`UNITED STATES PATENT AND TRADEMARK OFFICE
`______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`______________
`
`TALARI NETWORKS, INC.,
`Petitioner,
`
`V.
`
`FATPIPE NETWORKS INDIA LIMITED,
`Patent Owner.
`______________
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`______________
`
`Record of Oral Hearing
`Held: August 14, 2017
`______________
`
`
`
`
`Before STACEY G. WHITE, MICHELLE N. WORMMEESTER, and
`CHRISTA P. ZADO, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`
`APPEARANCES:
`
`FOR THE PETITIONER:
`
` THOMAS F. FITZPATRICK
` ANDY H. CHAN
` CHARLES F. KOCH
` Pepper Hamilton LLP
` 333 Twin Dolphin Drive, Suite 400
` Redwood City, California 94065
` 650-802-3600
` fitzpatrickt@pepperlaw.com
` chana@pepperlaw.com
` kochc@pepperlaw.com
`
` FOR THE PATENT OWNER:
`
`SAMEER GOKHALE
` ROBERT MATTSON
` Oblon, McClelland, Maier & Neustadt, LLP
` 1940 Duke Street
` Alexandria, Virginia 22314
` 703-413-3000
` sgokhale@oblon.com
` rmattson@oblon.com
`
`
`
`
`The above-entitled matter came on for hearing on August 14, 2017,
`commencing at 10:00 a.m. at the U.S. Patent and Trademark Office, 26 S.
`4th Street, San Jose, California.
`
`
`
`
`
`
`
`
`2
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`
`
`P R O C E E D I N G S
` Monday, August 14, 2017 - 10:00 a.m.
` ---oOo---
` JUDGE ZADO: Please be seated. Good morning,
`everyone.
` ATTORNEYS: Good morning.
` JUDGE ZADO: This is the consolidated oral hearing
`in IPR 2016-00976 and IPR 2016-00977 between petitioner
`Talari Networks, Incorporated, and patent owner FatPipe
`Networks India Limited involving challenged claims of U.S.
`Patents 6,775,235 and 7,406,048.
` Judges White and Wormmeester are joining us
`remotely from our Dallas and Alexandria hearing rooms
`respectively, so please speak into the microphone at the
`podium and identify any exhibits or demonstratives that you
`are using so that they can be sure to pull them up.
` Each party will have 90 minutes of total argument
`time. Petitioner will go first and present its case with
`regards to the instituted claims. Petitioner may reserve
`time for rebuttal.
` Thereafter, patent owner will argue its opposition
`to petitioner's case, and if there is any rebuttal from
`petitioner, we will hear it after patent owner's opposition.
` Will petitioner's counsel please introduce
`yourself.
` MR. FITZPATRICK: Good morning, your Honor. Tom
`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
`26
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`Fitzpatrick from Pepper Hamilton on behalf of Petitioner
`Talari.
` With me, by the way, is Andy Chan, my colleague
`from Pepper Hamilton, and Chuck Koch.
` JUDGE ZADO: And would petitioner like to reserve
`time for rebuttal today?
` MR. FITZPATRICK: We would, your Honor. I would
`like to reserve a half hour.
` JUDGE ZADO: Will patent owner's counsel please introduce
`yourselves.
` MR. MATTSON: Good morning, your Honor. Robert
`Mattson with the Oblon firm for patent owner FatPipe.
`Presenting for FatPipe today will be my partner Sameer
`Gokhale.
` JUDGE ZADO: Thank you. We're ready to begin
`whenever petitioner is ready.
` MR. FITZPATRICK: Thank you, your Honors.
` Thank you, your Honors and may it please the
`Board -- can you hear me?
` JUDGE WHITE: I can hear you.
` JUDGE WORMMEESTER: I can hear you.
` MR. FITZPATRICK: Thank you.
` May it please the Board, we'll go straight to slide
`2. I don't believe that the -- thank you. We'll go straight
`to slide 2 where you will notice that we have a table of
`contents of presentation for our presentation.
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
` Each presentation is rather extensive, and, your
`Honor, we do have a separate physical copy for your Honor.
`May I approach?
` JUDGE ZADO: No, thank you. I have it
`electronically. Thank you very much.
` MR. FITZPATRICK: Each presentation, as extensive
`as they are, the slides are intended to provide us a means to
`address any questions you may have. We plan on focusing on a
`very small subset of these slides today.
` And before we dive into those slides, I think it's
`important that we describe the context of what we're talking
`about today.
` First is the level of skill in the art. It's
`undisputed that the level of skill in the art in this matter
`is relatively high. It's a Bachelor of Science in computer
`science, computer engineering, electrical engineering or
`equivalent field as well as at least two years of academic or
`industry experience in any type of networking field.
` By the way, that's Talari's view. FatPipe's level
`of skill for a person of ordinary skill is a bit broader to
`accommodate its expert, but, again, it's similar, and, again,
`it's relatively high.
` And this is important because it highlights what
`was known and what is not in dispute today.
` Routing traffic between disparate parallel
`networks, that was known. That's not in dispute.
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
` Frame relay on the one hand and VPN on the other
`constituting disparate parallel networks, that's not in
`dispute.
` Frame relay as an example of a CO network, a
`connection oriented network, and VPN as an example of a CL or
`connectionless network, again, not in dispute.
` All of these things are well-known, even according
`to the 235 and 048 patents, in the prior art and conceded by
`FatPipe's expert, Mr. Williams.
` It's for those reasons that FatPipe does not
`contest the prior art as to many of the claimed elements of
`the 235 and 048 patents, and that's why I would like to
`address now the ones that are disputed starting with the
`dispute over per-packet basis.
` If we could go to slide 77.
` The per-packet basis issue applies to claims 4,
`that's an apparatus claim, and claim 9 of the 235 patent, and
`claim 7 and 19 of the 048 patent.
` Now, our briefing on this issue is comprehensive
`with its evidence and argument, and that's because the
`intrinsic record is clear about what the patent means when it
`says "per-packet basis."
` Karol plainly discloses, and FatPipe previously
`told Talari, the District Court and this Board that no
`construction was necessary.
` But now FatPipe seeks not only to construe 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
`
`6
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`phrase that we see here on per-packet basis, but then to add
`a series of limitations that are not supported by the
`intrinsic record but also to construe their own construction.
` If we turn to slide 78, this construction fails on
`its face because these words "for each packet makes a
`discrete choice" appear nowhere. They are not in the claims.
`They are not in the spec. They are not in the file history,
`and FatPipe introduces this term "discrete choice." As
`opposed to an indiscrete choice?
` It's for this reason that FatPipe requires a
`construction of this construction adding numerous -- at least
`three limitations that likewise have no basis in the
`intrinsic record.
` So there is these three improper limitations. The
`first -- and we'll take them in turn, but the first is
`FatPipe construes its construction to exclude the patent's
`own embodiment in which packets are routed based on a single
`selection -- that is a quote -- based on a single selection
`that applies to multiple subsequent patents, a situation
`where the decision is made to route packet A down 101 and
`packets B, C, and D also go down 101 even though 280 -- these
`are highways here locally that I'm sure some of us took to
`get here today, your Honors.
` So that's one. It would exclude it in embodiment.
`The second --
` JUDGE WHITE: Let's talk about that for a second.
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`When I look at the specification of the 235 in column 14,
`there's this disclosure column 14 around line 44 or so.
` MR. FITZPATRICK: 44 to 46, your Honor?
` JUDGE WHITE: Yes. It says the past selecting step
`908 may be performed once per packet or given selection may
`pertain to multiple packets.
` MR. FITZPATRICK: That's exactly where we were
`going, your Honor.
` JUDGE WHITE: So I'm looking at that, and I see
`that there are it looks like two separate ways disclosed
`there. Once per packet which seems to line up roughly with
`the construction that patent owners putting forward or that
`multiple packets which is the construction which is what
`you're trying to encompass as well.
` Why is that we can't look at the claim language and
`say that they made a choice to claim the once per packet
`embodiment in claim 4?
` MR. FITZPATRICK: Because that's not what's
`described in the figure that that reference is discussing,
`your Honor.
` If we look at the totality of figure 9 -- this is
`critical. If you look at the totality of figure 9, it shows
`how -- what they're trying to do to rewrite these claims.
` So when you look at figure 9, the first thing that
`you see and -- if we can go to slide 80 for a second.
` If you go to slide 80, it shows that according to
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`the 235 patent, figure 9 illustrates methods -- this is at
`the top -- methods of the present invention, and then again
`we cite the exact reference that your Honor pointed out, and
`it's clear from the specification and both the left and right
`sides of figure 9 that this figure 9 is an embodiment in the
`invention.
` The specification states one example is the
`inventive method in which load balancing or reliability
`criteria cause an initial path selection to be made, 908, and
`then a loop occurs in which multiple packets are received,
`904, and then sent, 918, over the selected path without
`repeating the select -- the selecting steps, 908, for each
`receive, 904, send, 918, pair.
` These descriptions amplify that when you look at
`figure 9 you are not limiting your gaze to half of figure 9.
`You're not looking at it and just saying, well, as FatPipe
`wants you to do which is to say, well, they don't mean what
`they say in claim 4. They don't mean what they say in figure
`9 or in column 14, 44 through 46 or column 16, 22 through 27,
`and instead only half of figure 9 represents the invention.
` That's contrary to the intrinsic record. It's
`contrary to common sense when you look at any figure in a
`patent, and nowhere in the specification does it say, well,
`there is a red half on the left and a blue half on the right
`as FatPipe depicts on page 14 of its patent owner's response.
` And nowhere does the specification say 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
`
`9
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`per-packet basis is that red half on the left. Right?
` And that figure 9's box 904 to 918, that's never
`described as anything other than selecting between network
`interfaces on a per-packet basis.
` So, for the same reason we talk about a round
`robin, and is that on a per-packet basis, and the answer
`would be yes because in a round robin for each packet a
`network interface path is chosen.
` So I think that looking at the figure as a whole
`and the specification as a whole there's not this carving,
`this carving this crevice in claim construction that you
`would need to do to do what FatPipe is seeking from your
`Honors.
` JUDGE WHITE: But are we not reading in things from
`the specification a little too much if we say that this
`figure defines what the claim has to be limited to?
`Especially when there is that language that says, you know,
`what was it, one packet or a group of packets?
` I'm paraphrasing here.
` Should we not look at this figure as just one
`particular embodiment and that it discloses more than one
`embodiment in the specification?
` MR. FITZPATRICK: So, your Honor, I'm not trying to
`limit the claims to this embodiment. What I am trying to say
`is that what FatPipe is doing is it's excluding its own
`embodiment. It's saying that you can't have a situation
`
`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
`
`10
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`where one decision is made for a packet, and that applies to
`multiple subsequent packets. That's what they are saying.
` So when you look at their constructions, you
`remember, the first time they said it doesn't need a
`construction. The second time they said, well, now we need a
`construction, and this is what it means.
` Then they say, well, it needs this discrete choice,
`and then they tell you what it means to have this discrete
`choice, and this discrete choice that they're saying is
`required would mean that you cannot have a situation that's
`described in this embodiment.
` So FatPipe is the one that's trying to read out
`this claimed embodiment, this embodiment.
` JUDGE WHITE: I guess the question coming to my
`mind is the word "per" so specific that it focuses on one of
`the embodiments? Is it -- it's not that they are trying to
`burden the language. That's what the language naturally
`means when they say per packet.
` MR. FITZPATRICK: I'm not -- I may not be following
`your Honor.
` JUDGE WHITE: Let me rephrase that. I think what
`I'm getting at is the language "per-packet basis" so specific
`as to exclude this other embodiment or is it that patent
`owner is trying to make the term "per packet" that specific?
` MR. FITZPATRICK: I --
` JUDGE WHITE: Is the term "per-packet basis," is it
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`that narrow on its own or are they trying to make it?
` MR. FITZPATRICK: I think they're trying to make
`it, your Honor. Now I absolutely understand where the Board
`is going, your Honor is going.
` That would in a sense say the per-packet basis
`requires a network path selection be made for each packet,
`and this embodiment clearly says that's not the case. Right?
` What they're importing into that claim -- into that
`claim construction is the frequency at which the network path
`decision is being made. That whenever there is a packet
`being received by their controller, we're assuming it is by
`their controller, there is a decision tree that that network
`path selection step has to be repeated, and that's not the
`case and that would rule out this embodiment from figure 9
`and from the column 14 and column 16 that the court
`referenced -- that your Honor referenced.
` JUDGE WHITE: And I think I'm looking also at the
`language. It says language -- the language in column 14 is
`may be performed -- may be performed once per packet, and
`here again in the claim language we're getting per packet.
` That seems to line up in a way as if they are
`trying to claim what is here in the spec, just the once per
`packet.
` MR. FITZPATRICK: But it would rule out the or a
`given selection may pertain to multiple packets.
` JUDGE WHITE: But don't they have the ability to
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`claim one or the other if that's what they're trying to
`claim?
` MR. FITZPATRICK: Not if it would rule out that
`embodiment, and that's what would happen. You would
`essentially erase figure 9 in its entirety. You couldn't
`have figure 9, and you couldn't have this embodiment because
`it would be excluded if per-packet basis means each and for
`every single packet this decision tree is being made, is
`being run through.
` Your Honor, I think it also -- our reading --
`actually, well, FatPipe's reading is consistent with the
`other things that it's trying to do. So if I can go to those
`other limitations that would go hand in glove with their
`exclusion of this embodiment.
` FatPipe construes its construction such that the
`discrete choice, this is the second improper limitation, it's
`discrete choice or distinct selecting cannot be based on a
`precomputed route.
` Now, if we turn to slide 84, now, if it was
`precomputed, why does Karol say a determination is made?
`It's simply not right. The plain disclosure in Karol says
`for each packet at the CL-CO gateway that a determination is
`made if that packet should be carried by CO network 160.
` That's not a precomputed route.
` Karol also says the determination of which network
`path to select is based on service requirements and 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
`
`13
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`quote -- the traffic situation in the CL and CO networks to
`evaluate whether a particular packet may be redirected.
` And that's -- that's directly from Karol. Because
`the determination is based on the traffic situation, the
`determination could not be precomputed.
` I thought I was going to take 280 coming down
`today. 280 looks congested. I take 101.
` Again, this is the same example that we used
`earlier.
` That's something I -- that's not precomputed.
`Traffic changes over time. So to read FatPipe's construction
`of discrete choice into these claims, they have no intrinsic
`record to support any of these claims, you would have to buy
`that it excludes this idea of this precomputed route.
` The other thing, if you go to slide 85, FatPipe's
`construction would require that the router updates have it
`faster than once per second.
` Now, patent owner argues that Karol doesn't have a
`packet path selector which selects between network interfaces
`on a per-packet basis because essentially what it is, your
`Honors, is that what's limited to OSPF, and OSPF is too slow.
` This dispute the fact that there is no minimum
`packet rate required anywhere in the patent. Mr. Williams,
`FatPipe's expert, confirmed that there is no minimum packet
`rate required by the patents.
` Under FatPipe's contention, by the way, Karol's
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`routing data basis which are updated approximately once every
`second, that would mean that Karol meets the claim element
`when the packet rate is one packet per second or slower.
` But the fact is that Karol is not limited to OSPF,
`and we'll see this in several of the arguments today.
` Karol specifically says that a person of ordinary
`skill in the art, which is again why I mentioned at the
`outset that the level here is relatively high, but Karol
`specifically says that a person of ordinary skill in the art
`can implement faster protocols.
` Mr. Williams admits that Karol is not limited to
`OSPF, so there is no reason that -- to require into this
`claim that doesn't say anything about packet rate
`requirements that you would include one.
` So if the court has no other questions on per
`packet, I would like to turn to a related dispute over per
`session. We actually linked them together.
` JUDGE ZADO: Actually, I do have one question on
`per packet, and this goes back again to the specification of
`the 235 patent at column 14.
` It's the selection that Judge White had pointed out
`starting around line 44.
` MR. FITZPATRICK: Yes.
` JUDGE ZADO: That the selection can be performed
`once per packet or if they pertain to multiple packets, and
`I'm still struggling with the idea that the claim language
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`does say specifically that there's a selection on a
`per-packet basis. The claim does not recite per-packet basis
`or per multiple packets, and so my question is that the way
`I'm reading this is it doesn't really look like the patent
`owner is attempting to read out a preferred embodiment, but
`rather there are two embodiments given in the specification,
`and the patent owner chose to claim one of those.
` And so if multiple packets were intended to be
`covered by this claim, wouldn't the claim have just said on a
`per-packet basis or per multiple packets?
` MR. FITZPATRICK: I think it's to the contrary,
`your Honor. If they wanted to say that every time a packet
`came in to the controller that packet was analyzed and that
`the decision tree was gone through, they could have said
`that.
` They didn't. And, in fact, when you look at,
`again, that same part of column 14, 44 through 46, I'm not
`sure how it couldn't read that out.
` There is certainly an embodiment here. It's clear
`here that this path selecting step, 908, may be performed
`once per packet, it could happen, or a given selection may
`pertain to multiple packets.
` So if claim 4 is read the way that FatPipe seeks to
`read it, you couldn't have that or a given selection may
`pertain to multiple packets because the selection for one
`would not apply to others.
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
` And the same thing when you again look further down
`on column 16, 22 through 27, where -- when they're talking
`about load balancing and the reliability criteria where an
`initial path selection is made, and then the loop occurs in
`which multiple packets are received and then sent over the
`selected path without repeating the selecting step.
` If they wanted that selecting step to be done every
`time a packet came in, they would have claimed it
`differently. They certainly wouldn't have said what they
`said, and they certainly wouldn't have to then retrofit their
`construction.
` It certainly couldn't be the broadest reasonable
`interpretation to read out this embodiment that is now not
`only shown in figure 9 but also then described twice at both
`column 14 and column 16.
` JUDGE ZADO: So, if I understand correctly, it
`sounds like what you're saying is that the “or a given
`selection may pertain to multiple packets” is also part of the
`invention. The fact that you could do one or the other?
` MR. FITZPATRICK: I think that's certainly part of
`the embodiment. I think that's certainly part of what is
`certainly part of what's claimed, yes.
` JUDGE ZADO: And if the claim had read, rather than
`on a per-packet basis, if the claim had read “select between
`network interfaces once per packet?”
` MR. FITZPATRICK: I think that would be different.
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`If it was once per packet, I think that would be a different
`situation, yes.
` JUDGE ZADO: Okay. Thank you. I had one other
`question because there was a lot of discussion so far about
`the -- I believe it was termed the high level skill of the
`art. Do you believe there is a dispute between the parties
`as to what the level of skill in the art is?
` MR. FITZPATRICK: So, the answer is really no, your
`Honor. I think the different -- we said that it's Talari's
`position that it is a Bachelor of Science in computer
`science, computer engineering, electrical engineering or
`equivalent field as well as at least two years of academic or
`industry experience in any type of networking field, and
`FatPipe stated that the experience would also include network
`design, configuration and/or diagnosis.
` I'm not sure exactly what those things add, but,
`no, I don't think the parties have a dispute into -- in fact,
`I think they agree that it's relatively similar and
`relatively high.
` JUDGE ZADO: Okay. Thank you.
` MR. FITZPATRICK: So a related -- if we could go to
`the next disputed issue, and it's related. Here this is the
`per-session dispute.
` And this is where FatPipe just ignores the explicit
`decision in Karol which shows a packet carrying the UDP
`segment as part of a session.
`
`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 IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
` Now, FatPipe concedes, by the way, that Karol did
`discloses per session per claim 10. This is important.
`Instead of per-packet basis, what we just discussed, claims
`10 and claim 19 claim a per-session basis. And it's telling
`that FatPipe's only challenge to the per-session basis
`pertains to claim 19.
` For claim 10 it doesn't challenge that Karol's
`disclosure of a selecting step to make network path
`selections on a per-session basis isn't met.
` So claim 19 requires a packet path selector which
`selects between the network interfaces on a per-session
`basis, and then it adds to promote load balancing.
` Now, your Honors I'm sure saw that the 235 patent
`says very little about sessions. FatPipe really says nothing
`at all.
` What the 235 patent does disclose is that load
`balancing is preferably done on a per-packet basis for site
`to site data traffic or on a TCP or UDP session basis for
`Internet traffic and it goes to prior approaches of the per
`department, the per accounting, the per sales department or
`per router basis for dividing traffic.
` Now, if we turn to slide 72, we can see that Karol
`likewise discloses that certain packets carrying either TCP
`or UDP segments with certain sessions or applications, and it
`lists several, web access, Telemet and the like, are
`appropriate for a flow to the CO network and others for 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
`
`19
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`CL network as shown in figure 6 of Karol.
` So in figure 6 what we see is step 635
`determines -- can we get that -- can your Honor see that?
` JUDGE WHITE: I have Karol up in front of me, so
`I'm looking at figure 6.
` MR. FITZPATRICK: Figure 6 of the step 635
`determines if the packet associated with UDP is associated
`with a session, and if yes, the CO network is selected in
`step 637.
` So there Karol selects between the network
`interface on a per-session basis.
` Now, obviously also should probably add packets
`carrying TCP segments within a session or application are
`routed to the CL network in figure 6 at step 625.
` But, you know, when you talk about per-session
`basis to promote load balancing, which is what the claim
`requires here, and if we can turn to slide 88, this is
`exactly what Karol discloses.
` Karol discloses these sessions or flows can be
`directed to one network based on bandwidth availability
`criterion, and when one path runs out of available bandwidth
`traffic, it can be put on the next best path. So that's
`Karol disclosing the load balancing associated with per
`session.
` This is also exactly how FatPipe read its claim
`when making its infringement allegation. FatPipe satisfied
`
`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
`
`20
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`this claim element in the District Court litigation saying
`that traffic is put on the best path until that path runs out
`of available bandwidth. The additional traffic will then
`spill over to the next best path.
` So I think there's less of a dispute on per session
`than there was on per packet.
` The next area I would like to turn to is one where
`there is probably more dispute, and that is site interface.
`If we could turn to slide 90.
` The site interface issue applies to claims 4 and 9
`of the 235 patent and claims 1 and 13 of the 048 patent.
` Now, Karol disclose a site interface connecting a
`controller to a site. FatPipe really wants to ignore the
`description in Karol and look only at figure 1 of Karol.
` So, FatPipe is saying that the controller is a
`single device that has to be directly connected to the site.
` The claims certainly do not require the controller
`to be a single device directly connected to the site.
` Likewise, as we see in slide 90 and the 235 patent
`discloses, the controller can be and may be directly
`connected to the site, but it also can be connected to the
`site via a LAN.
` And interesting, again, FatPipe did not need a
`direct connection to satisfy its infringement allegations.
`This is exactly when FatPipe -- and that's what you see below
`when FatPipe served its infringement contentions, it
`
`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
`
`21
`
`
`
`Case IPR2016-00976 (Patent 6,775,235 B2)
`Case IPR2016-00977 (Patent 7,406,048 B2)
`
`contradicted their argument to the Board.
` In there it said that if the controller is Talari's
`device, there is a LAN and a LAN switch between the Talari
`device and the site.
` But really besides the fact that the claims do not
`require that the controller be a single device directly
`connected to the site. Even if a direct connection is
`required, Karol discloses it.
` The express -- there is an expressed disclosure in
`Karol -- there are actually two that I would like to point
`out to the Board. One is Karol at column 2, lines 31 through
`34, where it states it is to be noted that the gateway of the
`present invention can be incorporated within a CL node. And
`in that sense, it would be directly connected.
` And, secondly, Karol also says at column 5, lines 5
`through 8, that the source or destination may be directly
`connected to the CL-CO gateway as opposed to being connected
`through a CL node.
` So this disclosure is referenced in the embodiment
`shown in figure 1 which we cited repeatedly in the position
`in the reply and in Dr. Ney's declaration, and, of course,
`figure 1 describes that embodiment where the controller
`really can be anywhere.
` And so I think that there is also this idea that
`the router source can create packets, and the CL node 111
`could also be the source.
`
`1
`2
`3
`4