[00:00:01] Speaker 02:
Judge Hughes will not be with us this morning, but he will be listening to the tape and fully participating in the case.

[00:00:18] Speaker 02:
We have five cases on the calendar this morning, three from the PTAB and two from district courts, one of which is being submitted on the briefs and will not be argued.

[00:00:32] Speaker 02:
The first case is Hacken Intelligence versus Juniper Networks and Palo Alto Networks, 2022, 1938, 1400, 1401, et cetera.

[00:00:52] Speaker 02:
Mr. Bullwicker, please proceed.

[00:00:59] Speaker 00:
Good morning, Your Honors, and may it please the Court.

[00:01:02] Speaker 00:
Alan Bullwinkle on behalf of Packet Intelligence.

[00:01:04] Speaker 00:
There are three main points that I would like to address today.

[00:01:07] Speaker 00:
First, the Board erred in its construction of conversational flow by omitting portions of the patentee's definition.

[00:01:15] Speaker 00:
And second, based upon this erroneous construction, the Board improperly found that Riddle discloses conversational flows.

[00:01:23] Speaker 00:
And finally,

[00:01:24] Speaker 00:
the board properly found that petitioners failed to carry their burden to show that Wakeman discloses the associative cash subsystem.

[00:01:33] Speaker 00:
Turning to my first point, the board erred by failing to adopt the patentee's explicit definition of conversational flow.

[00:01:41] Speaker 00:
Instead, the board only adopted a portion of that definition.

[00:01:45] Speaker 01:
The portions that the board did not include are all exemplary.

[00:01:50] Speaker 01:
They don't seem to be limiting in any way.

[00:01:52] Speaker 01:
How would the scope of the

[00:01:54] Speaker 01:
claim change at all if the board had added the exemplary language you wanted?

[00:02:01] Speaker 00:
The portions that the board excluded highlighted the importance of participants to a conversation.

[00:02:08] Speaker 00:
The language excluded provided an example of an activity.

[00:02:13] Speaker 00:
For instance, the running of an application on a server as requested by a client.

[00:02:19] Speaker 00:
And then later it explains that

[00:02:21] Speaker 00:
Some conversational flows involve more than one exchange of packets between a client and a server.

[00:02:27] Speaker 00:
So these portions that the board excluded highlight the conversational nature of a conversational flow.

[00:02:34] Speaker 00:
They have particular participants, a server and a client.

[00:02:37] Speaker 01:
Is highlighting the importance of certain elements that are already in the construction that the board did adopt, does it do anything to claim scope?

[00:02:49] Speaker 00:
It does here, Your Honor, because the board found, based upon its construction, that conversational flow is not limited to a single user or client.

[00:02:58] Speaker 00:
So in essence, the board found that a conversational flow doesn't have any relation to the participants in the conversation.

[00:03:06] Speaker 00:
Instead, it only relates to the activity, which it viewed as the type of application or protocol at issue in the network communications.

[00:03:16] Speaker 01:
How does your construction limit to just a single user?

[00:03:26] Speaker 01:
What portion of your proposed construction would have that impact?

[00:03:31] Speaker 00:
For instance, the running of an application on a server as requested by a client describing an activity.

[00:03:39] Speaker 00:
And the board's understanding of an activity is so broad that it ignores the participants in the conversation.

[00:03:47] Speaker 00:
The board is viewed a conversation as being any network traffic related to a particular application, regardless of the participants.

[00:03:56] Speaker 00:
And the specification confirms that the participants in the conversation are an important part of the conversational flow.

[00:04:04] Speaker 00:
And this is shown in the SAP example, which is what the board used as its justification for its construction and its interpretation that conversational flows are not limited to users or clients or the participants.

[00:04:18] Speaker 00:
And in the SAP example, this involves two connections.

[00:04:25] Speaker 00:
And the idea here is that within SAP, the applications that are available

[00:04:33] Speaker 00:
They have a dynamic address.

[00:04:34] Speaker 00:
So if you want to print, you have to ask the SAP server, where is the print service?

[00:04:39] Speaker 00:
And you get an address.

[00:04:40] Speaker 00:
And once the client has that address, it can then make a print request.

[00:04:45] Speaker 00:
And so client number one can make a request, get the address of the print server, and then submit its print request.

[00:04:52] Speaker 00:
That would be two different connections.

[00:04:55] Speaker 00:
But the packet monitor would recognize these two connections are related.

[00:04:59] Speaker 00:
and therefore are both part of the same conversational flow.

[00:05:03] Speaker 00:
And this is explained in the provisional at appendix 2503, lines 25 to 30.

[00:05:10] Speaker 00:
It says these two packet exchanges, these two connections, would then be correctly identified as being part of the same flow if the clients were the same.

[00:05:20] Speaker 00:
So it was important to note that the clients there were the same.

[00:05:23] Speaker 00:
And the 099 patent specification also clarifies this same point.

[00:05:28] Speaker 00:
at appendix 1319, column three, lines one through six, it says, if the clients were the same, the two packet exchanges would then be correctly identified as being part of the same conversational flow.

[00:05:42] Speaker 00:
So that's highlighting the importance of the participants in the conversation.

[00:05:47] Speaker 00:
Now another scenario, these same two packet exchanges, you have client number one requesting the address of the print server.

[00:05:55] Speaker 00:
it gets that address.

[00:05:57] Speaker 00:
Well, client number two overheard that address, and so now client number two knows the location of the print server.

[00:06:04] Speaker 00:
Because it knows that location, it doesn't need to ask for it.

[00:06:08] Speaker 00:
So client number two can now submit a print request directly without asking for that address.

[00:06:15] Speaker 00:
And the specification talks about that sequence of exchanges, and it explains that

[00:06:21] Speaker 00:
The packet monitor would recognize that client number two's communication with the print server.

[00:06:27] Speaker 00:
It would recognize that as a print activity because the packet monitor also previously saw that the SAP server said if this port gets used, it corresponds to the print activity.

[00:06:39] Speaker 00:
So that's a dynamically assigned port.

[00:06:42] Speaker 00:
It can change over time.

[00:06:44] Speaker 00:
And the importance here is that the monitor has the ability

[00:06:48] Speaker 00:
to recognize that client number two is engaged in a print activity.

[00:06:52] Speaker 02:
You want to talk about Riddle, the reference?

[00:06:56] Speaker 00:
Yes, your honor.

[00:06:56] Speaker 00:
So the way Riddle works is Riddle organizes traffic according to what it calls traffic classes.

[00:07:07] Speaker 00:
And traffic classes are defined according to a traffic specification.

[00:07:14] Speaker 00:
And so figure 2a at appendix 1900 of RIDDLE shows an example of some traffic classes.

[00:07:22] Speaker 00:
One of the traffic classes there is a web traffic class.

[00:07:26] Speaker 00:
Another one is FTP, or File Transfer Protocol.

[00:07:31] Speaker 00:
And so the way RIDDLE works is it gets a packet, and it will compare that packet to the traffic specifications for these traffic classes.

[00:07:39] Speaker 00:
And when it gets a match, that packet belongs to that traffic class.

[00:07:43] Speaker 00:
So to the extent that Riddle sees web traffic from various clients on the network, it would recognize that web traffic and put it in the web traffic class.

[00:07:55] Speaker 00:
The problem with Riddle and why Riddle doesn't teach or disclose conversational flows is that Riddle does not discern between the participants in the conversation.

[00:08:05] Speaker 00:
Riddle doesn't really understand the notion of conversations.

[00:08:08] Speaker 00:
Instead, it's just looking at a packet

[00:08:11] Speaker 00:
Does it match this traffic specification?

[00:08:13] Speaker 00:
And if it does, it belongs to that traffic class.

[00:08:18] Speaker 00:
Now petitioners have highlighted that RIDL can also have traffic classes that are specified for a predefined client.

[00:08:25] Speaker 00:
But that doesn't cure RIDL's deficiencies because even if you take, for instance, that FTP traffic class in Figure 2A, you could define that traffic class to only relate to a specific client.

[00:08:39] Speaker 00:
Well, at that point, anytime it sees FTP traffic involving that client, it's going to recognize that traffic is belonging to that class.

[00:08:48] Speaker 00:
But that client could engage in multiple file transfer conversations throughout a given day.

[00:08:54] Speaker 00:
So maybe there's a file transfer in the morning.

[00:08:56] Speaker 00:
That's a conversation.

[00:08:58] Speaker 00:
There's another file transfer in the afternoon.

[00:09:00] Speaker 00:
That's a different conversation.

[00:09:02] Speaker 00:
Riddle doesn't make that distinction.

[00:09:04] Speaker 00:
It would treat them all as being the same.

[00:09:07] Speaker 00:
And so for that reason, Riddle does not recognize conversational flows.

[00:09:12] Speaker 00:
And an independent reason that Riddle does not teach conversational flows is Riddle teaches that its traffic classes are directional by nature.

[00:09:22] Speaker 00:
At appendix 1907,

[00:09:27] Speaker 00:
column 8, lines 50 through 53, Riddle says traffic classes have the property of being directional, i.e.

[00:09:36] Speaker 00:
all traffic flowing inbound will belong to different traffic classes and be managed separately from traffic flowing outbound.

[00:09:44] Speaker 00:
So Riddle says the traffic classes are directional and it confirms that

[00:09:49] Speaker 00:
At another location, column 10, starting at line 59, it says network traffic is automatically classified under existing classes, beginning with the broadest classes.

[00:10:01] Speaker 01:
Didn't the board read Riddle as encompassing both unidirectional and bidirectional traffic?

[00:10:07] Speaker 00:
It did, your honor.

[00:10:09] Speaker 01:
And what's our standard of review of the board's reading?

[00:10:12] Speaker 00:
the board's factual findings are reviewed for substantial evidence.

[00:10:17] Speaker 01:
So isn't there substantial evidence to support the board's conclusion there?

[00:10:21] Speaker 00:
No, Your Honor, because RIDDLE defines its traffic classes as being directional and the board didn't even address that definitional language.

[00:10:38] Speaker 00:
So because RIDDLE is unable to

[00:10:42] Speaker 00:
group traffic that is bidirectional, that's an independent reason that it doesn't disclose conversational flows.

[00:10:49] Speaker 00:
Because the board's construction and patent owners' proposed construction say that conversational flows are the sequence of packets in any direction.

[00:10:59] Speaker 00:
And so that means to be able to recognize the conversational flow, the packet monitor has to be able to recognize and organize traffic going in either direction.

[00:11:09] Speaker 00:
Unless your honors have any further questions, I'll reserve my time for rebuttal.

[00:11:12] Speaker 02:
We will save it for you.

[00:11:14] Speaker 02:
Thank you.

[00:11:15] Speaker 02:
Mr. Hallward-Dreammeyer, you're going to take seven minutes.

[00:11:19] Speaker 02:
Yes.

[00:11:21] Speaker 04:
Thank you, your honors.

[00:11:23] Speaker 04:
May it please the court, Doug Hallward-Dreammeyer, on behalf of Palo Alto Networks.

[00:11:27] Speaker 04:
I'm going to be addressing packet intelligence's appeal.

[00:11:31] Speaker 04:
My colleague will be addressing our cross-appeal.

[00:11:36] Speaker 04:
With respect to the claim construction,

[00:11:39] Speaker 04:
issue, as Your Honor recognized just start, the language that was supposedly omitted is exemplary only.

[00:11:47] Speaker 04:
It is not limiting.

[00:11:50] Speaker 04:
We know this in part because not only does the claim language emphasize the relationship to an activity, we also have the primary example that the patent gives and the specification gives, which is the SAP example.

[00:12:10] Speaker 04:
If that discussion of SAP stopped where PAC Intelligence's counsel stopped with referencing the recognition or identifying of two connections between the same client server as being a conversational flow, they might have a good point.

[00:12:30] Speaker 04:
But that's not where it stops.

[00:12:33] Speaker 04:
It continues on column three after discussing

[00:12:39] Speaker 04:
other examples beyond SAP that also reflect disjointed connections, disjointed being in the example of SAP that second connection, which the specification references as being independent of the first because it's a different client.

[00:12:59] Speaker 04:
And so it recognizes this as an example of a disjointed connection, discusses other examples of

[00:13:08] Speaker 04:
applications that have disjointed flows, and then comes back to SAP.

[00:13:15] Speaker 04:
And this is in column three at line 44.

[00:13:19] Speaker 04:
Considering the previous SAP example again,

[00:13:23] Speaker 04:
because one of the features of the invention is to correctly identify the second exchange as being associated with a print service on that server.

[00:13:32] Speaker 04:
Such exchange would be recognized if the clients were not the same.

[00:13:37] Speaker 04:
So on the first instance, referred to by packet intelligence, there's a conversational flow recognized where the client and server are the same.

[00:13:47] Speaker 04:
But in the latter, they are recognized as associated even though

[00:13:53] Speaker 04:
the second client is independent or disjointed.

[00:13:57] Speaker 04:
And then it continues to specify that this is, in fact, what is claimed as being inventive.

[00:14:04] Speaker 04:
What distinguishes this invention from prior art network monitors is that it has the ability to recognize disjointed flows as belonging to the same conversational flow.

[00:14:17] Speaker 04:
So it doesn't just recognize them as associated in some other ambiguous way.

[00:14:24] Speaker 04:
It recognizes them as associated in a conversational flow.

[00:14:30] Speaker 04:
So the second independent, i.e.

[00:14:32] Speaker 04:
disjointed connection between the second client and the server is part of the same conversational flow as

[00:14:43] Speaker 04:
together with the first.

[00:14:45] Speaker 04:
And that's what's claimed as being inventive.

[00:14:49] Speaker 04:
And so the board followed that construction, followed the specific definition that's given.

[00:14:56] Speaker 04:
Of course, that's also consistent with how we would generally understand the nature of the term conversation.

[00:15:03] Speaker 04:
We are engaged in a conversation, even though there are three participants, two judges and one counsel.

[00:15:12] Speaker 04:
We aren't having one conversation and we having a second conversation.

[00:15:17] Speaker 04:
It's a single conversation.

[00:15:19] Speaker 04:
And if they had wanted to somehow limit this.

[00:15:24] Speaker 04:
to distinguish it from that common understanding, they would have to have been clear.

[00:15:27] Speaker 04:
And they didn't, because they don't reference the notion of specific endpoints in the definition.

[00:15:34] Speaker 04:
Instead, again, they reference the activity.

[00:15:38] Speaker 04:
In our example, the SAP print request, in which there are two clients each making print requests.

[00:15:44] Speaker 01:
In your view, would the claim scope be any different under the patent owner's proposed construction than under the board's construction?

[00:15:52] Speaker 04:
Well, under the PAN owner's discretion, they're trying to exclude the instance in which there is the second client who joins the conversation.

[00:16:03] Speaker 04:
They're trying to exclude that because they wrongly believe that Riddle doesn't disclose it.

[00:16:10] Speaker 04:
The result is going to be the same because, as the board found, Riddle also discloses

[00:16:17] Speaker 04:
connections on multiple flows, multiple connections between a single client and a single server.

[00:16:25] Speaker 04:
So even on their definition, Riddle discloses this.

[00:16:31] Speaker 04:
And I can turn to that as well.

[00:16:34] Speaker 04:
The board found this is on A39, that Riddle's disclosure of all traffic in a particular conversation between a client and a server

[00:16:46] Speaker 04:
is classified into a single service aggregate teaches bi-directional exchange.

[00:16:54] Speaker 04:
This characterization is clearer even than the definition in reflecting the particular limitation.

[00:17:03] Speaker 04:
The board cites for that proposition Riddle at A1909, this is column 11, lines 10 to 19.

[00:17:14] Speaker 04:
A service aggregate, which is how Riddle characterizes a conversational flow, is provided for certain applications that use more than one connection in a particular conversation between a client and a server.

[00:17:30] Speaker 04:
For example, an FTP client in conversation with an FTP server employs a command channel and a transfer channel.

[00:17:43] Speaker 04:
13, lines 54 to 59, there's a further explanation of the command channel, which discloses that an FTP session has one flow that is used to exchange commands and responses and a second flow that is used to transport data files.

[00:18:00] Speaker 04:
So Riddle is explicitly disclosing between a particular client and a particular service

[00:18:08] Speaker 04:
a conversational flow that has multiple connection flows, one of which is clearly bi-directional because it's the exchange of commands.

[00:18:17] Speaker 01:
The patent owner says that there are some institution decisions, maybe in other IPRs, where the board said Riddle's classification occurs serially on individual packet-by-packet basis rather than

[00:18:31] Speaker 01:
relying on state transitions across multiple packets.

[00:18:34] Speaker 01:
Is it correct that the board said that elsewhere of Riddle?

[00:18:37] Speaker 01:
And does that impact our decision?

[00:18:39] Speaker 04:
The board in an institution decision, which of course is by definition non-final, there's no opportunity to seek review of that.

[00:18:48] Speaker 04:
And it's just their view at the particular time before having a full trial.

[00:18:53] Speaker 04:
So I would distinguish it and say it's not binding on this court or even on the board as the board panel recognized here.

[00:19:01] Speaker 04:
But further, the board distinguished those earlier non-institution decisions because they concerned a different limitation.

[00:19:11] Speaker 04:
The limitation at issue there, and this is in the 099 patent, column 35, lines 31 to 39, was to store a set of predefined state transition patterns and state operations such that traversing

[00:19:29] Speaker 04:
particular transition pattern as a result of a particular conversational flow indicates to particular conversational flow.

[00:19:39] Speaker 01:
Totally different language there.

[00:19:41] Speaker 04:
It's all about transitions between flows.

[00:19:45] Speaker 04:
and between states, which is not, of course, an issue here.

[00:19:49] Speaker 04:
And so whatever the board had said with respect to the prior art not disclosing that has nothing to do with these limitations.

[00:19:57] Speaker 02:
Council, I think I have to interrupt this conversational flow.

[00:20:03] Speaker 02:
Your time has expired.

[00:20:05] Speaker 02:
Thank you, Your Honor.

[00:20:23] Speaker 03:
Good morning, Your Honors, and may it please the Court.

[00:20:25] Speaker 03:
I'm Bill Sigler on behalf of Juniper, and I'll be arguing the cross-appeal on behalf of both petitioners here today.

[00:20:33] Speaker 03:
Substantial evidence doesn't support the board's finding that Wakeman fails to teach an associative cache.

[00:20:39] Speaker 03:
That finding is contrary to the undisputed evidence.

[00:20:43] Speaker 03:
It's undisputed that Wakeman teaches a CAM cache.

[00:20:47] Speaker 03:
It's also undisputed that associative caches were well-known and not novel.

[00:20:52] Speaker 03:
It's also undisputed between the experts that an associative cash would be selected when the need for performance outweighs costs.

[00:21:01] Speaker 03:
Patent owner's expert, Ms.

[00:21:03] Speaker 03:
Quigley, stated that at A, 1768 to 69.

[00:21:07] Speaker 03:
And our expert, Dr. Weitzman, agreed with that point at A, 1694.

[00:21:12] Speaker 02:
You didn't spend a lot of time talking about weight.

[00:21:17] Speaker 03:
Well, Your Honor, I know that was a

[00:21:19] Speaker 03:
piece of the board's decision here, and respectfully, Your Honor, the board overlooked the discussion of Wakeman in discussing the prior element of claim 1.3.

[00:21:30] Speaker 03:
There was a good deal of discussion there.

[00:21:32] Speaker 03:
The petition, for example, provided an overview of Wakeman at A782 to 85.

[00:21:40] Speaker 03:
And then detailed Wakeman's CAM cash scheme for claim 1.3 at A812 to

[00:21:49] Speaker 03:
And then our discussion with respect to the claims at issue here such as claim three on the 646 patent referred back to that earlier discussion.

[00:22:01] Speaker 03:
And of course our expert provided conclusions there based on Wakeman and that earlier discussion.

[00:22:07] Speaker 03:
And here, Your Honors, it's also important to note that the undisputed evidence shows that Wakeman itself prioritizes performance.

[00:22:15] Speaker 03:
Wakeman talks about providing faster forwarding speeds through its CAM cache.

[00:22:20] Speaker 03:
Thus, Wakeman teaches an associative cache here.

[00:22:23] Speaker 03:
That's what the undisputed evidence shows.

[00:22:26] Speaker 03:
An associative cache, as the experts agreed, is defined by an absence of a policy that restricts where data may be cached.

[00:22:35] Speaker 03:
And the experts also agreed that Wakeman doesn't make that restriction.

[00:22:39] Speaker 03:
Again, that's found in the appendix, Dr. Quigley's opinion at 1768 to 69, and Dr. Weissman as well at A1734.

[00:22:53] Speaker 01:
So even if Wakeman discloses associative cash, what about the findings on the no motivation to combine

[00:23:00] Speaker 01:
I think Wakeman talks about the, and you've already alluded to it, the trade-off perhaps between performance and expense.

[00:23:07] Speaker 01:
Isn't that a fair consideration?

[00:23:09] Speaker 01:
And isn't that part of what the board's analysis was?

[00:23:14] Speaker 03:
Well, two responses to that, Your Honor.

[00:23:16] Speaker 03:
First, the board found that there was motivation to combine Riddle and Wakeman with respect to other claims here.

[00:23:22] Speaker 03:
So that was part of the board's finding.

[00:23:26] Speaker 03:
Secondarily, the board did also find that,

[00:23:29] Speaker 03:
Wakeman doesn't teach away.

[00:23:31] Speaker 03:
The board rejected that argument in its opinion.

[00:23:34] Speaker 03:
And as we detailed in our briefing, there's multiple opinions from this court holding that

[00:23:41] Speaker 03:
cost isn't you know significant sufficient itself to outweigh the considerations of performance this is just but it's a fact it's a permissible factor and assessing motivation combined sure your honor and in here we're just talking about a simple binary design choice that one of ordinary skill in the art face is that

[00:24:00] Speaker 03:
It's the usual, the paradigm design choice in computer science, where we're trading off cost versus performance.

[00:24:08] Speaker 01:
Did the board make a factual finding that there were only those two options?

[00:24:13] Speaker 03:
The board did not, Your Honor, but the record evidence below shows that there are only those two options.

[00:24:18] Speaker 03:
And logically, there's only those two options, Your Honor.

[00:24:20] Speaker 01:
Well, I think for that, are you just relying on the expert statement that you quote in the briefs?

[00:24:25] Speaker 03:
Both expert statements, Your Honor.

[00:24:27] Speaker 03:
Both experts are expert, and in our petition,

[00:24:31] Speaker 03:
Address that an associative cache was well known and using one was a routine engineering design choice.

[00:24:36] Speaker 03:
Their expert came back in the response and raised the alternative design choice of a direct mapping cache.

[00:24:43] Speaker 01:
But I didn't see, but you can point me on the record if I missed it, where their expert said those are the only two options.

[00:24:50] Speaker 03:
She did not say that, your honor, not explicitly.

[00:24:53] Speaker 03:
But logically, that's the case here.

[00:24:55] Speaker 03:
You have two different choices here.

[00:24:57] Speaker 03:
You have a cash that limits where cash values are stored.

[00:25:00] Speaker 03:
That's direct mapping.

[00:25:02] Speaker 03:
And you have a cash that doesn't limit where cash values are stored.

[00:25:06] Speaker 03:
That's associative.

[00:25:07] Speaker 03:
Either it does or it doesn't.

[00:25:09] Speaker 03:
And there's no other evidence in the record of any other cash that would apply to this situation.

[00:25:15] Speaker 03:
We're talking about a cash mapping scheme.

[00:25:17] Speaker 03:
It either restricts or it doesn't.

[00:25:20] Speaker 03:
Here, one of ordinary skill in the art

[00:25:24] Speaker 03:
would understand from Wakeman that it doesn't restrict in its teaching and associative cash.

[00:25:28] Speaker 03:
That's what the undisputed evidence shows.

[00:25:31] Speaker 03:
And I see that I'm eating into my rebuttal time.

[00:25:32] Speaker 03:
I'm happy to answer further questions now, otherwise we'll reserve.

[00:25:36] Speaker 02:
We'll save you rebuttal time.

[00:25:38] Speaker 02:
Of course, it's only be used in response in reply on the cross field.

[00:25:45] Speaker 03:
Thank you, Ernest.

[00:25:48] Speaker 02:
Mr. Bowling.

[00:25:57] Speaker 00:
Thank you, Your Honors.

[00:25:58] Speaker 00:
Alan Boynck on behalf of Hackett Intelligence.

[00:26:02] Speaker 00:
I'll start first with the cross-affiliate issues that were raised.

[00:26:09] Speaker 00:
First of all, it's not undisputed that an associative cache was well known.

[00:26:16] Speaker 00:
There's no record evidence to support that position.

[00:26:21] Speaker 00:
In the petition,

[00:26:25] Speaker 00:
The petitioners say an associative cash was a well-known cash design paradigm, and that's at Appendix 830.

[00:26:33] Speaker 00:
To support this proposition, they cite to the 646 patent specification, which is describing an embodiment of the cash, and that description doesn't say that this is a well-known cash design at all.

[00:26:47] Speaker 00:
And the other support they point to is their expert's testimony.

[00:26:50] Speaker 00:
Their expert uses the exact same language, and associative cash was a well-known cash design paradigm.

[00:26:57] Speaker 00:
But there's no support for that statement.

[00:27:00] Speaker 00:
Their expert also cites to the 646 patent specification, which can't be used to support that, and it doesn't support that statement.

[00:27:08] Speaker 00:
Council referred to the fact that Wakeman was discussed regarding previous limitations,

[00:27:13] Speaker 00:
But if you go look at the discussion of Wakeman, their expert's testimony on that fact, and even Wakeman's teachings, it never talks about an associative cache or that an associative cache is well known.

[00:27:29] Speaker 00:
And then also regarding the there's only two design choices, that's not supported by the record either, Your Honor.

[00:27:37] Speaker 00:
The petitioner's expert said,

[00:27:41] Speaker 00:
an associative cache or any other type of well-known cache.

[00:27:46] Speaker 00:
He didn't say that there's only one other type, he said any other type, and the implication there is that there are many other types.

[00:27:53] Speaker 00:
He didn't enumerate them.

[00:27:58] Speaker 00:
At the end of the day, the board found that petitioners failed to substantiate conclusory testimony regarding whether or not an associative cache was well-known.

[00:28:11] Speaker 00:
that finding should be affirmed.

[00:28:15] Speaker 00:
Turning back to the SAP example that counsel was discussing, the board, when it was dealing with the second client getting involved, the board

[00:28:32] Speaker 00:
pointed to the language that counsel identified.

[00:28:35] Speaker 00:
It said they would even be recognized if the clients were not the same.

[00:28:39] Speaker 00:
But that statement doesn't say that they would be recognized as being in the same conversational flow.

[00:28:45] Speaker 00:
The first statement said if the clients were the same, then they would be recognized as being in conversational flow.

[00:28:52] Speaker 00:
The second statement is really speaking to the fact that the packet monitor knows that when a client communicates with this dynamic port over here,

[00:29:01] Speaker 00:
that that is a print activity.

[00:29:03] Speaker 00:
That's the recognition, the recognition of the application type.

[00:29:09] Speaker 00:
Council also analogized the conversation that we're having here to what Riddle is recognizing.

[00:29:17] Speaker 00:
But really, in that context, the way Riddle works is we are having a conversation here.

[00:29:23] Speaker 00:
And there are specific participants in that conversation.

[00:29:27] Speaker 00:
And you could say the type of conversation we're having is a Federal Circuit oral argument.

[00:29:33] Speaker 00:
While downstairs in a couple of other courtrooms, there are also some federal circuit oral arguments that are going on.

[00:29:40] Speaker 00:
Those are different conversations and they have different participants.

[00:29:44] Speaker 00:
Riddle would treat them all as the same conversation.

[00:29:48] Speaker 00:
And that's not the case.

[00:29:49] Speaker 00:
There's multiple conversations going on here.

[00:29:51] Speaker 00:
We have one going on here.

[00:29:53] Speaker 00:
There's another one downstairs.

[00:29:55] Speaker 00:
And Riddle would not be able to distinguish between those.

[00:29:59] Speaker 00:
It would say, these are Federal Circuit oral arguments.

[00:30:01] Speaker 00:
They all belong in the same bucket.

[00:30:10] Speaker 00:
So in conclusion, the board's finding on conversational flow completely ignored the conversational nature of a conversational flow.

[00:30:21] Speaker 00:
A conversation has particular participants, and the board wiped that out.

[00:30:26] Speaker 00:
First, by omitting the patentee's full definition, which highlighted the importance of those participants, and second, by applying that construction so broadly to find that riddle teaches conversational flows.

[00:30:43] Speaker 00:
But the board's finding as to riddle cannot be supported in light of its erroneous construction.

[00:30:48] Speaker 00:
Thank you, Your Honors.

[00:30:49] Speaker 02:
Thank you, Council.

[00:30:50] Speaker 02:
Mr. Stigler has a few minutes on the cross appeal.

[00:31:04] Speaker 03:
Thank you, Your Honors.

[00:31:05] Speaker 03:
I'd like to address three items in response to Pat and others' counsel.

[00:31:11] Speaker 03:
First of all, an associative cache was well known.

[00:31:13] Speaker 03:
The specifications here of the patents at issue show that.

[00:31:18] Speaker 03:
They don't explain what an associative cache is or say how to implement one.

[00:31:24] Speaker 03:
They just say it's an associative cache and the expectation is that one of ordinary skill in the art would know what one of those is and know how to implement it in this process.

[00:31:33] Speaker 03:
It's just like what we saw in this court's decisions in the C, RFD, and Uber cases.

[00:31:40] Speaker 03:
It's a well-known concept in the art.

[00:31:43] Speaker 03:
And one of ordinary skill in the art would have known how to implement it.

[00:31:46] Speaker 03:
Secondly, Your Honors, our expert Dr. Weissman did explain below that a person of ordinary skill in the art would have understood that Wakeman provides an associative cash

[00:32:01] Speaker 03:
due to not disclosing any restrictions, as I described before.

[00:32:05] Speaker 03:
You can find that at A, 1734 to 35.

[00:32:08] Speaker 03:
Third, counsel referenced a snippet from Dr. Weissman's declaration below, where he mentioned other types of caches.

[00:32:21] Speaker 03:
And if you look at where he made that statement, that's on A, 1735.

[00:32:25] Speaker 03:
What he's talking about there is just that a person of ordinary skill in the art would have understood

[00:32:30] Speaker 03:
and had the ability to implement many different types of caches.

[00:32:36] Speaker 03:
Caches were well known.

[00:32:37] Speaker 03:
He's just making the statement that, in general, caches were well known.

[00:32:40] Speaker 03:
And one of ordinary skill in the art would have had it in their skill set to implement them.

[00:32:47] Speaker 03:
With that, Your Honors, I'm happy to answer any further questions.

[00:32:50] Speaker 03:
Otherwise, I'd close by, again, stating that there's not substantial evidence to support the board's finding here and that reversal is the appropriate remedy on these two claims.

[00:33:00] Speaker 02:
Thank you, counsel, the case is submitted.