`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`_____________
`
`
`ZSCALER, INC.,
`
`Petitioner,
`
`v.
`
`SYMANTIC CORPORATION,
`
`Patent Owner.
`_____________
`
`IPR 2018-00920
`Patent 9,525,696 B2
`_____________
`
`RECORD OF ORAL HEARING
`Held: August 8, 2019
`_____________
`
`
`
`
`Before JEFFREY S. SMITH, BRYAN F. MOORE, and NEIL T. POWELL,
`Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`DONALD DAYBELL, ESQUIRE
`JARED BOBROW, ESQUIRE
`
`Orrick, Herrington & Sutcliffe LLP
`2050 Main Street
`Suite 1100
`Irvine, CA 92614
`
`
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`BRYAN PARRISH, ESQUIRE
`CHAD WALTERS, ESQUIRE
`Baker Botts, LLP
`2001 Ross Avenue
`Dallas, TX 75201
`
`
`
`
`
`
`
`The above-entitled matter came on for hearing on Thursday, August 8,
`2019, commencing at 3:22 p.m., at the U.S. Patent and Trademark Office,
`600 Dulany Street, Alexandria, Virginia, before Chris Hofer, Notary Public.
`
`
`
`
`
` 2
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`
`
`
`
`
`
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE SMITH: Welcome back everybody. We're here for
`
`
`the second case today. This is IPR 2018-00920. The procedure for this
`hearing will be the same as the last one. Each side will have 30 minutes to
`make their case and each side can reserve time for rebuttal. Petitioner will
`go first. Petitioner, do you wish to reserve time for rebuttal?
`
`
`MR. DAYBELL: Yes, Your Honor. We'll reserve ten minutes
`for rebuttal.
`JUDGE SMITH: Okay. And Patent Owner, do you wish to
`
`
`reserve time for rebuttal?
`
`
`MR. WALTERS: Yes, Your Honor, five minutes please.
`
`
`JUDGE SMITH: Five minutes? Okay. And Petitioner, can
`you step up to the podium and state your appearance.
`
`
`MR. DAYBELL: Yes, Your Honor. Don Daybell appearing
`for Petitioner Zscaler.
`
`
`JUDGE SMITH: Thank you, and Patent Owner can you step
`up to the podium and state your appearance.
`
`
`MR. WALTERS: Your Honor, this is Chad Walters and my
`colleague Bryan Parrish will be handling the argument for this IPR.
`
`
`JUDGE SMITH: Thank you.
`
`
`JUDGE MOORE: And Petitioner, just since it's a different
`record you might want to mention who you have with you just so it's on the
`record.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
`
`
`
`
` 3
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`MR. DAYBELL: Thank you, Your Honor. With me is my
`
`
`colleague, Jared Bobrow.
`
`
`JUDGE SMITH: Okay. Petitioner, you have 20 minutes to
`make your presentation and you may begin when ready.
`
`
`MR. DAYBELL: Thank you, Your Honor. So there are four
`grounds at issue in this case but there's really only one disputed issue and it
`has to do with the network processing module limitation of claims 1 and 13.
`Patent Owner highlights the subscriber profile as well but as we will show,
`the dispute there is largely subsumed within the dispute over this claim
`limitation, and the dispute is whether the cited prior art teaches what is
`referred to in claim 13 --
`
`
`JUDGE SMITH: Sorry, can you -- when you use your
`demonstratives can you refer to --
`
`
`MR. DAYBELL: I apologize, Your Honor. Yes, referring to
`slide 3, the dispute is over the language in claims 1 and 13 referring to a
`network processing module and in particular the steps for identifying one or
`more of the plurality of application processing modules for processing the
`identified data packets based on an association of the application configured
`on each application processing module with the subscriber profile.
`
`
`On slide 4 it is Petitioner's contention that Nortel which is the
`ground 1 and 2 in view of the knowledge of a PHOSITA as well as the
`combination of Alles and Lin which is grounds 3 and 4 disclose this
`limitation. Now, as the parties have noted throughout their briefing Nortel is
`essentially, it's not word for word identical but it's essentially the same
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 4
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`substantive disclosure as the combination of Alles and Lin so all of the
`arguments that both sides make that apply to Nortel apply largely equally to
`Alles and Lin and we're going to focus on Nortel in this discussion with
`citations to the Nortel reference primarily.
`
`
`So a quick sort of summary of how the Nortel reference works
`is it teaches an internet service node which is used to police traffic flow in
`between access ports and trunk ports. The access ports are where the users
`are sitting and the trunk ports go back to the internet. So this traffic is
`processed by what Nortel refers to as a packet service card or packet service
`cards and those cards receive the traffic from the switch fabric 440 as shown
`in figure 4 on slide 5. This switch fabric is configured or controlled by a
`router service management card RNC460 and that card in turn is configured
`by a configuration manager 470, and as we will show or have shown in our
`papers that configuration manager uses service policies.
`
`
`So the Nortel reference also teaches a CAM circuit. Cam
`stands for --
`JUDGE SMITH: And this is page 6 --
`
`
`MR. DAYBELL: Sorry, slide 6, yes. The Nortel reference also
`
`
`teaches a CAM circuit, CAM being a content addressable memory. This
`circuit sits on the data path between the ports and the switch fabric as can be
`seen in figure 8. The data path starts on the right hand side. The data comes
`in through a framer 810 and then is received by a sign of logic 850 that uses
`the CAM 820 to identify a processor that is going to process the traffic and it
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
`
`
`
`
` 5
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`makes that identification based on information about the subscribers and the
`services that they wish to use.
`
`
`The CAM and the assignment logic are also configured by the
`RNC that we talked about previously on slide 5, the RNC 460 and that
`configuration goes through this processor interface 830 which you can see in
`the upper right hand corner of figure 8 on slide 6.
`
`
`So essentially what we have here is a configuration manager
`470 and now on slide 7, a configuration manager 470 that communicates
`configuration parameters which are determined by the service policies and
`this is a quote from Nortel, it communicates those parameters to the RNC
`460 and then the RNC in turn configures the different components to
`provide the service policies. So the CAM and the switch fabric are both
`configured by the RNC to have these service policies and the CAM therefore
`implements the service policy.
`
`
`Now you question about what is a service policy? Well, Nortel
`shows us that it has a figure 6A on slide 8 I am now. It shows what the
`service policies look like. Each of the rows in figure 6A is a service policy
`and it's made up of two parts. The first part on the left is a series of columns
`that Nortel uses as classifiers which contain information about the
`subscribers. For example, the source ITSRC is the source IP address for the
`subscribers. The DST is the destination IP address for the traffic going back
`and forth. SDC is the type of traffic. Nortel describes this as illustrating
`subscriber policies for a subscriber, this figure, and showing how you can
`use policy rules to implement features such as security.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 6
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`So let's look at each row. So for example, row 1 of figure 6A
`
`
`on slide 9 I am now, shows that the service policy which is configured into
`the CAM identifies the data flow using in this case three classifiers. The
`first one is if the source IP address is either subsection or substation A or if
`it's office 1 and the destination is the opposite of that. So this is looking at
`traffic between substation A and office 1 and in particular it's looking at
`IMAP, which is a protocol for email, it's looking at email traffic between
`these two groups and it says if you have email traffic between these two
`groups, if that's your subscriber profile then the application or the action you
`should perform is to accept the traffic and encrypt it using what's called
`triple DESs, the DES algorithm applied three times and this is the basic
`functionality of the Nortel (indiscernible.)
`
`
`So the dispute between the parties centers on how this CAM
`that we talked about works and we've explained in our papers that this CAM
`implements the service policy as provided through the configuration
`manager 470, and I'm on slide 10 now, and it does this using a classifier
`which is the first three columns are classifiers and it does this to identify a
`processor and the processors are the things that run the applications on the
`right hand side of slide 8 -- and switching back to slide 10 -- it runs this
`application that's specified by the service policy.
`
`
`Now let's look a little bit about how the CAM works. The
`CAM itself is a content addressable memory, so it receives -- it has a search
`field which is figure specified as No. 711 and that is the field that contains
`the information about the subscribers, the subscriber profile information, and
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 7
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`has an output field which is shown here in 715 on slide 11. That's going to
`have the processor identifier. It's going to identify which processor is
`matched to that particular subscriber or the particular group of subscribers
`and the way it works is it receives data, input data 701 that comes in and it
`compares the input value 701 with the search field 711, goes to until it finds
`a match and when it finds a match in the search field which contained the
`subscriber information it outputs a processor number from the column 715,
`outputs that on line 749 and tells the switch to send this traffic to that
`particular processor and that's again cited in Nortel. It says,
`
`
`"When the CAM generates the output field, that field identifies
`the processor to which the cell or the packet may be assigned for further
`processing."
`
`
`So on slide 12 we know that this processor in some instances is
`a specialized processor and Nortel teaches us this and it teaches that in those
`situations the CAM identifies the single specialized processor that is
`dedicated to providing a specific application. That's the nature of a
`specialized processor. So for this situation at least the CAM is aware of the
`specialized processor and since the processor is specialized it can only do
`one application. It's also aware of the application.
`
`
`Now there are other situations where the processor is not
`specialized but even in those cases the CAM will direct a subscriber's traffic
`to the processor that is identified in the service policy that we see in figure
`6A of this reference and the way the -- so what we have here is we have this
`CAM circuit which is configured as I indicated before with a service policy
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 8
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`and it's configured by the configuration manager to identify a processor to
`process the identified data packets based on the service policy.
`
`
`JUDGE SMITH: Let me just -- if I maybe just kind of jump to
`and to maybe anticipate Patent Owner's argument, it sounds like from what I
`can tell what Patent Owner is saying is these actual policies that associate
`the user with the action and is doing that using these source and destination
`addresses and an action with it. That whole thing is the policy and that
`whole thing is not implemented in the CAM. The only thing the CAM does
`is identify the user using the IP address and then from the IP address there's
`a corresponding module to send it to -- and that corresponding module is --
`well let me back up. Initially from what I understand in Nortel there are
`users that are associated with service policies --
`
`
`MR. DAYBELL: Yes.
`
`
`JUDGE SMITH: -- so for each module that can implement that
`service policy to that user, that's what's programmed into the CAM, the
`actual modules.
`
`
`MR. DAYBELL: Right. The CAM --
`
`
`JUDGE SMITH: So the CAM has information about the
`modules and then it routes the data to that module --
`
`
`MR. DAYBELL: Correct.
`
`
`JUDGE SMITH: -- and then that module then again looks at
`the IP address and from there looks at the rules to figure out what action to
`apply to it and then applies that action. So I think from what I understand
`Patent Owner's argument is that because the action itself is not included in
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 9
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`the CAM, it's not implementing the service policy but, you know, looking at
`the claim I don't see where the claim requires the actual -- it just says
`network processing module to identify an application module to process the
`data packet based on an association of the application with the subscriber,
`and that's done in the CAM with the IP address and then the corresponding
`module.
`MR. DAYBELL: You are correct, Your Honor. The claim
`
`
`does not require the action to be in the network processing module. The
`claim simply requires that the network processing module decide which
`processor to send the subscriber's information to based on what it says is an
`association between the subscriber profile and the action. The service policy
`is that association and the CAM's decision is based on the information in the
`service policy.
`
`
`JUDGE SMITH: Okay. So I mean what, you know, I guess
`there's a disconnect between what I'm reading about from you in your briefs
`and what I'm reading about from Patent Owner in Patent Owner's brief. Like
`the disconnect is you're saying the CAM implements the service policy.
`Patent Owner's saying the CAM doesn't implement the service policy.
`Maybe --
`MR. DAYBELL: Sure. To be clear --
`
`
`JUDGE SMITH: -- the CAM is implementing determining
`
`
`who the subscriber is from the IP address and then sending it to the
`corresponding processing module.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
`
`
`
`
` 10
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`MR. DAYBELL: Correct. And to be clear what we are saying
`
`
`is that the CAM implements a portion of the service policy.
`
`
`JUDGE SMITH: Okay.
`
`
`MR. DAYBELL: It implements the classifier portion which I
`have in slide 88 (phonetic).
`
`
`JUDGE SMITH: By classifier you mean it identifies the user
`from the IP address.
`
`
`MR. DAYBELL: Among other things, yes. It's more than just
`the IP address but yes, it uses the source and destination IP address as it uses
`the type of application, the service. It uses a number of other pieces of
`information in the IP header.
`
`
`JUDGE SMITH: To identify the --
`
`
`MR. DAYBELL: The (indiscernible.)
`
`
`JUDGE SMITH: -- the user. To identify the user.
`
`
`MR. DAYBELL: Or a group of users, but yes.
`
`
`JUDGE SMITH: Okay.
`
`
`MR. DAYBELL: Yes, yes it does. And then once it's
`identified the user or a group of users which is the subscriber profile it uses
`that information to decide which packet service card which is where the
`processors are.
`
`
`JUDGE SMITH: Okay.
`
`
`MR. DAYBELL: So yes, so the CAM implements a portion of
`the service policy and therefore the decision the CAM makes is based on the
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
`
`
`
`
` 11
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`service policy and the service policy provides the association between the
`subscriber profile and the application.
`
`
`So slide 14 is just an example of some of the fields that the
`CAM can look at. It can look at the destination IP address, it can look at the
`source address, it can look at port numbers, port numbers correlate to the
`services being provided. It can look at protocol types, it can look at other
`information.
`
`
`JUDGE SMITH: This is to identify the subscriber?
`
`
`MR. DAYBELL: Correct.
`
`
`JUDGE SMITH: And this would -- okay. And this would, I'm
`just mapping this back to the claim. So the claim has subscriber profile for
`identifying data packets associated with the subscriber profile and that's
`what's being done here.
`
`
`MR. MAYBERRY: Correct. The subscriber profile is the
`classifiers that are shown in figure 6A of Nortel.
`
`
`JUDGE SMITH: Okay.
`
`
`MR. MAYBERRY: As seen on slide 8. So speaking of
`classifiers, one of Patent Owner's arguments is that we fail to identify a
`subscriber profile because we fail to identify a classifier in Nortel's CAM.
`This is just incorrect. I mean they assert that Nortel teaches some special
`definition of classifier that requires a data flow and the conditions under
`which the action is applied and that's just not true. I mean Nortel clearly
`teaches that some classifiers contain no conditions as we see --
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
`
`
`
`
` 12
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`JUDGE SMITH: If you could, I mean this is another thing I
`
`
`struggled with reading the briefs from both parties. The claim itself talks
`about a subscriber profile for identifying data packets associated with the
`subscriber profile and there's a lot of argument about does Nortel disclose a
`classifier or not disclose a classifier. Maybe if you could discuss this in
`terms of what the claim language actually says and kind of -- so the
`subscriber profile is some type of data structure that maps an IP address to a
`subscriber so the --
`
`
`MR. DAYBELL: Correct.
`
`
`JUDGE SMITH: -- is that right? So you have this data
`structure that says if the IP address is such and such, then this is the
`corresponding subscriber.
`
`
`MR. DAYBELL: That's one example, yes. That would be for
`example the second to last entry on slide 8.
`
`
`JUDGE SMITH: Well I'm just talking about the claim right
`now, just the scope of the claim. Because I mean this limitation itself is kind
`of difficult to understand. So you have a subscriber profile which to me just
`seems to be data by itself, it's not stored in memory, it's not read from
`memory and executed by a processor. You just have this data, so you have
`data for identifying data packets associated with itself. I'm just trying to
`make sense out of this and then map that on to Nortel.
`
`
`MR. DAYBELL: I agree with you that the language of that
`limitation is a little obtuse. At this risk of putting words in the patent
`Applicant's mouth I suspect they may have wanted that to say data packets
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 13
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`associated with the subscriber in a stream of data packets. But you're right,
`it is self-referential.
`
`
`JUDGE SMITH: But I mean to meet this limitation, if you had
`some type of data structure for identifying, maybe some type of data. Like it
`could be an IP address that identifies data packets associated with a
`subscriber.
`MR. DAYBELL: Yes.
`
`
`JUDGE SMITH: That would meet this limitation?
`
`
`MR. DAYBELL: That would meet this limitation.
`
`
`JUDGE SMITH: Okay.
`
`
`MR. DAYBELL: Yes. And that is our point is that the
`
`
`classifier is not something specific that requires conditions as well as
`information specifying a data flow, it just has to have enough information to
`identify the subscriber which is the data flow pertains to subscriber so it has
`to identify the subscriber.
`
`
`I think my time is nearing an end so in the time I have left --
`
`
`JUDGE SMITH: You have three minutes.
`
`
`MR. DAYBELL: Okay. So Nortel itself has a description in
`two different places. The first of how the CAM entries are configured and
`the second of how service policies are implemented, this is slide 17, and
`these two descriptions are very similar to each other. It makes it clear from
`these that the CAM when it is configured is implementing a service policy.
`If we compare the language on the bottom is the service policy
`implementation. It talks about an IP address to be generated dynamically
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 14
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`and we see that same language in the CAM configuration where the entry of
`a CAM may be configured dynamically. The service policy it speaks of if
`it's being used to establish a dial in connection, the same dial in connection
`is used in the CAM configuration example. They both speak of assigning
`the user a new IP address and the last one speaks of generating the
`processing rules which flow from the service policies and then the first one
`speaks of generating a CAM entry. So we can see from this that the CAM is
`doing all of the things that are tied up with implementing a service policy.
`
`
`Now the Patent Owner has a lot of discussion about all of the
`things that a packet service card does and they say that the packet service
`card must determine the flows and they rely on a passage which talks about
`the packet service card determining flows. But that's not the only place that
`this flow determination is done and in fact Nortel itself teaches that using a
`design with a packet service card determines flows led to problems with the
`assignment of processors and it teaches that that assignment has to be done
`quickly and for speed you base the assignment on examination of cell
`headers, and then the very next thing it teaches is that their invention
`examines cell headers using the CAMs and it uses the CAMs for
`determining the subscriber and thus the processor force for processing.
`
`
`So at the end of day, the way this is all put together is you have
`the CAM and the packet service card that together implement the service
`policy. The CAM uses the subscriber profile to decide which subscribers or
`groups of subscribers are going to be routed to each processor and then it
`routes them to the processor and the processor in the packet service card
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 15
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`performs whatever the action is that's specified in the service policy, and I
`think with that I am at my time. I'll reserve the rest of my time for rebuttal.
`
`
`JUDGE SMITH: Thank you. Patent Owner, you have 25
`minutes. You may begin when you are ready.
`
`
`MR. PARRISH: Good afternoon, Your Honors. My name is
`Bryan Parrish, may it please the Board. As Your Honors are probably well
`aware now the parties have a fundamental disagreement about what these
`references teach so what you're going to see today in my presentation is as
`much as possible focus on the actual language of what these references say.
`Petitioner went ahead and went over some of the introductory information so
`I'm going to skip ahead to slide 10 where we're going to go through what
`Nortel actually describes is the operation of the internet service node.
`
`
`When data flows come into the internet service node they come
`in through the access ports and the access ports described as the physical
`interface to the internet service node. On slide 11 what we see is Nortel
`describes the CAM as being located in the ports of the internet service node
`and that makes sense. It's the job of the CAM to efficiently and quickly
`route data flows to the packet service cards that they're supposed to go to
`and what we see from Nortel is that the CAM has three fields. It has a
`search field, it has a mask field and it has an output field which you don't see
`described at all in Nortel as the CAM having an application field. What we
`see is that when bits come into the CAM from the access ports, those bits are
`masked. The masked portions of the bits are compared to the search field
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
`
`
`
`
` 16
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`and if there's a match then the value in the output field is output on to the
`upper (phonetic) bus.
`
`
`JUDGE SMITH: Can you just tell us why the claim requires
`that? Why does the claim require that?
`
`
`MR. PARRISH: I'm sorry, Your Honor? What does the claim
`require?
`JUDGE SMITH: You're saying that the CAM in the prior art
`
`
`reference doesn't have an application field; is that what you're saying?
`
`
`MR. PARRISH: Yes. One of the --
`
`
`JUDGE SMITH: Does the claim require that? Does the claim
`require the network processing module to have a field for an application?
`
`
`MR. PARRISH: We contend that it does, Your Honor. What
`we see is the --
`
`
`JUDGE SMITH: Where is the support in the spec for this
`limitation?
`MR. PARRISH: So I can fast forward to the actual claim itself
`
`
`and I can go through the claim language.
`
`
`JUDGE SMITH: Okay.
`
`
`MR. PARRISH: So the claim requires a network processor --
`
`
`JUDGE SMITH: What slide is this?
`
`
`MR. PARRISH: This is slide 5. The claim requires a network
`processing module for identifying one or more of a plurality of application
`processing modules for processing the identified data packets and the key
`part here is based on an association of the application configured on each
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 17
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`application processing module with a subscriber profile, and so what this
`claim element requires is several things. It requires a network processing
`module. It requires a subscriber profile which we saw Petitioner maps to
`classifiers and processing rules as applied by the CAM and we see that the
`third requirement is there has to be an association of the application
`configured on each application processing module with a subscriber profile.
`
`
`JUDGE SMITH: Well no, that's not what the claim says. The
`claim just says it has to identify the application processing modules based on
`an association.
`
`
`MR. PARRISH: Yes, Your Honor.
`
`
`JUDGE SMITH: But you don't have to make that association
`to determine an association. It just has to identify a module based on the
`association.
`MR. PARRISH: It's our contention that the CAM can't make
`
`
`that association unless it has any knowledge of the application. How can the
`CAM decide where to send it based on the association --
`
`
`JUDGE SMITH: How does your spec decide where to send it?
`
`
`MR. PARRISH: Sorry?
`
`
`JUDGE SMITH: How does your spec -- where is the support
`in the spec for this limitation?
`
`
`MR. PARRISH: Yes, Your Honor. I'll back up to the overview
`of the 696 patent. What we see in the 696 patent is the description of what
`the subscriber profile does and this is the response on page 89. What we see
`is the present invention may include a subscriber profile. This profile may
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 18
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`relate in application to a subscriber. So what we have, as you alluded to in
`your questioning earlier, is we have a subscriber profile and the whole
`concept of the subscriber profile is to associate data flows associated with
`that subscriber to applications. So the subscriber profile itself has this
`collection of applications and data flows and maps those two things together.
`
`
`JUDGE SMITH: So what you're quoting from the spec is not
`really -- so this is for the subscriber profile limitation. This subscriber
`profile in the spec is not really in the claim. The claim says the subscriber
`profile identifies the data packets associated with the subscriber profile.
`What you're pointing to in the spec says the subscriber profile relates an
`application to the subscriber so this is not the subscriber profile that's in the
`claim. The subscriber profile in the claim identifies data packets associated
`with the subscriber profile. This column 37 that you're pointing to, I don't
`see anything in here about identifying data packets. Is there anything in this
`column 37 about identifying data packets with the subscriber?
`
`
`MR. PARRISH: Yes, Your Honor. If you look at the actual
`column itself this portion of the spec talks about how data can be processed
`by the application processing modules you see in figure 2 -- by the network
`processing modules you see in figure 2 and what happens is that when the
`data flows come in they need to be associated with a subscriber and when
`they're associated with a subscriber they're associated with this subscriber
`profile and so that's what we see is in the claim is there has to be the
`subscriber profile and the network processing module and it's or identifying
`data packets associated with the subscriber profile.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 19
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`JUDGE SMITH: Does the subscriber profile have to be in a
`
`
`network processing module?
`
`
`MR. PARRISH: Sorry, I misspoke. The flow processing
`facility has to have a subscriber profile and if you look at figure 2 --
`
`
`JUDGE SMITH: Flow processing facility has to have a
`subscriber -- where's the flow processing facility?
`
`
`MR. PARRISH: The flow processing facility is element 102.
`It's that entire box in figure 2.
`
`
`JUDGE SMITH: No, in the claim I'm asking?
`
`
`MR. PARRISH: In the claim it's in the preamble, Your Honor.
`
`
`JUDGE SMITH: Okay. So it's 402 you said? So 402, that
`flow processing facility looks at the IP addresses and determines that the
`associated subscriber from the IP addresses; is that right?
`
`
`MR. PARRISH: What's described in the claim and in the
`patent is that the network processing module takes in these data flows.
`You'll see here it's element 210 in the patent and the figure, and those are
`associated we see in the claim as that information is associated with a
`subscriber profile and there has to be an identification of which application
`processing module to send the data flow to and you have to make that
`identification based on an association of the application configured on each
`application processing module with a subscriber profile.
`
`
`JUDGE SMITH: The claim says you have to do that?
`
`
`MR. PARRISH: Yes, Your Honor. The network processing
`module, it's the second highlighted box.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 20
`
`
`
`IPR 2018-00920
`Patent 9,525,696 B2
`
`
`JUDGE SMITH: The claim just says, well okay. I was looking
`
`
`at column 90 in the patent as support for this claim limitation that the
`network processing module -- is column 90 not the support for it, because
`the column 37 does not say anything about routing the packets to an
`application processing module.
`
`
`MR. PARRISH: As you can tell this patent has I think over
`100 columns. There's multiple inventions listed in this patent.
`
`
`JUDGE SMITH: But there's just one limitation that we're
`talking about. I'm just asking where is this one limitation supported?
`
`
`MR. PARRISH: This network processing module limitation?
`
`
`JUDGE SMITH: Yes.
`
`
`MR. PARRISH: The network processing module and the
`hardware of figure 2 are described in more detail, I believe it's roughly
`starting in column 30. If you look at line 19 it says as a data flow is received
`at the network processing module and it talks about packet headers and
`payloads may be inspected, and then if you skip over the next couple of
`columns discuss different ways you can process the data flow.
`
`
`JUDGE SMITH: Let me ask you about column 90.
`
`
`MR. PARRISH: Ninety?
`
`
`JUDGE SMITH: Yes, 90. And I don't want to take up too
`much of your time.
`
`
`MR. PARRISH: No, I'm here to answer your questions.
`
`
`JUDGE SMITH: I just want to understand what you're -- so
`column 90, the first full paragraph