`571-272-7822
`
`
`
`
`Paper No. 38
`Entered: December 21, 2021
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`_______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_______________
`
`DISH NETWORK L.L.C., DISH TECHNOLOGIES L.L.C., and
`SLING TV L.L.C.,
`Petitioner,
`
`v.
`
`SOUND VIEW INNOVATIONS, LLC,
`Patent Owner.
`_______________
`
`IPR2020-01276
`Patent 6,757,796 B1
`_______________
`
`Record of Oral Hearing
`Held: November 18, 2021
`_______________
`
`
`
`
`Before JAMESON LEE, DEBRA K. STEPHENS, and
`JOHN A. HUDALLA, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`ELIOT WILLIAMS, ESQ.
`Baker Botts LLP
`1001 Page Mill Road
`Building One, Suite 200
`Palo Alto, California 94304
`650-739-7511
`eliot.williams@bakerbotts.com
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`PARHAM HENDIFAR, ESQ.
`Lowenstein & Weatherwax LLP
`1880 Century Park East
`Suite 815
`Los Angeles, California 90067
`310-307-4510
`hendifar@lowensteinweatherwax.com
`
`
`
`
`The above-entitled matter came on for hearing on Thursday,
`
`November 18, 2021, commencing at 10:00 a.m. EST, via Video-
`Teleconference.
`
`
`
`
`
`
`
`
`
`
`
`2
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`
`P R O C E E D I N G S
`- - - - -
`
`10:00 a.m.
`JUDGE HUDALLA: Good morning, everyone. This is the oral
`hearing in IPR2020-01276. This is Judge Hudalla. I have with me Judges
`Lee and Stephens.
`I'd like to start first with appearances, starting with Petitioner first,
`
`please.
`
`MR. WILLIAMS: All right. Good morning, Your Honor. This is
`Eliot Williams for Petitioner.
`JUDGE HUDALLA: Good morning, Mr. Williams. Are you by
`yourself this morning?
`MR. WILLIAMS: I will be presenting by myself, although I think I
`have some backup counsel and perhaps the client on the public line.
`JUDGE HUDALLA: Okay. Thank you.
`And for Patent Owners?
`MR. HENDIFAR: Good morning, Your Honor. Parham Hendifar
`for Patent Owner. I'm here with the lead counsel, Mr. Kenneth Weatherwax.
`And we believe we also have Patent Owners as participants on the public
`line.
`
`JUDGE HUDALLA: Thank you very much. Good morning to you.
`Okay. So, per our trial hearing order, each side is going to have one
`hour to argue. Petitioner has the burden of proof and will go first, and may
`reserve some rebuttal time. Patent Owner will go second, and may reserve a
`brief sur-rebuttal.
`
`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
`
`
`
`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
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`I want to remind you, as you've already alluded to, that there's a
`public line, so this is a hearing open to the public and a full transcript of it
`will be made part of the record.
`And it goes without saying that you shouldn't discuss any
`confidential information, although I don't know there could be anything
`under seal in this case.
`So, we have your slides. And we can go ahead and refer to them by
`slide number. I think that would probably be the easiest way to do it. We'd
`prefer to see you on the video if we could. But, please do be careful to
`remind us along the way what slide you're talking about.
`And, also, if you could occasionally give us a moment to interject
`with questions, we would appreciate that as well.
`The video operations folks have also reminded us to ask you to
`identify yourself so that there's a clear transcript after the fact.
`So, with that, Mr. Williams, I think we can start with you. Would
`you like to reserve some rebuttal time?
`MR. WILLIAMS: Yes, Your Honor. If you can let me know when I
`have 20 minutes left, that would be great.
`JUDGE HUDALLA: That sounds fine. You can begin then.
`MR. WILLIAMS: All right, terrific. Thank you, Your Honor.
`Again, this is Eliot Williams for the Petitioner. Let me, I guess we
`can just first begin by turning to Claim 1, which is on Slide 4 of Petitioner's
`demonstratives, just to remind everyone about what the scope of the claim is
`that we're dealing with.
`I'll note that in essence this invention is relatively simple. We are
`here -- what's claimed on Claim 1 is essentially serving a request from a
`
`4
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`proxy where data has been pre-stored, or at least some of the data has been
`pre-stored in a buffer so that it can be provided to the end user, you know,
`faster, without, without essentially they're decreasing latency by transmitting
`at a higher data speed than it would be if it wasn't pre-stored in a buffer.
`JUDGE HUDALLA: Mr. Williams.
`MR. WILLIAMS: I apologize for --
`JUDGE HUDALLA: Yeah, I think we have a problem with
`somebody's audio. Go ahead. I mean, go ahead to the video operation, our
`court reporter.
`Okay. Sorry about that, Mr. Williams, if you could please just start,
`I mean briefly start over, that would be great.
`MR. WILLIAMS: Sure. Yeah. Yeah, yeah, we'll start over.
`Okay. Everyone can hear me now. Is that right? Yeah, okay.
`Thank you, Your Honor. Again, Eliot Williams for Petitioner.
`Just beginning with Slide 4 which is the claim language. The
`invention here is essentially serving a request for live media content from a
`proxy server where at least some of the content that's being served has been
`pre-stored on a proxy server in a buffer, and then served to the end user such
`that the end user gets it faster than they would if it hadn't been pre-stored.
`That idea is well-known in (audio interference) -- which there are
`many in the prior art. In this case we present Geagan as sort of the typical
`example of that type of architecture.
`Now, one of the claim terms here which we'll get to, has been
`construed by at least one of the court to require what is, in essence, that the
`buffer on that proxy be a circular buffer, although that's not exactly the claim
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`5
`
`
`
`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
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`language, and that's not exactly the construction, but more or less a circular
`buffer is going to fit into that construction.
`Those types of circular buffers are also well-known to server audio-
`video content. As it happens, Geagan doesn't specifically describe his buffer
`as being circular. He just calls it a buffer. And so, therefore, we've turned to
`Vishlitzky to help fill in the details there as to how you would implement the
`buffer of Geagan. And one obvious way to implement that buffer is with a
`circular buffer.
`And so, more or less that is the ground 1 that we're presenting in this
`case. And so, with that overview, perhaps I should just step into the claim
`limitations. And I'll talk through what I think are the issues in the case.
`JUDGE HUDALLA: Mr. Williams.
`MR. WILLIAMS: And then, of course -- Yes?
`JUDGE HUDALLA: Sorry. This is Judge Hudalla. I just wanted to
`ask you, you're talking about ground 1. Would you agree that Petitioner has
`effectively abandoned ground 2?
`MR. WILLIAMS: I don't think we've abandoned ground 2. And
`ground 2 is certainly still relevant. The only -- I mean, Patent Owner made
`an argument about ground 2 in their Patent Owner response that I didn't -- I
`thought was a bit of a non-sequitur. So, we didn't really respond to it in the
`reply because Petitioner has already addressed the question that they raised,
`which is the use of ATM networks to establish the connection between the
`client and the proxy.
`But, so I can certainly get to that and I can talk about that ground in
`particular. I haven't, haven't abandoned it. The only real relevance of that
`ground, which involves Zheng, is for the notion that when you have data at a
`
`6
`
`
`
`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
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`proxy that you want to send down to the client, you can send that data from
`your essentially a buffer data proxy, you can send that buffer data down to
`the client at a maximum data rate, essentially changing the instantaneous
`data rates as fast as the bandwidth allows, and then slow to a rate that's more
`consistent with the bit rate of the video. That's an idea that's in Zheng and
`we point out in ground 2, but that applies to any proxy. We could apply it
`just as easily to Geagan.
`So --
`JUDGE HUDALLA: Mr. Williams.
`MR. WILLIAMS: -- it's not abandoned.
`Yes?
`JUDGE HUDALLA: I was going to say, though, you certainly didn't
`respond to our findings in the institution decision where we rejected your
`construction. And I believe ground 2 is contingent on that construction.
`So, I'm wondering what might be left after that, given that you
`haven't made any argument?
`MR. WILLIAMS: Yeah, well, I believe that's right. I mean, Your
`Honors, we noted that in the institution decision. And if the Board continues
`to stick with the construction in the institution decision, I agree, I don't think
`Zheng would be relevant to this case.
`JUDGE HUDALLA: Okay. Thank you.
`MR. WILLIAMS: We'll see how the argument progresses here. But,
`you know, as between those two different data rates if, if in fact there is no
`need for the instantaneous data rate to change, then, you know, Geagan and
`Vishlitzky by themselves certainly practice that element and there's no need
`to turn to Zheng.
`
`7
`
`
`
`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
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`Okay. Okay, now back into the preamble.
`So, okay, again just looking at the -- we can turn to Slide, I guess, 7,
`where we start talking about the preamble. You know, there's a couple
`things in this preamble that I don't think are disputed that are present in the
`combination, including a network with a content server and a plurality of,
`you know, live streaming media broadcast objects. Those are essentially the
`live videos that are going to be transmitted by Geagan and via Vishlitzky.
`There's also a plurality of helper servers that connect to a plurality of
`clients. And those are all shown, again, in Slide 7, which is an overview of
`the Geagan system.
`The proxy servers there are, sort of, diagram P in the middle of that
`cloud. And as Geagan explains, those helper servers or proxy servers can all
`talk to each other and communicate between each other when necessary,
`which is part of the construction of helper servers that we apply in this case,
`which comes from, I think, some earlier decisions at the Board as well as
`district courts.
`So, I don't, I don't think there's any real controversy about most --
`about any of those elements essentially that Geagan teaches, the architecture
`of what is Claim 1.
`So, then moving to perhaps the next part of the preamble, or perhaps
`the first element, depending on how you look at it, we can turn to Slide 7.
`Slide 7 requires a method of reducing start-up latency associated
`with distribution of these objects. And, you know, here we have pointed to a
`couple things.
`One is just that the structure of Geagan itself, the fact that it's got
`these proxies that are close to the client, or as close to the client as you can
`
`8
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`reasonably get. That will have the effect of reducing start-up latency
`because of especially where the data has been precast on those helper servers
`in the case of the preconfigured buffers in this combination. You will avoid,
`these will avoid the need to have this round-trip time of it's, you know,
`request going all the way to the back end server, and then getting this data
`from the server and then coming back to the client.
`So, it's that savings in time that we point to primarily as being the
`reduction of start-up latency when you have a preconfigured buffer there,
`i.e., you know, when you have the data already stored on that local proxy
`server in some way.
`And the Board, in its institution decision, essentially agreed to that
`part of the analysis. And, you know, we continue to stand by that decision.
`Obviously, we agree with it, think it's correct, and continues to show why
`Geagan would meet this element of reducing start-up latency.
`Now, we have a second argument in the petition about how Geagan
`meets this element under sort of an alternative mapping, which has to do
`with seaming itself. So, as you recall, in Geagan he teaches the ability to
`seam multiple streams of content together ––which would protect against
`packet loss.
`So if, for instance, the Crosby server's connection to the back-end
`content source is noisy such that a packet or some packets get lost in the
`transmission, then Geagan tells you you can open up alternative connections
`to that server to try to, you know, get, get that data through a different
`channel. And the channels will be less noisy or where the noise wouldn't
`affect the particular packet that was sent.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`9
`
`
`
`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
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`And when you, when you do that it's, of course, possible in Geagan
`then that some users, some early-on users for whom the data wasn't seamed
`yet would get missing packets, and some later users would get those packets,
`later users meaning someone who sent request a few milliseconds after the
`first user, would have been able to take advantage of that seaming and,
`therefore, would get essentially more data in the same amount of time that
`the fist user did. So that's a second reason why we say --
`JUDGE HUDALLA: Mr. Williams.
`MR. WILLIAMS: Yes?
`JUDGE HUDALLA: Judge Hudalla again. Could you address
`Patent Owner's argument on this? They argue that the whole seaming
`method that's disclosed in Geagan would actually result in a longer start-up
`latency because it takes time for there to be multiple streams received at the
`proxy server.
`Why? Why would that reduce start-up latency?
`MR. WILLIAMS: So, thank you, Your Honor, for that question.
`Again, a couple responses. First of all, there's no requirement in
`Geagan that you do seaming. It's clear that that's something that happens
`after a certain threshold amount of packet loss occurs. So, but by no means
`does the petition depend on seaming. So, just to be clear about that.
`JUDGE HUDALLA: Wait a minute though. You rely expressly on
`seaming, though, in your petition, don't you?
`MR. WILLIAMS: We relied on that as the second argument, yes, the
`faster seaming. So, I just made two arguments that are articulated in the
`petition. One is the fact that it's proxied at the edge and will, obviously,
`have a huge savings in round-trip time.
`
`10
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`And the second is that because of seaming there would be additional
`data that a second user could get that the first user doesn't. So, we do rely on
`it, but it's not necessary for our mapping.
`JUDGE HUDALLA: It is -- it either is or is not your position. I
`mean, I, you know, I hope you don't expect us to go through and pick and
`choose amongst your positions. Are you relying on seaming or are you not?
`MR. WILLIAMS: Well, Your Honor, the combination teaches the
`patent under two theories that are articulated in the petition, both of which
`work. So, we are relying on it, but it is not necessary to our position because
`we have a second position that doesn't depend on seaming.
`JUDGE HUDALLA: Okay. Well, I want you to know that we will
`consider it to be part of your position because we're not going to go ahead
`and, you know, we're basically not going to be in a position to pick and
`choose amongst your arguments to find the best one. I don't think that that's
`fair to ask of us.
`MR. WILLIAMS: Well, no, I'm not asking you to pick and choose
`the best one. I mean, we make two, both of them -- I mean, prior art, you
`know, the prior art can involve with the claim in multiple ways. There's
`certainly nothing wrong with that. And the petition presents two ways. But
`if you want to just focus on the seaming one, I'm happy to talk about, about
`that.
`
`I mean, even if you assume Geagan always says seaming, which is,
`of course, not what we say in the petition, but let's assume that that's how the
`Board wanted to interpret the petition, it still would, it still would practice
`the claim for the reasons we set forth in the petition. Because, you know,
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`11
`
`
`
`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
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`the additional latency that's incurred by seaming would only, only apply in
`the worst case to the first user.
`So, the first user who sends this request to get the packet, where
`Geagan determines that there is a lost packet and, therefore, decides to start
`performing its seaming technique, that first user might experience a slight
`additional delay while the proxy of Geagan reorders the seamed packets to
`send them out.
`But, of course, Geagan doesn't, doesn't hold on to those packets
`forever waiting for it to come. It specifically points out that there would still
`be gaps in the seamed stream, and that some users may have less and some
`users may have more, just depending on whether they're able to take
`advantage of additional seamed streams or not.
`And our expert Dr. Negus analyzes this issue in his declaration at
`great length, and concludes that even in the case where there is some
`additional delay due -- due to the seamed streaming, that delay would be
`dwarfed by the advantage you get from having pre-stored content on the
`proxy server, so that certainly, the data rates would increase.
`And, moreover, it's not just Dr. Negus who says that, Patent Owner
`has essentially admitted as much as well. Slide 29 shows that section from
`their Sur-reply. What they say is "Geagan might at best, under certain
`conditions, happen to reduce start-up latency." So, they essentially concede
`that Dr. Negus' analysis is right, at least under, you know, "certain
`conditions."
`And on Slide 28 you will see Patent Owner's expert on cross-
`examination says, yes, in this case where they're seaming, when the second
`user can take advantage of a seamed stream that's already present on the
`
`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
`26
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`proxy server that at least could reduce start-up latency. Even he was willing
`to admit that that's certainly a possibility.
`So, therefore, so I think the answer of the seaming, the seaming
`question is sort of irrelevant to the data rate or the -- sorry, to the increase in
`lat -- decrease in latency because any cost of that seaming would be more
`than dwarfed by the advantage you get from having pre-stored content there,
`at least for subsequent users to that, to that stream.
`JUDGE HUDALLA: And just to be clear, there's no analysis in the
`petition where Petitioner attempts to show how the reduction of start-up
`latency may be affected by seaming vis-a-vis the closeness of the proxies?
`MR. WILLIAMS: I'm not sure I agree with that, Your Honor. Let
`me just look at the petition for one second.
`(Pause.)
`MR. WILLIAMS: I mean -- let me make sure I'm in the right section
`first of all.
`I'm looking at -- so, this is analyzed really more in detail in the data
`rate section than it is in the latency section, although the technical principles
`are essentially the same. The question is do you get data faster at one end in
`one circumstance than the other?
`So, but in terms of the preamble analysis, yeah, the preamble tends
`to -- so, the preamble analysis in the proxy of this element certainly focuses
`on the delay between the proxy and the client. And, you know, our analysis
`there is based on Kevin Negus' conclusion at paragraph 264 that Geagan's
`disclosure that you are streaming close to the client would teach that the
`delay between the proxy and the client is substantially less than the delay
`between the content source and the client's, thus reducing latency.
`
`13
`
`
`
`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
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`So, I'd say that analysis does subsume everything Geagan's doing
`with seaming. But perhaps the more detailed analysis happens when we get
`to Element 1C that talks about the different data rates that could happen
`under the two scenarios: one where you've got content pre-stored, and one
`where you don't. And that's where the difference in seaming is more
`nuanced. And so, he does explain that in more detail.
`Oh, I should say, yeah, I think he also teaches -- sorry, give me one
`more second on the first point, Your Honor.
`I think in the petition, yeah, so let me turn to 1C. So, here on that
`one when we're analyzing on 1C, for instance, on page 45 after talking about
`the difference between the seaming rates, the different seaming scenarios,
`you know, we, we point out that either one of them understood that the
`higher retrieval speed that serves to reduce latency perceived by users is a
`fundamental advantage of proxy caching, such as Geagan.
`So, again, that subsumes everything Geagan's doing about seaming.
`The fact that you've got this thing at a proxy close to the serv -- close to the
`edge is itself, you know, fundamentally the purpose of proxying, and would,
`you know, certainly reduce latency. Otherwise you would never do this,
`there's no reason to introduce additional latency.
`I think going back to analyzing the different --
`JUDGE HUDALLA: Actually, --
`MR. WILLIAMS: Sorry. I was just pointing out element 1C before
`that. Yeah.
`JUDGE HUDALLA: Okay. I was just going to say thank you. I
`think we understand your position. But I just want to make sure you have
`enough time to get into some of your other arguments.
`
`14
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`MR. WILLIAMS: Oh no, yeah, that's fine. This is obviously a large
`one. So, we might as well get to it now in terms of element 1C where the --
`how the seaming works. Then I can come back to element 1B in a moment.
`So, the only other thing I was going to, was going to point out in
`terms of our analysis here, again, is that, you know, their expert has
`conceded the analysis is true, that there's at least cases where this would
`reduce latency despite the seaming. And if you read the petition from pages
`43 to, you know, 46, we have a fairly detailed analysis of the seaming
`scenarios and the different ways that could play out, and showing that in,
`you know, all those cases discussed there that you would have higher data
`rates for the pre-stored data than you would for the non-pre-stored data even
`when seaming is applied.
`Okay. I can do -- let's just move on, I think, to the other element 1A,
`which receiving the first request. I don't think there is a dispute about that
`element. But to the extent there is, our analysis there is at page 27 and 28,
`showing that, you know, in this combination what would happen is the
`request is going to come through the proxy server. Right? It's going to come
`through Geagan. We make that explicit.
`Patent Owner had made an earlier argument in their Patent Owner
`response that somehow in the combination there would need to be a central
`server where these requests came in. And, you know, I think it's clear in our
`combination that's not the combination we're envisioning.
`JUDGE HUDALLA: Could I ask you about that, Mr. Williams?
`This is Judge Hudalla again.
`Do you, well, could you address the fact that Petitioner relies on
`"stream servers controlled by controller servers" and the fact that Petitioner
`
`15
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`also relies on the admission control program of Vishlitzky? Doesn't that
`support the notion that there has to be central control?
`MR. WILLIAMS: Well, I don't know what -- again, this is the whole
`problem with central control. I don't know where that context, what that
`context actually means.
`Within the proxy server of Geagan, yes, there is a central controller
`that's going to do these things. And that's our combination sharing.
`So, the point is the request comes into a proxy which needs to figure
`out what to do with it. And so that's why when we talk about Vishlitzky's
`use of the central controller where we kind of combine together the
`controller server and the stream server in that green circle, which you can
`see on, for instance, on Slide 8, the point is that, you know, Vishlitzky
`teaches functionality about how to allocate and send client requests to the
`preconfigured buffers. And, yeah, you would need that kind of functionality
`in the Geagan proxy servers. So, that's the combination.
`JUDGE HUDALLA: So, you --
`MR. WILLIAMS: Central control. That's what, that's what we were
`talking about.
`JUDGE HUDALLA: Why don't I just start with a few basic
`questions.
`You do agree that --
`MR. WILLIAMS: Okay.
`JUDGE HUDALLA: -- Petitioner relies on stream servers controlled
`by controller servers from Vishlitzky; right?
`
`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
`26
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`MR. WILLIAMS: We, yes, well, we agree that that functionality
`from Vishlitzky would be put into the Geagan architecture. That's the
`combin -- Yes.
`JUDGE HUDALLA: Okay. And you agree that Petitioner relies on
`the admission control program which is shown in Figure 17 of Vishlitzky?
`MR. WILLIAMS: Yeah, the logic of that Figure 17 would be part of
`assigning requests that came into the individual proxies to, to buffers in
`those proxies. That's right.
`JUDGE HUDALLA: Okay. So, that being the case, doesn't that
`basically support Patent Owner's argument that there has to be some sort of
`central control aspect, i.e., the admission control program which sends out
`the video streams, or sends out requested video streams to various stream
`servers?
`MR. WILLIAMS: Yes. In the individual proxies there would be a,
`you know, central control.
`I mean, every, every server, the Geagan proxy servers included,
`needs a processor controlling what it does, yes. But it's not central in the
`sense that there is some overarching, there's some super central controller
`that's controlling all the proxy servers. So that, I think, is where the
`difference in the positions is.
`Patent Owner's argument seems to be that, therefore, there must be
`some super central controller controlling all of these proxies. And that's
`where we, we disagree. And that's certainly not the combination we
`proposed in the petition.
`JUDGE HUDALLA: Okay. So, I guess that's the problem the panel
`is having here, which is: Where in the petition does it tell us how to import,
`
`17
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`as you're suggesting, the functionality of these stream servers controlled by
`controller servers and admission control program into individual helper
`servers? Because I think that almost kind of cuts against the way that
`Vishlitzky operates.
`MR. WILLIAMS: Well, again, Vishlitzky is a server. Right? He
`doesn't have -- there's not an edge server in Vishlitzky. This is just a content
`server that's got data stored thereon.
`Our point is that what Vishlitzky teaches you is two things: it teaches
`you that you can have this notion of popular and unpopular movies, and then
`you might serve those requests differently. And in the case of Vishlitzky,
`you would pre-populate buffers for popular movies, and you wouldn't pre-
`populate those buffers if it was unpopular.And it teaches you that those
`buffers can be circular buffers.
`So, those are the two teachings of Vishlitzky that are relevant to our
`combination. And those are applicable to the proxy servers updating.
`JUDGE HUDALLA: Well, here's my question. You just, you just
`said that you serve popular and unpopular movies differently. You're taking
`that functionality from Vishlitzky, right? So, how would that apply? How
`could I, at multiple different helper servers, try to serve, you know, an
`unpopular versus a popular movie at multiple different helper servers? I
`mean, that seems to be like a central control aspect that Patent Owner is
`talking about.
`MR. WILLIAMS: So, that's not a part of central control. That, okay,
`so in our, in our combination and in Vishlitzky, the decision to make a
`movie popular or not, that's an annual decision that you make to set up 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
`
`18
`
`
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`buffers. So, yes, in that sense there would be an administrator of the system
`who would make that decision. That's discussed in our petition.
`Is that what, is that the question?
`JUDGE HUDALLA: Sure. I'd love to see where you're referring to,
`
`please.
`
`MR. WILLIAMS: Yeah, okay. Let me take you there. One second.
`Looking for that.
`(Pause.)
`MR. WILLIAMS: All right. So, first would be on petition page 34.
`So, on that page we say, a POSITA would find it obvious to
`implement the buffer of Geagan as a preconfigured buffer because
`Vishlitzky teaches it is advantageous to preconfigure a player buffer, et
`cetera, to serve as a popular stream such as Geagan's popular live serving
`content. For example, a POSITA would have understood that establishing a
`preconfigured player history buffer for popular live events would be
`advantageous because those events are likely to draw many viewers.
`Geagan, therefore, discloses a buffer that is manually configured before the
`live event is requested and permanently maintained as memories thereafter.
`So, that's one example.
`And then on page 36 they explain again, however -- this is about four
`lines down -- however, if a request is for a less popular live event than with
`Geagan, in view of the disclosure of Vishlitzky, would either do other
`processes such as allocate non-preconfigured player history buffer to service
`the request, as will be discussed.
`So, that's an example of 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
`
`19
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`IPR2020-01276
`Patent 6,757,796 B1
`
`JUDGE HUDALLA: But in that case nothing is orchestrating that
`choice between what you're calling preconfigured and non-preconfigured
`buffers?
`MR. WILLIAMS: I'm not sure I understand the question, Your
`Honor. What do you mean nothing is orchestrating the choice between?
`JUDGE HUDALLA: So, there must be, first, a decision as to
`whether, whether a certain movie is going to be a preconfigured or certain,
`in this case a livestream program is going to be preconfigured or not.
`MR. WILLIAMS: Right. So, there does need to be a manual
`decision about that by, by the system administrator, yes.
`JUDGE HUDALLA: And that choice can't be done by the individual
`helper servers or proxies in the case of Geagan?
`MR. WILLIAMS: I mean, that's not how -- right, that's not the
`argument we've made in the petition though, that it's not an automated
`process of deciding whether something is popular or not. It is, that is a
`manual decision, obviously, for an administrator.
`JUDGE HUDALLA: Okay. So the once, once the system starts
`running, then what is making that choice as to whether you're serving the
`live content from a preconfigured buffer versus a non-preconfigured buffer?
`MR. WILLIAMS: So that, that is what -- so that is where the logic
`from I guess Figure 17. You see the figure, for instance, on page 35 of the
`petition. So this is that step 171 of Vishlitzky. Where you can request and
`you're asking is that a request for a popular movie or not.
`So that's, that's where that decision is made. But that is a step that's
`being practiced in the, in the proxy server of Geagan of which there are,
`obviously, many. I think three are shown, but obviously there could be
`
`20
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`