2,
`
`2018, commencing at 10 a.m., at the U.S. Patent and Trademark Office, 600
`Dulany Street, Alexandria, Virginia.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`2
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE McSHANE: Good morning. We are here this morning to
`conduct an oral hearing in Unified Patents Inc. v. Plectrum, LLC. That's
`IPR2017-01430. Can we have appearances, please.
`MR. LEARY: Your Honor, my name is Matthew Leary. With me
`I have Dave Cavanaugh on behalf of the petitioner, Unified Patents.
`MR. PINCKNEY: Your Honor, Ryan Pinckney for patent owner,
`Plectrum, LLC.
`JUDGE McSHANE: Thank you. We included a general order of
`the arguments for today's hearing in the trial hearing order. And to review,
`petitioners will present their case or petitioner will present its case and will
`have 45 minutes, and you may reserve rebuttal time. Patent owner will then
`argue its opposition. You will have 45 minutes. And then petitioner can
`come back and rebut its -- present any rebuttal in its remaining time.
`A couple of reminders here, if you will, I don't know if you have
`been here before, perhaps you have, but if you are using demonstratives,
`please try to identify the page number you are on or the slide number you are
`on. It helps the record, and also it helps our judge, in particular Judge Quinn
`is in Dallas today, and she's attending remotely. So it helps her because
`she's looking at paper copies of the demonstratives. If you could, please,
`again, use the microphone, and that would help not only the court reporter,
`but Judge Quinn as well. Please don't interrupt the other party if you have
`objections. You can address your objections in your argument time. So just
`hold fire on 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
`
`3
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`
`One more thing. We received an e-mail on July 26, 2018, where
`you stated that both parties had objections to the demonstratives. Are those
`objections still standing?
`MR. LEARY: Yes, Your Honor.
`MR. PINCKNEY: Yes, Your Honor.
`JUDGE McSHANE: So I'm just going to make a comment or two.
`Let's start with the understanding that demonstratives are not evidence and
`are used as visual aids only, and any arguments and evidence that are not in
`the record will not be considered by the panel in rendering a final decision.
`That said, we are going to decline to strike some of the demonstratives
`today. We note that petitioner's objections to patent owner's demonstratives
`are directed to alleged new improper arguments in the petitioner's reply, and
`the objection there is that the assertions are late raised. Here the Board can
`discern the appropriateness of the arguments when we are looking at them.
`Patent owner objects to alleged new arguments in some of the
`petitioner's demonstratives, and we note that some of the slides at issue
`include figures that do not seem to appear in the record. If some of the
`figures and arguments in the set are not in the record, they will be
`disregarded. But again, the Board can discern the appropriateness of the
`arguments and the evidence. Any questions on that?
`MR. LEARY: No, Your Honor.
`MR. PINCKNEY: No, Your Honor.
`JUDGE McSHANE: Thank you. With that, we may proceed.
`And petitioner, do you wish to reserve rebuttal time?
`MR. LEARY: Yes, Your Honor, we expect to -- per your
`instruction, let me get to the microphone. We expect to use about
`
`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
`26
`
`4
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`35 minutes for our principal presentation. And if it's okay with the Board,
`we would like to reserve the final 10 minutes for rebuttal and possibly split
`that time with my colleague, Dave Cavanaugh.
`JUDGE McSHANE: Okay. And on that, what I'm going to do is
`set it for the full 45 and then we'll see how we go. So just let me know when
`you are ready to go.
`MR. LEARY: May it please the Board, good morning. My name
`is Matthew Leary for the petitioner. I put up on the screen petitioner's
`exhibit Unified 1023, and hopefully for Judge Quinn, she's got that. Turning
`to slide 2, in addition to some background today, there are two main topics
`that the petitioner will discuss. First I'll discuss claims 8 and 11. I'll explain
`why it would have been obvious to have used a CRC hash such as that in the
`Jain article in place of the XOR hash that is disclosed in the Cheriton patent.
`I will also explain why the Cheriton patent discloses both an input and an
`output packetizer.
`Second, separately I would like to address the remaining
`row-based claims and explain why it would have been obvious to have used
`row-based SRAMs in the system disclosed in the Cheriton patent and how
`doing so would meet each of the row-based limitations.
`If at any point the Board has any questions, we would hope that
`you would just ask at that time and we could address it as they come up.
`Turning now to slide 4, I would like to start with an overview of
`the '951 patent. That patent is directed to providing memory cache to store
`and look at network addresses. That cache is situated within a network
`device such as a switch or a router or a bridge. The top half of slide 4 shows
`Dr. Seshan's high-level explanation of these types of network devices. 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
`26
`
`5
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`basic problem that they solve is how to route an incoming packet to get to
`where the packet is going.
`As shown by the red arrows here in the top of slide 4, a data packet
`enters the device from the left-hand side and the device has to determine
`internally which output port to route the packet to in order to get to where
`it's going. And as the summary of the invention of the patent states, which is
`shown here at the bottom of slide 4, the system performs three basic actions
`to perform that routing. First, it parses out address information from the
`incoming packet; two, it looks up that information in a cache to try to
`determine, for instance, what output port the destination device is on; and
`finally, it routes the packet to that port.
`JUDGE QUINN: Counsel, I think that maybe you are overstating
`the invention. The way I read it is that the meat of the inventive concept
`here has to do with how the addresses are looked up because that tends to be
`the most computation-intensive activity for a packet to be processed so
`quickly to achieve line speed. So with a cache organized the way it is so that
`it could be -- hold a large number of addresses and look them up efficiently,
`that's what made this useful since packet processing has been around before
`that. So you know, with particular to the specific structure that this patent
`discloses, I'm interested in your thoughts as to how this four-way associative
`-- four-set associative cache, the improvement of that is what we are really
`looking at with respect to the prior art.
`MR. LEARY: Absolutely, Your Honor. And that's our
`understanding as well. It's the cache that's situated in this processing
`environment. That's the focus of, in fact, the very next slide I was going to
`put up. As Your Honor noted, it's the cache that is the focus of the patent
`
`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
`26
`
`6
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`claims within this processing environment. We just wanted to provide a
`quick overview of the application in which that cache works.
`JUDGE QUINN: Well, and my question really goes to, you know,
`the arguments about the rows of the cache and the looking up of the cache.
`We denied institution because we didn't see Cheriton as disclosing the rows.
`And as I see the claims, do you think that this is something that needs claim
`construction in order to really define the invention with respect to this
`four-way set associative cache with the rows organized in the manner which
`are described, which is to have four sets within each of the rows?
`MR. LEARY: Your Honor, we don't see a reason for claim
`construction at this point. What we would like to do, if we could, would be
`to focus first on claims 8 and 11 which, as you know, are the claims that
`don't deal with the row operation, and then address the row-based claims
`second. But I would like to address both here in this example of the
`high-level operation. In fact, we are setting this example in the context of
`claim 2 which, as Your Honor knows, is one of the row-based claims.
`JUDGE QUINN: What I'm trying to figure out here, we have to
`issue a final written decision on these claims, okay. So in order to
`distinguish how we see the claims vis-a-vis what you have presented to us, is
`this an issue of claim construction? And if it is, what should be the
`construction in that situation. And that goes to the independent claims,
`obviously. I know you want to discuss the dependent claims, but to help us
`out in understanding how you see the case and with respect to that claim
`term.
`
`MR. LEARY: So we do not think that any claim construction is
`necessary to resolve the issues here. We think, as I will present later, 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
`26
`
`7
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`the Cheriton patent discloses the actions and the actions on the data that is
`claimed in the row-based claims and that one of ordinary skill in the art
`would know how to implement the Cheriton patent using the known
`row-based SRAMs such as those that were described in the Fujishima
`patent. It was my hope that we could discuss at a high level the patent and
`then move to those specific disclosures, if that would be okay with Your
`Honor in terms of the order of our presentation.
`JUDGE QUINN: Just go ahead. I guess we'll get to it later, then.
`MR. LEARY: I'm sorry, Your Honor. I want to make sure I
`answer your question. We didn't think that claim construction is required to
`resolve the issues. Is there another part of your question that I can answer?
`JUDGE QUINN: Well, I mean, I want to understand. You are
`telling us now that someone of ordinary skill in the art would understand
`Cheriton to have it, but if we don't agree that it has it because the claim
`construction does not lend itself to that understanding, then we would have
`to resolve that conflict. And you are saying Cheriton has it. I want to know
`how is that different than what we already told you in the decision on
`institution?
`MR. LEARY: Sure, Your Honor. So I should clarify that it is not
`the petitioner's position that Cheriton explicitly discloses row-based
`SRAMs. And if Your Honor turns -- just to base this in the petition, if Your
`Honor would turn in the petition to page 23, on page 23, about four lines
`from the top, it says the petition noted that one of skill, quote, would
`organize Cheriton's SRAMs in a row-based manner. We didn't say, you
`know, that it would have to be organized in that way or some other
`talismanic signal that this was talking about inherency. And in so doing,
`
`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
`26
`
`8
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`when the petition said that in page 23, the petitioner recognized that
`Cheriton did not explicitly disclose row-based SRAMs. And instead
`page 23 tried to make clear that the element-by-element mapping in the
`subsequent pages were each directed at showing how the claims were
`unpatentable under that statutory ground of single-reference obviousness
`which was flagged on page 21 as the section header for all of these
`independent claims. So I just wanted to make sure that we were clear that
`the statutory ground was single-reference obviousness based in the Cheriton
`patent. And as we've said, we were not claiming anticipation or anticipation
`via inherency for these claims.
`Specifically on page 23, it said that the row-based SRAMs were
`known. They were -- and there's corroborating evidence there in terms of
`Fujishima. And what followed page 23 in each of the element-by-element
`mappings was an explanation of how the claims were obvious in view of
`Cheriton as used with row-based claims, which would be obvious.
`JUDGE QUINN: Well, there's a difference here as far as I'm
`understanding what you have in the petition, is that there's a difference
`between saying it would be obvious to have values indexed and stored in
`rows in an SRAM. It is different to say like what I read in the patent that
`you have different -- the four sets are in the same row. So you are trying to
`equate storing sets within a row with storing rows, four rows in four sets of
`memory. Is that how you are seeing it?
`MR. LEARY: No, Your Honor. I guess the rubric in which we
`put this is if you take Cheriton with its existing actions which, for instance,
`include storing multiple entries in a single row -- sorry, in a single set of
`SRAMs, if you take Cheriton's teachings of storing entries and SRAMs 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
`26
`
`9
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`you implement Cheriton as we believe would be obvious using row-based
`SRAMs and in particular in the way shown in Figure 6 of the Cheriton
`patent, that that same action of storing entries with the same data, including
`network address, is stored within the rows of that SRAM -- of those SRAMs
`because that's the way the SRAMs work.
`And Your Honor, I wanted to point out that for each of the
`mappings when we discuss what it is Cheriton discloses, that discussion is in
`the context of Cheriton as implemented with row-based SRAMs. And I
`recognize, Your Honor, that in some places the resulting language in the
`petition might have said things along the lines of Cheriton discloses that data
`is stored in a row or something to that effect. But that disclosure was always
`in the context of Cheriton as implemented using row-based SRAMs.
`So for instance, when Cheriton discloses taking a network address
`and storing it in an SRAM, if that SRAM was a row-based SRAM, which
`was known and disclosed in references such as Fujishima, then the data that
`Cheriton stores necessarily is in a row because that's how row-based SRAMs
`work. Row-based SRAMs store their data in rows. They arrange it in rows.
`And when you are looking to identify or locate data in those row-based
`SRAMs, it's performed using a row address or index. And all of that
`disclosure of the structure of the claims comes from the knowledge of one of
`skill in the art of row-based SRAMs such as those in Fujishima.
`So with respect to your original question, which as I understood,
`was does Cheriton store its entries in four -- a different entry in each row,
`that's not our point. Cheriton, when used and implemented with a SRAM
`would store its entries exactly the way that Cheriton always describes storing
`
`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
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`its entries, which is it stores it in the SRAM, and for row-based SRAMs
`there would be a row in which that would be stored.
`Your Honor, that's one of the reasons we noted had that there are
`no modifications to Cheriton that need to take place in order for it to use
`row-based SRAMs, because all of the actions and the data in Cheriton are
`agnostic as to how the SRAMs are set up. They just know that they are
`going to take, for instance, four entries or a destination address and store it
`in an SRAM. And if that SRAM was row-based, that storage, of course,
`occurs in a row. And if you need to locate information that Cheriton
`discloses using a hash index to look up a location in the SRAM, if you are
`using row-based SRAMs, that lookup is performed by looking up a
`particular row number at an address or an offset.
`Does that better answer your question, Your Honor?
`JUDGE QUINN: Yeah, I understand your point. I know where
`you are coming from and I don't think you are going to tell me differently,
`but I just wanted to kind of play with the understanding here that the patent
`describes a very interesting way of storing cache entries, four of them per
`row. And when you index into a row, you have four possible sets from
`which to look for, and that's why the patent talks about the first entry and
`things like that. Whereas, Cheriton doesn't have that as a parallel search into
`four different caches in parallel. So it is different. It's a question of how you
`are looking at it to make it fit the claim language. So trying to figure that
`out.
`
`MR. LEARY: Thank you, Your Honor. And I think we'll get into
`that in more detail here in a minute. With that, I would like to turn to, in the
`interest of time, turn to claim 8. That's on slide 6 of our presentation. 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
`26
`
`11
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`this slide just highlights some of the higher -- I'm sorry. This slide, slide
`number 6, describes the high-level features of independent claim 8 which the
`Board instituted on originally.
`And two of the things I wanted to point out here was, one, that
`claim 8, of course, does not deal with row-based -- does not claim row-based
`access or storage. Instead, claim 8 is directed mostly at how the data is
`processed within the system. For instance, the incoming parsed address
`information is hashed to form an address using a CRC generator. Whereas,
`some of the other claims don't include that. And that hash then is used along
`with address information in an input packetizer to group that information
`together, send it to the cache for a lookup, and then the resulting information
`of that lookup is packaged by an output packetizer to send the resulting, for
`instance, address report information to the output side of the switch.
`So it's independent claim 8 that we wanted to focus on first
`because it doesn't have these row-based limitations and of course sort of
`different issues attached given that it was originally instituted.
`JUDGE QUINN: I guess I should apologize for hijacking your
`presentation.
`MR. LEARY: That's quite all right, Your Honor. We like to
`answer your questions first and foremost.
`So I would like to move to slide 22 and just discuss particular
`issues raised by the Board and by the patent owner. And I'll start with the
`Board's decision on institution, which of course, included claims 8 and 11.
`And when the Board instituted on those claims, of course it evaluated those
`claims in light of the patent owner's preliminary response, and presumably
`the Board didn't agree with the arguments in that preliminary response.
`
`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
`26
`
`12
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`
`And that's notable, Your Honors, because nothing has happened
`since that preliminary response with the Board changing its initial position
`that those claims would be unpatentable. If anything, the case now is even
`stronger for these two claims.
`As you are aware, the patent owner produced no new documents in
`this case. They didn't bother to depose the petitioner's expert. Nor did they
`provide an expert of their own to rebut Dr. Seshan. They developed no
`discovery to rebut either the petitioner or Dr. Seshan. And importantly,
`there's no secondary considerations here in this proceeding.
`But most importantly, the patent owner's response after the
`institution also did nothing to change the state of play following that. If we
`move to slide 23, this lists the various issues raised by the patent owner or
`by the Board. For claims 8 and 11, of course, the issues raised were brought
`by the patent owner. For these claims, the response essentially repeated
`verbatim the same arguments made in the preliminary response. And for
`example, regarding the obviousness of using a CRC hash such as that in Jain
`instead of an XOR hash, if the Board were to compare pages 13 and 14 of
`the POPR to pages 5 and 6 of the POR, you'll see it's the identical argument,
`and for the most part it's verbatim.
`JUDGE QUINN: The way I see it, though, and we see this all
`time, is that we may institute based on the evidence of record at the time and
`as a preliminary finding, of course, that you have some likelihood of
`prevailing, a reasonable likelihood of prevailing. But it's another kind of
`analysis we do at the end when we are looking at “is there a preponderance
`of the evidence.” So with regards to what patent owners decide to do in this
`
`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
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`situation, I see them calling into question whether there is sufficient
`evidence in the petition as it stands to go the distance.
`So I looked at what your rationales were, and it looked like you
`relied on the substitution rationale, substitute the XOR with a CRC hashing.
`And they have called that into question. So your response to that is, what I
`see is, more rationales to combine. So what do we do with these extra
`rationales when your initial rationale was the substitution?
`MR. LEARY: First of all, I think the initial rationale, which was a
`substitution, was sufficient in and of itself not only to make our prima facie
`case under the burden that we had at the initiation of these proceedings, but
`also more than sufficient to also meet our burden now as Your Honors are
`deciding in a final written decision how to dispose of those claims.
`And I wanted to talk both about the original petition and what it
`gave in terms of rationales and what it gave in terms of a motivation for the
`replacement or substitution of the CRC with an XOR hash.
`As background, of course, the patent owner didn't dispute that the
`Jain reference discussed the CRC hash function as an alternative to the
`XOR. There's no real dispute that it would have been possible to have made
`that substitution. I think the thrust of the argument that the patent owner
`made regarding these claims in its response was that they argued that there
`wasn't a sufficient reason to make that substitution.
`And what I wanted to point out for the Board initially in the
`petition is that in the petition, and this is cited among other places at
`page 57, the Jain patent not only stated that CRCs were known and that they
`were an alternative to an XOR, the Jain patent explicitly stated that the CRC
`was, quote, a preferred hash algorithm for use in its network address cache
`
`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
`26
`
`14
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`even though it said that an XOR could also be used. For Your Honors, that
`would be in the Jain reference at page 6, lines 1 through 4.
`So that preference in our petition alone implicates several
`rationales and motivations to use CRCs instead of the XOR in Cheriton.
`And that preference is also evidence of an expectation of success in making
`that substitution because you have Jain, which is an application that is
`almost identical to Cheriton's application. Specifically it's looking up
`network addresses in a network device. And Jain has a teaching that says as
`between XORs and CRCs, we prefer a CRC. That alone, I believe, is
`enough to make out both the rationale and also an expectation of success.
`The fact that in our reply brief we also added additional rationales
`and pointed out additional motivations, Your Honor, we think that's entirely
`permissible given that that was a direct response in rebuttal to the patent
`owner's assertion that there were no such rationales or motivations. And so
`we believe, Your Honors, when evaluating this for the final written decision
`should be able to properly look at both the prima facie case, which was
`already quite strong in the petition, as well as our direct responses to the
`patent owner in our reply.
`I'm sorry, Your Honor?
`JUDGE QUINN: Well, I'm going to disagree with you on whether
`we can do that or not. But my concern is first what is your evidence in the
`petition. You said that you have provided the substitution and then you
`supported that with your expert declaration. Was there a cross-examination
`of the expert where he expounded on that the substitution would work in the
`same manner with this reasonable expectation of success or is the testimony
`all we have?
`
`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
`26
`
`15
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`
`MR. LEARY: So the testimony from Dr. Seshan that that
`substitution would have been routine and the teaching in the Jain patent that
`he corroborated his testimony with saying that the CRCs would, in fact, be
`preferred for this exact same application of looking up network addresses,
`that testimony, Your Honor, is completely unrebutted. There was no
`deposition taken of Dr. Seshan, although, of course, he was offered. He was
`also offered, just so you know, in July even after the supplemental reply, and
`patent owner elected not to do so. So all of that evidence, that testimony is
`completely unrebutted at this point, at this state of the evidence in the record.
`JUDGE QUINN: Thank you.
`MR. LEARY: Turning to the expectation of success we were just
`discussing, both in the petitioner's reply and also in the petition, as I said, I
`think there was more than sufficient expectation of success.
`In addition, in the reply we pointed to other references by the same
`author which, in fact, were noted by the author of the Jain reference, which
`provided even more evidence saying that a CRC hash would be known to
`work efficiently in the same application as Cheriton in which that
`substitution would be made.
`If Your Honors have any other questions with respect to claims 8
`or 11, I would like to answer them now. If not, I would probably move on to
`row-based claims to address Judge Quinn's questions.
`JUDGE McSHANE: Do you have anything else on the
`packetizing issue?
`MR. LEARY: Your Honor, yes. So I'm looking now at slide 32.
`And the only other dispute for claims 8 and 11 was the input and output
`packetizer terms. We believe the Board correctly described those terms in
`
`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
`26
`
`16
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`its decision on institution as involving the assembly and reformatting of
`information that's transmitted between components of the device. And that
`is exactly what the packetizers in the '951 packet are described as doing.
`For example, as the Board correctly recognized, the '951 describes
`the input packetizer as constructing a packet, which the Board said didn't
`have to be in any particular form, but it described constructing a packet with
`information such as the source and the destination address from the
`incoming packet along with the generated hash function and sending that
`package of information up to the lookup unit where that information, of
`course, is used to do the search in the cache, and also on the output side that
`there's circuitry that says, as an example in the '951 patent, that sends the
`resulting information from the search to the output side of the hardware.
`And I'm going to now switch to claims 33 and 34. I'm going to
`touch on them briefly, but Your Honor, the main point of each of these
`slides is that the operation of the input and output packetizers that are
`disclosed in Cheriton is exactly near the operation of the description of those
`input and output packetizers in the patent. In particular here for the input
`packetizer, as shown on slide 33, there is a section of the switch hardware
`and also a part of the hash function logic that takes exactly the same data as
`described in the patent which is source and/or destination information and a
`hash and sends it to the lookup unit, to the portion of the device that's
`performing the lookup using that information.
`Also on the output packetizer side, again, the information or the
`disclosure that we pointed to in Cheriton is identical to the examples given
`in the '951 patent. It's circuitry that takes the resulting address information
`and related information like the output port, packetizes it and sends it from
`
`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
`26
`
`17
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`these buffers. For instance, like a buffer is labeled 621 through 624. It
`sends that information to the output side of the switch hardware. Again, all
`of this is completely unrebutted in this case in terms of expert, other
`documents or other expert deposition.
`Your Honor, does that take care of the input and output packetizer?
`JUDGE McSHANE: Yes, thank you.
`MR. LEARY: Now turning to slide 35, I would like to address the
`row-related claims. And especially for Judge Quinn, to preface this, we
`want to let you know that we appreciate where the Board is coming from,
`and we recognize that the Board's initial impression was that the petition was
`deficient for these row-based claims. So in preparing for today we went
`through very carefully the Board's responses and statements it made in its
`earlier findings, and we wanted to answer what we recognized as being some
`of the concerns or questions that the Board had about the petition before we
`even started talking about any of the details on the merits for these claims.
`The first question was whether the petition was legally sufficient.
`And as Your Honors are well aware, for a petition to be sufficient, it requires
`an identification of the challenged claims, the statutory grounds for those
`claims and an application of the evidence to whatever issues are raised by
`those grounds. And we believe we did exactly that.
`As I pointed out a minute ago, on page 21 of the petition, it noted
`this was a single-reference obviousness ground. And consistent with that
`ground, on page 23 it provided the context for the rest of the mapping which
`was that one of skill in the art, quote, would organize Cheriton's RAMs in a
`row-based manner and in so doing, recognized that Cheriton, although it
`didn't explicitly disclose row-based SRAMs, it would be obvious to use
`
`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
`26
`
`18
`
`

`

`Case IPR2017-01430
`Patent 5,978,951
`
`row-based SRAMs and then each of the additional claim limitations
`discussed in that context how one of skill in the art would view Cheriton
`when used with the row-based RAMs.
`Second, the second question we understood the Board had was
`whether the petition had adequately provided rationales for why it would be
`obvious to use row-based SRAMs in Cheriton. And the answer to that
`question, we believe, is yes. And the answer to this question starts also on
`page 23. We wanted to point out for the Board that the petition didn't just
`allege simply that row-based SRAMs were known. Instead the petition
`emphasized that row-based