`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`RPX CORP., ERICSSON INC.,
`TELEFONAKTIEBOLAGET LM ERICSSON,
`Petitioners,
`
`v.
`
`IRIDESCENT NETWORKS, INC.,
`Patent Owner.
`____________
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`____________
`
`Record of Oral Hearing
`Held: September 24, 2018
`____________
`
`
`
`
`Before THOMAS L. GIANNETTI, MATTHEW R. CLEMENTS and
`SCOTT B. HOWARD, Administrative Patent Judges.
`
`
`
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`ADAM C. FOWLES, ESQUIRE
`J. ANDREW LOWES, ESQUIRE
`Haynes and Boone, LLP
`2505 N. Plano Road
`Suite 4000
`Richardson, TX 75082-4101
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`ROBERT R. BRUNELLI, ESQUIRE
`JASON H. VICK, ESQUIRE
`Sheridan Ross, PC
`1560 Broadway
`Suite 1200
`Denver, CO 80202-5141
`
`
`
`
`
`The above-entitled matter came on for hearing on Monday, September
`24, 2018, commencing at 3:05 p.m., at the U.S. Patent and Trademark
`Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`
`
`2
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE GIANNETTI: Please be seated. So we're back from our
`break and this is the second case we will be hearing this afternoon, case IPR
`2017-01661. The panel is the same panel as we had for the previous case,
`myself, Judge Giannetti, Judge Clements and Judge Howard, and again
`Judge Howard will be presiding. Judge Howard.
`JUDGE HOWARD: Thank you. We'll skip the preliminaries and get
`right down to the argument. Mr. Lowes, will you be arguing again?
`MR. LOWES: No, Your Honor. My colleague Adam Fowles will
`argue.
`JUDGE HOWARD: Okay. And there's 45 minutes for this argument.
`How much time would you like to reserve?
`MR. FOWLES: Your Honor, I'd like to reserve ten minutes of time
`for rebuttal.
`
`JUDGE HOWARD: Okay. And I will let you know when you come
`within five minutes of that rebuttal time.
`
`MR. FOWLES: Thank you.
`
`JUDGE HOWARD: You may begin when you're ready.
`
`MR. FOWLES: Thank you, may it please the Board, and as already
`indicated my name is Adam Fowles. I'm representing Petitioners RPX and
`Ericsson, just for the record.
`Today we'll be focusing on the areas of dispute between the parties
`and as was indicated in the previous proceeding, because claim construction
`is shared between these two I'll be skipping that portion of my presentation
`today unless the Board has any additional questions on that point.
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`
`Otherwise, turning to slide 2. The other areas of discussion we plan
`on addressing today include first, that a POSITA would have been motivated
`to combine Lee and Fichou with Golden, second, that a Golden combination
`with Fichou and Lee renders obvious every limitation of independent claim
`1. Independent claim 1 is the independent claim for which all the limitations
`were argued in Patent Owner's reply briefing and the limitations are in
`common with those in independent claim 13.
`So turning to slide 3. This reproduces the independent claim 1
`language and specifically starting at line 62 there the language of “providing
`by the controller to the portal routing instructions for traffic corresponding to
`the connection so that the traffic is directed by the portal based only on the
`routing instructions” is one point that we'll be talking about today. The next
`is the "required route supported by the portal and dynamically provisioned
`by the controller.” And finally, about the teaching of control paths for the
`connection are supported only between each of the originating and
`terminating end points and the controller.
`Just to complete this overview, turning to slide 4 there were three
`instituted grounds for this IPR. The first ground is the combination of
`Golden, Fichou and Lee with respect to the independent claim and the other
`grounds rely on Golden, Fichou and Lee in combination as well with the
`addition of Har and Pillai for certain aspects of those additional claims.
`Turning now to slide 8. We'll just dive right into an overview of
`Golden and the other references. Golden describes ways to provide a
`guaranteed quality of service between any end stations providing on demand
`reserve band with connections, what it calls virtual circuit connections, that
`are set up when requested and when done they are released or torn down.
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`
`JUDGE GIANNETTI: Pardon me, counsel. Which slide are you on
`
`now?
`
`MR. FOWLES: I'm still on slide 8.
`JUDGE GIANNETTI: Okay fine, I have it.
`MR. FOWLES: Okay. We have another copy if you need it as well.
`JUDGE GIANNETTI: I have it.
`MR. FOWLES: So, for example, Golden teaches that one of the types
`of services that it supports for which applications may request service is
`video conferencing which is one of the types of high quality service
`connections that the 119 patent itself indicates as an example. Golden
`teaches managing these requests with what it calls an enterprise control point
`and I've reproduced here figure 9 from Golden which illustrates this
`enterprise control point 50 which is a controller and as well in this figure we
`see on the left hand side what is called an upgraded host 102 that is in
`communication with this ECP50 and we also see the ECP50 in
`communication with switches 56 along a path, and finally we see here on the
`right hand side a host/router 94 and you'll notice that this host/router 94 does
`not have a signaling channel 58 to the ECP50.
`If we could turn to slide 9, however, we see that Golden teaches that
`this host 94 can be similarly upgraded as the host 102 on the left hand side,
`and what would that mean? That would mean that this host 102 on the right
`hand side when similarly upgraded would likewise include a daemon
`process 106 and a signaling interface 104.
`JUDGE GIANNETTI: So counsel, let me just stop you there. Is it
`correct that this figure that you've shown in slide 9 did not appear in the
`patent?
`
`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
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`
`MR. FOWLES: That's correct.
`JUDGE GIANNETTI: So this is your rendition of what it would take
`or what it would look like if you replaced that element; is that correct?
`MR. FOWLES: This is Dr. Reddy's rendition of what a POSITA
`would understand that rendition to look like.
`JUDGE GIANNETTI: Thank you.
`MR .FOWLES: Yes. Finally, Dr. Reddy testified that a POSITA
`would understand from this that there would be a signaling channel 58 from
`the right hand side post 102 when similarly upgraded to the ECP50. Golden
`also teaches the desirability of determining whether a connection should be
`admitted in the network but it does not provide details about how to do so.
`So if we turn to slide 10 Fichou provides details about how to admit
`connections within a network. Like Golden, Fichou describes a system that
`uses a centralized controller to provide a required quality of service for a
`requested connection from a source work station. So in Fichou's source
`work station that needs a reservation of a connection to a destination work
`station sends that request to a central controller and Fichou teaches that the
`server first authenticates those requests and then follows it up with what it
`calls a user rights verification where it checks a database that defines what
`types of requests each user is allowed to perform and then verifies based on
`parameters such as bandwidth, request of quality of service, and so forth.
`Now upon verification, Fichou teaches a distribution of what it calls a
`flow ID which can be, according to Fichou, an MPLS label. However
`Fichou and Golden do not provide details about that MPLS operation.
`So turning to slide and looking at Lee with an E -- slightly confusing
`from the prior proceeding, but this is a U.S. printed publication -- looking at
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`Lee it teaches centralized controllers controlling an MPLS network. In Lee,
`the centralized controller calculates a label switch path and distributes the
`appropriate labels to those switches in that path with what it calls an LFIB, a
`label forwarding information base, and the information in that LFIB as Dr.
`Reddy testified that is the information used by MPLS switches when
`performing look-ups for forwarding packets received that include an MPLS
`label as opposed to using a routing table.
`So let's move to slide 12. Slide 12 just provides an overview of the
`architecture of Golden and how the hardware elements taught in Golden read
`on the recited elements of independent claim 1. So we see here on the left
`hand side an originating end point that is the host 102 that is upgraded. It is
`in communication by signaling channels 58, or control paths, to ECP50
`which is a controller and that controller is in turn also in communication via
`respective signaling channels to these switches in the path and the network,
`and we've identified here on the left hand side a portal, according to the
`claim, that receives packets from the originating end point and decides how
`to forward them based on the MPLS labels according to the combined
`teachings of Lee, and finally, on the right hand side we see the terminating
`end point which again is this similarly upgraded host 102 which used to be
`the host/router 94.
`So let's turn now to slide 13 and really the first area of discussion
`today, and that is that a POSITA would have been motivated to combine
`Fichou and Lee with Golden.
`Turning now to slide 14. The petition demonstrated at first that a
`POSITA would have been motivated to combine Fichou with Golden for
`multiple reasons based on the evidence of record, including Golden itself,
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`Fichou's teachings and the declaration testimony of Dr. Reddy. For
`example, both Golden and Fichou teach the use of a centralized controller
`architecture to control the quality of service connections within a network.
`Further, Golden and Fichou both teach the desirability of controlling what
`types of connections to admit within the network.
`Golden, again as I noted, doesn't teach exactly how that admission is
`done. That's because Golden was an architectural disclosure. It was focused
`on many other aspects and left open to a POSITA how to fill in admitting
`connections within the network and so a POSITA would have been
`motivated to turn to Fichou because Fichou teaches such details, as Dr.
`Reddy testified and further Fichou teaches that its admission approach is
`advantageous by providing a way for the customer to manage the
`authorization for each user of an end point.
`JUDGE GIANNETTI: Why Fichou though? I mean there are
`probably lots of places you could look for a network architecture in details.
`Why Fichou of all of them? I hear your theory that you’ve got this broad
`architectural disclosure and fill in the detail if you turn to Fichou, but why
`specifically Fichou?
`MR. FOWLES: Well as noted Fichou teaches that it's advantageous
`to do this user rights verification so that each user is controlled and managed
`by the customer and that's relevant, for example, where there's a source work
`station and different people can sit down at that same work station and try
`and log into the network and request a service.
`JUDGE GIANNETTI: But why? Why would they want to do that? I
`mean you're giving me the end result, you're not giving me the rationale for
`why one would do 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
`
`8
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`
`MR. FOWLES: Why one would turn to Fichou?
`JUDGE GIANNETTI: Well, why would one make this combination,
`this particular combination?
`MR. FOWLES: Because if they want to know how to control
`admission to the network like Golden suggests they want to turn to
`something else that teaches how that's done and I believe you're right that
`there are many different references that could have been turned to by a
`POSITA.
`JUDGE GIANNETTI: Right.
`MR. FOWLES: However, Fichou is exemplary of how that is done
`and Fichou as a U.S. printed publication was readily accessible by a
`POSITA and they would have been able to find it.
`JUDGE GIANNETTI: All of that is interesting but it still doesn't
`answer my question. What is it about Fichou that suggests that you could
`use that particular source of details for this architecture? Why would you
`use it? Is there some suggestion (indiscernible) that this is the
`(indiscernible)? I mean why would one do this?
`MR. FOWLES: I think the suggestion in Golden that it's desirable to
`control admission of connections in the network is that suggestion to point
`one to Fichou because Fichou has details about how to control admission of
`requests in the network.
`JUDGE GIANNETTI: Thank you.
`MR. FOWLES: Looking here on the slide, Patent Owner's arguments
`about why the petition failed to establish focused on this concept that they
`were nothing more than conclusory statements provided in the petition.
`However, the quotes that were provided in the Patent Owner's response do
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`nothing more than selecting portions from the petition and ignoring the
`surrounding evidence and rationale. For example, as we've already
`addressed, Golden itself teaches that it's desirable to determine whether to
`admit a connection and here from column 10, lines 9 through 12, provides
`the quoted language from Golden that ECP50 could communicate with a
`quality server within the network for further determination of whether to
`admit the connection.
`Looking at the next portion there, Fichou provides those details and
`there are advantages in doing so, and finally Golden provides the that
`statement a POSITA to apply common knowledge, now as something that
`Dr. Reddy testified to and testified that Fichou is representative of that
`common knowledge. All of these together provide evidence from references
`themselves and the understanding of the POSITA as declared to by Dr.
`Reddy.
`JUDGE HOWARD: What about Patent Owner's argument that the
`Petitioner hasn't supplied enough evidence of how the references should be
`combined? What evidence is there of the mechanics of how you would take
`these three different references and put them all together and why?
`MR. FOWLES: Yes, Your Honor. If we could turn to slide 16.
`There are a couple of reasons that this argument should fail and first because
`the petition explained that a POSITA would apply Fichou's verification
`teachings to Golden's enterprise control point which is an example controller
`and specifically Fichou taught that its teachings could be separated or
`incorporated into existing functionality and Dr. Reddy declared that that
`existing functionality, for example, is the ECP50 in Golden, and so that's
`one way in which how the teachings of Fichou would be incorporated into
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`Golden. It's modifying Golden's ECP50 with this further verification
`teachings of Fichou.
`As another example, the petition explained that a POSITA would
`understand Golden to teach that its ECP50, that its control includes control
`over MPLS network elements generally. It was a known quantity and was
`known to exist and Lee, turning to Lee for a second, Lee provides details
`about how a centralized controller, again modifying Golden's ECP50, would
`determine the label switch paths and provide those labels to the switches
`determined in the path.
`These are all examples about how a POSITA would have modified
`Golden with the different teachings of Fichou and Lee, and beyond that it
`would sound like Patent Owners are also requiring bodily incorporation
`which is not the standard as I understand it. Rather, the petition and Dr.
`Reddy put forth evidence about how the combined teachings, about what the
`combined teachings would suggest to a person having ordinary skill in the
`art.
`
`JUDGE HOWARD: Thank you.
`MR. FOWLES: Yes. Turning back to slide 15, just to provide some
`examples about why a POSITA would have been motivated to combine Lee
`with Golden and Fichou in particular. Both Golden and Lee again teach the
`use of centralized controllers to control different aspects of their operations.
`They both teach operability with MPLS but again Golden doesn't teach
`details about how the labels will be determined and distributed, and Lee
`provides such details. As was already discussed today, MPLS was well
`known and Lee is but one example of how this could be done and
`specifically a POSITA would have been motivated to make 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
`26
`
`11
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`combination, as Dr. Reddy testified to, because as Lee describes it doing so
`at a centralized controller reduces the complexity and reduces the load for
`the switches and the network and all this evidence from Golden, Fichou and
`Lee, as well as Dr. Reddy, provides the evidence and the rationale about
`why a POSITA would have been motivated to make this combination.
`
`Really again Golden is inviting a POSITA to fill in the details
`and Lee provides those details, as well as provides the assurance just like
`you saw with Surdila that the routing would be done based on the labels
`according to the quality of service connection determined instead of the IP
`lookup. That's just by way of example. Dr. Reddy declared to this as well
`that a POSITA would have known that with MPLS an IP lookup would be
`replaced by an MPLS lookup in the LFIB.
`
`Turning now to slide 17. Unless there are any other questions
`on that point we move on to another area of discussion today and that is
`specifically that Golden in combination with Fichou and Lee renders
`obvious directing traffic based only on the routing instructions provided by
`the controller with no independent routing of the traffic.
`Slide 18 shows the relevant claim language for the based only
`language, and we move to slide 19 Golden teaches that the ECP50, and this
`is looking at what was laid out in the petition, that the ECP50 sends
`reservation requests to switches 56 and specifically the portal identified in
`figure 9 that we annotated and saw on slide 12. But staying here on this
`slide, a POSITA would have appreciated that the ECP in Golden performs
`centralized control functions relating to bandwidth reservation, and so then
`the petition turned to Lee which teaches specifically how to generate MPLS
`labels at the centralized controller as part of the LFIB information that is
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`sent to the MPLS switches, and finally, as Dr. Reddy declared, the LFIB
`taught in Lee constitutes routing instructions as used in this claim.
`Let's turn to slide 20. So the limitation specifically recites that the
`traffic for the connection is directed by the portal based only on the routing
`instructions. Much has already been discussed in the earlier proceeding
`about the use of MPLS which replaces the regular IP routing lookup with the
`MPLS lookup in the LFIB.
`If we turn to slide 23, we'll address specifically the argument used in
`the Patent Owner's response on this point, and that first focused on Lee and
`its description of the LFIB as information and that because it's information it
`can't be instructions, and this ignores the substance of Lee and the
`understanding of a POSITA what MPLS teaches with the form of a word
`choice, and that's based on the declaration testimony of Dr. Reddy again
`both in the original declaration of Dr. Reddy and as put here Dr. Reddy's
`confirmation of that in his supplemental declaration.
`"The LFIB's inclusion of an input label and output label and output
`interface and so on means the LFIB has all the information the MPLS switch
`needs to now forward using the LFIB instead of a traditional lookup."
`Turning to slide 24. Interestingly, Dr. Sharony cited to in Exhibit
`2003 in his declaration. Exhibit 2003 is from Cisco Press entitled MPLS
`Fundamentals Forwarding Label Packets and there it actually confirms and
`is consistent with Dr. Reddy's declaration testimony and the position put
`forward in the petition that an IP lookup is replaced with a lookup of the
`label in the LFIB when it is MPLS operation and also here looking at this
`middle quote,
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`
`"When a router receives an IP packet the lookup done is a IP lookup.
`But when a router receives a label packet the lookup is done in the LFIB of
`the router."
`So here we see again that all the evidence concentrates to the
`conclusion that when there is an MPLS label, the label is used to determine
`how to forward that packet out of the switch instead of a traditional IP
`lookup.
`JUDGE GIANNETTI: Just going back to the previous case where we
`had this discussion about the role of MPLS, I believe that -- and Mr. Vick --
`that there was a table lookup involved, even in the MPLS operation. Is that
`your understanding or is that some other process going on here? I know
`your argument focuses on the labels rather than on the lookup process, but
`Mr. Vick's argument suggested that there was a routing table involved in
`both cases.
`MR. FOWLES: For MPLS Routing I do not understand that to be the
`case, but that is not an issue that was raised in Patent Owner's briefing.
`JUDGE GIANNETTI: Right.
`MR FOWLES: We'd like to have our expert be able to look at that in
`more detail so that we could inform our decision based on that, but talking
`here today when the label is there the lookup done to determine how to
`cause that packet to leave the switch is done in the LFIB instead of a
`traditional routing table lookup.
`JUDGE GIANNETTI: Okay. And tell me, remind me what LFIB
`stands for.
`MR. FOWLES: Label forwarding information base.
`
`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
`
`14
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`
`JUDGE GIANNETTI: Label forwarding information base. So this
`suggests there's still some sort of table lookup going on there, right?
`MR. FOWLES: Yes, Your Honor. It's in the LFIB.
`JUDGE GIANNETTI: But not a routing table.
`MR. FOWLES: That's right. The switch does not make any kind of
`independent decision on how to forward the packet because the centralized
`controller gave those labels that tell it how to forward that packet.
`JUDGE GIANNETTI: Okay.
`MR. FOWLES: Any additional detail is not something that the claim
`requires and the specification of the 119 patent does not explain how that
`forwarding is accomplished. It simply claims the resulting function of new
`independent routing, and Dr. Reddy testified that MPLS labels, because
`they're used, means that there is no independent routing because it doesn't
`look at what the traditional concept of routing is of that IP lookup.
`JUDGE GIANNETTI: So what is your rationale for using MPLS as
`opposed to IP routing? Is that something that's suggested by your principal
`reference here? What's your argument for why we should be looking in
`MPLS?
`MR. FOWLES: Well first Golden does teach the use of MPLS in
`some of its network elements and specifically teaches that its architecture is
`designed to be interoperable with multiple protocols and as we've discussed
`here today and both sides seem to be in agreement, MPLS was a very well
`known protocol even at the time of 2006.
`JUDGE GIANNETTI: And more efficient, faster? Is that its virtue?
`MR. FOWLES: Well it's definitely --
`
`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
`
`15
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`
`JUDGE GIANNETTI: What are the virtues of MPLS over IP
`routing?
`MR. FOWLES: One virtue is that it forces packets to adhere to a
`specific flow and here in this evidence, as Lee teaches, that flow is
`determined by the centralized controller and that centralized controller when
`those teachings of Lee are incorporated into Golden's ECP50 means that that
`ECP50 sends those labels to the switch and two things, 1) that flow now
`must follow those labels so that the requested quality service is actually
`achieved and 2) it actually, as Lee teaches, provides the advantage of
`reducing the computational burden at the switch itself, and so that does make
`the switch better, more efficient and able to better achieve its enhanced
`functionality goal of sending things to the centralized controller so that
`something that has a picture of the entire network is able to decide how best
`to meet a request of quality of service.
`JUDGE GIANNETTI: So why hasn't MPLS just replaced IP
`addressing? Why isn't that the standard now?
`MR. FOWLES: MPLS still exists, but beyond that I couldn't answer
`that question. I think that it's because there are different needs for different
`customers and so it's really tailored to what the particular customer wants for
`their network and for their operations.
`JUDGE GIANNETTI: Thank you.
`MR. FOWLES: Yes. I'll note that actually MPLS is still very much
`discussed in the literature even today. Turning now to slide 26. There's
`another argument that Patent Owner raised in their response that Lee does
`not suggest let alone teach that the MPLS OAM, which stands for operation
`administration and maintenance function, could be implemented exclusively
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`in the centralized control apparatus. This is apparently because, "Lee
`explicitly teaches that the switch can receive an LSP which stands for label
`switch path from the MPLS OAM function instead of from the centralized
`control apparatus."
`But what this argument fails to consider is that that description in Lee
`is of just one embodiment that addresses what happens in failure scenarios
`and in deposition Dr. Sharony acknowledged that in non-failure scenarios
`and normal operation just quote here, "In the teachings of Lee in a normal
`situation where there are no failures the central controller will calculate the
`LSPs."
`That's correct. That agrees with Dr. Reddy's testimony that what
`happens in regular operation is that Lee teaches a centralized controller,
`which is modifying Golden's teachings, determines this label switch path
`and sends the labels that are determined from that to the switches that are
`less effective.
`If there are no further questions on this point I'd like to move on to the
`next area of discussion. Let's move to slide 27 then. The next area of
`discussion touches on the fact that the petition established that Golden in
`combination with Fichou and Lee renders obvious a required route
`supported by the portal and dynamically provisioned by the controller.
`Let's turn to slide 29. This lays out how the petition established how
`the combination renders this limitation obvious and this slide focuses on the
`language about the dedicated bearer path and the required route. What's
`worth noting here is that Golden again teaches that it reserves this path from
`beginning to end as modified by Lee's teachings the MPLS labels. So that
`provides the required route that is at the portal that the portal supports, again,
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`because it receives these labels from the centralized controller per Lee's
`teachings.
`Turning to the next slide, slide 30, and focusing specifically on the
`language a required route that is dynamically provisioned. Each switch in
`the path, according to Golden's teachings, receives reservation commands
`from the ECP, which is the enterprise control point again, and each switch
`including the portal dynamically provisions its route that is part of the end to
`end path. That is a quote from Dr. Reddy's initial declaration testimony and
`Golden throughout teaches that it really is establishing what it calls an on
`demand reserve bandwidth virtual circuit connections and finally it teaches
`that you can take down these connections once they're done or released, in
`Golden's terminology.
`This is an example of dynamically provisioned and why is that the
`case? Let's turn to slide 32. Let's look first at the comparison that Patent
`Owner has made of what Golden has in its pre-computed path list to the
`claimed required route that is dynamically provisioned. This is again
`comparing apples to oranges because Golden doesn't describe those pre-
`computed lists as being allocated yet. This is merely Golden describing the
`centralized controller becoming aware of what's in the network. The petition
`never compared the pre-computed list to the required route. Rather it looks
`at the virtual circuit connection that is set up in response to a request. This is
`a dynamic provisioning of a required route because it is making a link
`available with requested parameters.
`I say that in that way turning now to slide 33, actually 34, because at
`deposition when Dr. Sharony was asked what provisioning means, for
`example, dynamic provisioning, he answered that it is, "You basically sort of
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`18
`
`
`
`Case IPR2017-01661
`Patent 8,036,119 B2
`
`set up certain parameters or metric, certain parameters in order to support a
`given performance," and this actually agrees with Dr. Reddy's testimony that
`what a POSITA would understand with dynamic provisioning is that it
`would involve insuring enough of a desired parameter, for example,
`bandwidth, would be made available for the requested purpose. Dr. Sharony
`never testified that dynamic provisioning would preclude looking to pre-
`existing paths to then set up a connection. Just because those paths are
`known, does not mean they're already reserved for the end point that's
`requested. In fact, Golden goes through this whole process to demonstrate
`how to actually get one of those paths reserved for its desired quality of
`service connection.
`If there are no further questions, I'll turn to the next area of discussion.
`JUDGE HOWARD: Counsel, just to remind you you have about five
`minutes left before you get into your rebuttal time.
`MR. FOWLES: Thank you, Your Honor. Slide 35 is the last area of
`discussion for today so I'll wrap up with that and that is this question
`whether Golden teaches a control path from the controller to the terminating
`end point. I've already explained how Dr. Reddy testified how a POSITA
`would understand Golden's teaching that figure 9 is host/router 94.
`Let's turn to slide 37, please. Reproduced again here this figure that
`that host/router 94 on the right hand side would be "similarly upgraded" as
`host 102 on the left hand side and let's turn to slide 38. Dr. Sharony --
`JUGE HOWARD: Just going back to 37 for a second.
`MR. FOWLES: Oh, yes please.
`JUDGE HOWARD: Is it fair to say that your argument on this
`limitation is dependent on our agreeing with the modification that Dr. Reddy
`
`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 IPR2017-01661
`Patent 8,036,119 B2
`
`made to figure 9? That if we didn't believe that supported making that
`modification, the limitation wouldn't be met?
`MR. FOWLES: The way that we set it up in the petition, yes, You