[00:00:00] Speaker 00:
Our next case is Central Petal Networks LLC versus Keysight Technologies, Inc.

[00:00:05] Speaker 00:
24-1930.

[00:00:06] Speaker 00:
Counselor, you may proceed.

[00:00:14] Speaker 02:
Good morning, Your Honors.

[00:00:16] Speaker 02:
Thank you very much, Judge Raina.

[00:00:17] Speaker 02:
May it please the court?

[00:00:19] Speaker 02:
The PTAB, in this case, made several fundamental errors.

[00:00:24] Speaker 02:
To start, it should not have combined four references that are in large part contradictory, cherry picking limitations using pure hindsight, and the patented issue here, the C572 patent as a roadmap.

[00:00:40] Speaker 02:
The PTAP's motivation to combine analysis is thoroughly lacking and entirely conclusory.

[00:00:46] Speaker 02:
The vast majority of it, or all of it actually, is at the appendix 22, one page at most, and the top of page 23.

[00:00:55] Speaker 02:
In fact, it is not even an analysis.

[00:01:01] Speaker 02:
It's not even a conclusory analysis.

[00:01:03] Speaker 04:
It is merely- I'm sorry.

[00:01:05] Speaker 04:
There was a prior IPR for a related patent, the 148 patent.

[00:01:13] Speaker 04:
And the board in that prior IPR

[00:01:17] Speaker 04:
found a motivation to combine these references.

[00:01:20] Speaker 04:
Is that right?

[00:01:21] Speaker 04:
It did.

[00:01:21] Speaker 04:
And we affirm that, right?

[00:01:23] Speaker 04:
Using rule 36, you did.

[00:01:25] Speaker 04:
OK.

[00:01:25] Speaker 04:
So I guess what I'm trying to understand is maybe that's already been a litigated issue, whether or not these references here, which are the same references there, as to whether there's a motivation to combine them.

[00:01:43] Speaker 02:
Well, I don't think so, Your Honor.

[00:01:44] Speaker 02:
And at least for three reasons, probably four.

[00:01:49] Speaker 02:
First of all, the claims themselves between the two patents are facially different.

[00:01:55] Speaker 02:
And there are meaningful limitations in the 572 patent, the patent at issue here, and the prior patent that go to the timing and directly to the court's, to the board's incorrect claim construction in this case that I'll get to in a minute.

[00:02:11] Speaker 02:
So they're facially different.

[00:02:12] Speaker 02:
Second, the standard of review.

[00:02:17] Speaker 02:
The standard of claim construction was different at that time.

[00:02:20] Speaker 02:
They used the broadest reasonable interpretation.

[00:02:24] Speaker 02:
And as Your Honor may recall, that was changed.

[00:02:27] Speaker 02:
And in this case, it's the regular Phillips standard.

[00:02:31] Speaker 02:
So that's very different.

[00:02:32] Speaker 02:
And in any event, this was a Rule 36 decision.

[00:02:37] Speaker 02:
We don't know on what basis, actually, the court decided on.

[00:02:42] Speaker 04:
I mean, hopefully you can understand my concern as to this one particular issue, which is the fact question of whether there was a motivation to combine references A, B, C, and D. Setting aside what the claims look like, setting aside how to construe those claims, this is more a separate line of inquiry, which is a fact matter.

[00:03:12] Speaker 04:
would a skilled artisan be motivated to combine these references?

[00:03:16] Speaker 04:
That seems like something that was in fact debated and needed to be resolved in the prior litigation and was resolved against your client.

[00:03:26] Speaker 02:
Again, a motivation to combine.

[00:03:29] Speaker 02:
First of all, it was decided at the board.

[00:03:31] Speaker 02:
This court did not rule directly on this issue.

[00:03:34] Speaker 02:
We don't know because of the Rule 36.

[00:03:36] Speaker 02:
But more importantly,

[00:03:37] Speaker 02:
in order to decide whether there is a motivation to combine references that meet those claims, not our claims here, those claims.

[00:03:47] Speaker 02:
A, we need to know what those claims are, and B, what the claim construction and what the standard for claim construction is of those claims.

[00:03:56] Speaker 02:
Just because that was litigated then,

[00:03:59] Speaker 02:
doesn't mean that our claims here, with their own words, with their own claim construction, would get the same analysis.

[00:04:08] Speaker 02:
And in any event, we're still entitled to a ruling from this court on that question.

[00:04:14] Speaker 04:
They look very, very similar to me with the claims from 148, at least the representative claims.

[00:04:22] Speaker 04:
It's the same idea of having two different rule sets and then applying the first rule set, and then in response to a need to swap the first for the second rule set, cease any processing, and then

[00:04:43] Speaker 04:
cash packets and then reconfigure the system so that you can use the second rule set.

[00:04:52] Speaker 04:
I mean, it looks really similar to me.

[00:04:54] Speaker 04:
I didn't quite see a difference.

[00:04:56] Speaker 04:
I'm not saying that there should be collateral estoppel, but I'm just saying at a minimum it would be a little peculiar, I think, for a full proceeding to conclude that the 148 claim is

[00:05:11] Speaker 04:
is not patentable, and then somehow this claim has got a game changer in it that suggests that it is patentable.

[00:05:19] Speaker 02:
First of all, the combination is wrong and lacks analysis.

[00:05:22] Speaker 02:
And just because there was a conclusion in the first case doesn't mean that the analysis was done here.

[00:05:29] Speaker 02:
They didn't do it.

[00:05:30] Speaker 02:
If we look at it at page 22, they just didn't do it.

[00:05:34] Speaker 02:
They said, by the way, the board, that there is no claim preclusion.

[00:05:38] Speaker 02:
They agreed with us.

[00:05:39] Speaker 02:
And they went through to do the analysis anyway, but they didn't do a motivation to combine analysis at all.

[00:05:46] Speaker 02:
Now, if you look at the claims here, they require clearly a temporal limitation.

[00:05:54] Speaker 02:
The steps are done in sequence here.

[00:05:59] Speaker 02:
And the claims are definitely different.

[00:06:01] Speaker 02:
So first of all, in the 148 patent, it requires pre-processing.

[00:06:06] Speaker 02:
It's at the very top of the claim.

[00:06:08] Speaker 02:
Here it talks about modifying and then configuring.

[00:06:12] Speaker 02:
In the 148 patent, it talks about signaling and being responsive to the signaling and ceasing caching of one or more packets.

[00:06:21] Speaker 02:
In our case here, it talks about which packets.

[00:06:26] Speaker 02:
It talks about based on a signal to process,

[00:06:32] Speaker 02:
It talks about processing the packets, caching the packets that were ceased to be processed, and then a timing limitation after completion of the configuring, which is missing from the prior patent.

[00:06:49] Speaker 02:
after completion of the reconfiguring, start processing those cash patents.

[00:06:54] Speaker 02:
So there are definitely differences.

[00:06:56] Speaker 02:
But the main point here is that the analysis wasn't done and this court didn't address it.

[00:07:02] Speaker 02:
And I think we're entitled to a review by this court.

[00:07:06] Speaker 02:
And frankly, much more on this issue should be required of the PTAB.

[00:07:12] Speaker 02:
When the PTAB invalidates and combines four references,

[00:07:17] Speaker 02:
and says that there is a motivation to combine.

[00:07:20] Speaker 02:
They need to tell us why.

[00:07:22] Speaker 02:
And they didn't.

[00:07:24] Speaker 02:
They didn't even conclude that the one of skill and the art should combine them.

[00:07:29] Speaker 02:
They just simply said once combined, the limitations are met.

[00:07:35] Speaker 02:
And by the way, so that's the first fundamental error.

[00:07:39] Speaker 02:
And the second fundamental error here is that even if combined, key limitations are missing.

[00:07:48] Speaker 02:
So specifically, even if combined, none of the four references disclose pausing and caching data packets during the rule swap.

[00:08:01] Speaker 01:
Well, Huema does disclose pausing.

[00:08:04] Speaker 02:
Huema discloses pausing in that particular application, but not caching.

[00:08:13] Speaker 01:
In 2019, these weren't particularly new concepts.

[00:08:17] Speaker 01:
No, no, that's each one individually.

[00:08:19] Speaker 01:
Each one individually.

[00:08:21] Speaker 01:
And caching was pretty common.

[00:08:23] Speaker 02:
Caching was common, of course, going back a long time, but not in this case.

[00:08:28] Speaker 02:
So when we're talking about the swapping rules dynamically and on the fly, the state of the art, and by the way, it's 2013 at the time of the application.

[00:08:39] Speaker 02:
The application thing.

[00:08:42] Speaker 02:
Yeah.

[00:08:44] Speaker 02:
So when you're talking about dynamically swapping rules on the fly, quickly in a system that receives rules and modifies them and then configures, the prior art simply didn't do it.

[00:09:02] Speaker 02:
The one reference they have is the Reese reference, which didn't

[00:09:08] Speaker 02:
didn't pause at all.

[00:09:10] Speaker 02:
So the packets just came.

[00:09:13] Speaker 02:
Now they're saying, let's combine that with Huima.

[00:09:16] Speaker 02:
No explanation as to why we're combining with Huima, because by the way, Reese doesn't need to pause because of its own system.

[00:09:23] Speaker 02:
But okay, so they say, let's combine with Huima.

[00:09:26] Speaker 02:
Huima

[00:09:27] Speaker 02:
doesn't cash and it just lets the packets go through.

[00:09:31] Speaker 02:
That threatens the system with those packets in the meantime or the packets get lost.

[00:09:37] Speaker 02:
And by the way, if Huima was interested in saving packets or avoiding the threat to the network, it actually proposes an alternative embodiment.

[00:09:50] Speaker 02:
Not caching, it proposes in that case to use

[00:09:56] Speaker 02:
parallel paging system.

[00:10:01] Speaker 02:
So they didn't do it either.

[00:10:05] Speaker 02:
Then you have to ask yourself, they're saying, okay, so let's now combine with a caching reference.

[00:10:11] Speaker 02:
Why?

[00:10:12] Speaker 02:
Again, we don't know.

[00:10:13] Speaker 02:
But then which caching reference?

[00:10:15] Speaker 02:
What they're saying is, let's combine with a hater reference.

[00:10:19] Speaker 02:
The problem is that hater and most caching references in this type of situation talk about caching during processing.

[00:10:28] Speaker 02:
The whole point of hater and other caching references for processing data packets is to speed up

[00:10:36] Speaker 02:
the processing of data to make it more efficient.

[00:10:40] Speaker 02:
Hayter talks about caching packets during active processing and not caching the whole packet because they need to go as fast as possible.

[00:10:53] Speaker 02:
Decidedly not during a pause for a rule swap.

[00:10:58] Speaker 02:
And by the way, they're also combining with a fourth reference.

[00:11:03] Speaker 02:
Okay.

[00:11:04] Speaker 02:
To talk about modifying and such.

[00:11:06] Speaker 00:
So, but then wasn't there also finding that Hader found that a packet having no payload data would be cash to me entirely.

[00:11:17] Speaker 00:
Isn't that speeding up the process?

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

[00:11:20] Speaker 02:
That's evidence of speeding up the process.

[00:11:22] Speaker 02:
Hader wants to speed up the process and cache his packets, parts of packets actually.

[00:11:28] Speaker 02:
It can cache entire packets too.

[00:11:30] Speaker 02:
No.

[00:11:31] Speaker 02:
It only can cache an entire packet, perhaps, we don't know for sure, but at best it can cache an entire packet if only that packet, if only if that packet has a header, no data.

[00:11:43] Speaker 02:
If it has data,

[00:11:46] Speaker 02:
The hater actually says to inhibit caching the whole packet.

[00:11:51] Speaker 01:
But most importantly... The board found that in some instances there was a header and at least one byte of payload that was cached.

[00:11:59] Speaker 02:
That's right Judge Freeman.

[00:12:00] Speaker 02:
So at least, so the hater does say that you can, if you have a header and some data you can cache at least one byte of the data, but clearly not the whole data payload of that packet.

[00:12:14] Speaker 02:
But the most important point here is that it's all about speeding up the process, not caching during a pause.

[00:12:25] Speaker 02:
I see the time in my rebuttal time.

[00:12:26] Speaker 04:
I just wanted to explore a couple things about Hader with you, or the invention.

[00:12:32] Speaker 02:
Here, is it the Reese or Rose opinion, a reference?

[00:12:37] Speaker 02:
I've been saying Roeasy, but then my client corrected me and said Reese.

[00:12:42] Speaker 02:
Let's just say ROEC.

[00:12:44] Speaker 04:
So we've got ROEC, and it talks about the concerns about latency and wanting to speed up processing.

[00:12:53] Speaker 04:
And caching is something that really helpfully contributes to high speed data packet processing.

[00:13:03] Speaker 04:
So I mean, there is grounds to use it in a situation where, assuming you want to pause,

[00:13:11] Speaker 04:
virus review of packets while moving from one rule set to the second rule set.

[00:13:20] Speaker 04:
Hader talks about caching portions of data packets, as you know, but it talks about

[00:13:30] Speaker 04:
that in terms of whatever is the necessary portion of the packet that is going to be operated on by the processor.

[00:13:39] Speaker 04:
So which parts do you need to have quick access to for doing processing on?

[00:13:48] Speaker 04:
And in the context of this combination, where you're swapping out rule sets

[00:13:54] Speaker 04:
and you're going to pause any processing of the packets, why wouldn't it be that you would be what's necessary to be processed are the entire packets before you let them go through the firewall?

[00:14:11] Speaker 04:
It seems like that's the most obvious outcome of the situation.

[00:14:15] Speaker 04:
And if Hayter says process whatever is necessary,

[00:14:19] Speaker 04:
Here in this combination, what is necessary would be the entire package.

[00:14:23] Speaker 02:
It's a completely different application, Judge Chen.

[00:14:26] Speaker 02:
In Hader, he wants to keep processing.

[00:14:29] Speaker 02:
He doesn't want to pause.

[00:14:30] Speaker 02:
He wants to keep processing and process as fast as possible.

[00:14:34] Speaker 02:
And for that, you cache only a portion of the data and the rest you put in memory and you let it go through.

[00:14:41] Speaker 02:
If you do that in our application,

[00:14:44] Speaker 02:
in a network security situation, you're defeating the whole purpose.

[00:14:47] Speaker 02:
Because now you're letting through the data that is potentially infected.

[00:14:53] Speaker 02:
HATER precisely says you should inhibit processing the whole data packet.

[00:14:59] Speaker 02:
Only what's necessary for fast processing.

[00:15:01] Speaker 01:
I don't think it qualifies for teaching away though.

[00:15:04] Speaker 01:
And that's really what you're arguing.

[00:15:06] Speaker 02:
Okay, I think it does because

[00:15:10] Speaker 02:
The whole point is not to cash the whole thing and to just speed up processing.

[00:15:13] Speaker 02:
But here's the most important part.

[00:15:16] Speaker 02:
Hater is about keeping processing.

[00:15:20] Speaker 02:
It has says absolutely zero about what to do when you're pausing and, and, and.

[00:15:26] Speaker 01:
This isn't an anticipation argument.

[00:15:28] Speaker 01:
It's a, it's an obviousness.

[00:15:30] Speaker 01:
So we're looking for just the motivation to combine.

[00:15:33] Speaker 01:
Not that hater satisfies all of the limitations.

[00:15:38] Speaker 02:
First of all, you need to combine four references, not just hater, but okay, but on hater, the fact that he says nothing about pausing and cashing during a pause, I think it takes it completely out of the realm.

[00:15:54] Speaker 02:
And I want to leave you, now I'm into my red time, I want to leave you two fundamental points.

[00:16:01] Speaker 02:
First, even if your honors think that there might be a motivation to combine, I really don't think it is,

[00:16:08] Speaker 02:
The record below does not conclude that.

[00:16:13] Speaker 02:
It does not give a proper analysis for that.

[00:16:16] Speaker 02:
The board needs to be held to a higher standard.

[00:16:20] Speaker 02:
It cannot just simply take the bold assertions by the petitioner and

[00:16:27] Speaker 02:
just conclude with no explanation.

[00:16:30] Speaker 02:
Second, it categorically got the claim construction wrong.

[00:16:34] Speaker 02:
And if you find that that's correct, that in fact they did get the claim construction wrong, you must reverse.

[00:16:41] Speaker 02:
The claim construction here, they decide that the claim does not have a temporal limitation and that the cashing can take place at any time, which is patently, so to speak, incorrect.

[00:16:56] Speaker 02:
on the face of the claim and with the specification.

[00:17:00] Speaker 02:
The whole point of these claims is to pause and cash during the rule swap, not at any time.

[00:17:07] Speaker 02:
They got that wrong, and at least for that, it needs to be reversed.

[00:17:12] Speaker 00:
So you did go well into your time, but we'll restore a little bit of time for you.

[00:17:20] Speaker 00:
Thank you, Your Honor.

[00:17:21] Speaker 00:
Thank you, Counselor.

[00:17:23] Speaker 00:
Let's hear from

[00:17:27] Speaker 00:
Dr. Dietrich.

[00:17:29] Speaker 03:
Thank you, Your Honor.

[00:17:34] Speaker 03:
Good morning, Your Honors.

[00:17:35] Speaker 03:
May it please the court, Jonathan Dietrich, for Keysight Technologies.

[00:17:39] Speaker 03:
This appeal presents a straightforward case for affirmance based on the substantial evidence supporting the board's factual determinations that underpin its conclusion of obviousness.

[00:17:51] Speaker 03:
I will start with centripetal's timing argument

[00:17:54] Speaker 03:
that the prior art does not render obvious cashing packets while rule sets are being swapped.

[00:18:00] Speaker 03:
First, the board was correct to reject centripetal's claim construction argument to limit cashing to occur only while rule sets are being swapped.

[00:18:10] Speaker 03:
But more importantly, contrary to what counsel just said, consideration of centripetal's proposed claim construction does not change the analysis here in any respect because, as centripetal admitted in its brief,

[00:18:23] Speaker 03:
The board found that the prior art vendors obvious the claims even under centripetal asserted claims go.

[00:18:30] Speaker 04:
So just so I understand the board's primary theory of what would be the result of the combination of all these references.

[00:18:40] Speaker 04:
Would the caching occur in the board's view completely before the reconfiguring, or would it be occurring maybe a little bit before the reconfiguring starts, but then continue to

[00:18:58] Speaker 04:
be caching during the reconfiguring?

[00:19:00] Speaker 04:
It wasn't clear to me what the difference was between what was in the board's mind as to the resulting combination versus what centripetal wanted, which appeared to be that the caching would occur only during the reconfiguration process from rule set one to rule set two.

[00:19:20] Speaker 03:
Yes, your honor.

[00:19:21] Speaker 03:
I agree that I believe the record is a little bit unclear, especially compared between what Central has argued in this appeal and what it argued before the board.

[00:19:29] Speaker 04:
Well, I need to understand what the board's thinking to figure out first of all, whether it was reasonable.

[00:19:34] Speaker 03:
So what was the board thinking?

[00:19:36] Speaker 03:
I think the board was thinking that just based on the layout of the claim, that the claim and the specification did not support limiting the timing

[00:19:48] Speaker 03:
to occur only during the reconfiguration step.

[00:19:51] Speaker 03:
For example, as the board said, it could occur after the signal to switch is received.

[00:19:57] Speaker 04:
Well, you were the petitioner, right?

[00:20:00] Speaker 04:
So you proposed a certain outcome from combining all these references.

[00:20:06] Speaker 04:
And so your proposed outcome was what?

[00:20:10] Speaker 04:
The cashing began before any reconfiguring started?

[00:20:14] Speaker 04:
No, Your Honor.

[00:20:16] Speaker 03:
To be clear, the board analyzed and the petitioner's analysis demonstrated exactly the claim scope that centrifugal is arguing for.

[00:20:26] Speaker 03:
to explain why the caching occurred while rulesets are being swabbed.

[00:20:31] Speaker 03:
In other words, between the timing of when the signal to switch rulesets is received and ending with when the signal is received to resume processing.

[00:20:45] Speaker 03:
why it doesn't matter, and as Drupal admitted in its opening brief, that's the analysis the board performed, and that's the analysis we performed.

[00:20:53] Speaker 03:
So it's unclear why they think that this claim construction matters when that was the analysis that actually was performed.

[00:21:01] Speaker 03:
So with that, getting into that analysis.

[00:21:05] Speaker 00:
Give us a quick summary of the motivation to combine analysis of Hayler, Roessi, and Huma.

[00:21:13] Speaker 03:
Yes.

[00:21:15] Speaker 03:
So I'll address those in two respects.

[00:21:17] Speaker 03:
I'll say Rose for shorthand, your honor.

[00:21:20] Speaker 03:
Rose plus Huma.

[00:21:22] Speaker 03:
And then I'll address Rose and Hader.

[00:21:27] Speaker 03:
So regarding Rose and Huma, as the board noted, Rose's paragraphs 47 and 52.

[00:21:33] Speaker 04:
Where are you in the board opinion?

[00:21:36] Speaker 03:
Appendix 22.

[00:21:38] Speaker 03:
Rose's paragraph

[00:21:40] Speaker 03:
47 and 52 teach a response system that initiates the process of swapping rule sets upon receiving a signal.

[00:21:49] Speaker 03:
And then looking at Huma, paragraphs 31 and 36, still on appendix 22, Huma teaches that in response to receiving a signal to change rule set, packet processing pauses to avoid processing of packets with outdated rules.

[00:22:05] Speaker 03:
And that after rules are updated, a signal is received

[00:22:09] Speaker 03:
for packet processing to continue.

[00:22:13] Speaker 03:
So still, Appendix 22, the board credited Dr. Jacobson's testimony based on these teachings that a person of ordinary skill in the art would have understood that Rose's network device should pause the processing of any packets that arrive after the instruction to change rule set and not resume processing until the rule swap is completed.

[00:22:36] Speaker 03:
Now addressing the combination of Rose and Hader,

[00:22:39] Speaker 03:
Rose's paragraph 13 discloses a rapid response system that has the same objective as what is disclosed in the 572 patent, namely swapping rule sets, quote, as quickly as possible, closed quote, to minimize harm from a network attack.

[00:23:00] Speaker 03:
Similarly, Rose's paragraph 18 discloses the goal of reducing latency during a potential or ongoing attack.

[00:23:07] Speaker 03:
That's discussed.

[00:23:09] Speaker 03:
at appendix 19.

[00:23:12] Speaker 03:
And these disclosures tie in with Hayter's disclosure at column 3, lines 53 through 59, of using a packet processor with an integrated cache to allow for high-performance packet processing when receiving, processing, and forwarding packets.

[00:23:29] Speaker 03:
That's discussed at appendix 22.

[00:23:32] Speaker 03:
And Dr. Jacobson explains that applying haters teachings of high performance packet processing to Rose would further Rose's goal of decreased latency and applying the rule sets to packets when switching rule sets.

[00:23:45] Speaker 03:
That's Appendix 270 paragraph 102, which is cited in the petition at Appendix 123.

[00:23:51] Speaker 03:
And similarly, at appendix 21, the board acknowledged Dr. Jacobson's testimony that it would have been obvious to a person of ordinary skill and the art to cache packets as hater teaches when pausing the processing of packets per rose and huma so that packets can be efficiently processed once packet processing resumes with the second rule set.

[00:24:13] Speaker 03:
And again, that furthers Rose's goal of improved efficiency and reduced latency during a time of a network attack.

[00:24:21] Speaker 03:
So with this record, the board's determination that the combination of Rose, Huma, and Hader teaches or suggests that any caching of packets would occur in the period between when a signal is received to change rule sets and when processing of packets reoccurs under the new rule set is grounded in the express teachings of the references and Dr. Jacobson's supporting testimony.

[00:24:44] Speaker 03:
And I'd also just add there was some discussion on this.

[00:24:49] Speaker 03:
with your honors that I think there's a pretty clear record for collateral stopple and the board found as much in its institution decision saying that centripetal likely would be a stop from arguing against motivation to combine if it were to press that issue.

[00:25:04] Speaker 03:
Ultimately, it decided based on the D2s in the petition and fact finding.

[00:25:10] Speaker 03:
And those fact finding are the same as it made before with respect to the 148 and supported by substantial evidence.

[00:25:17] Speaker 03:
But the relevant claim limitations here are identical.

[00:25:21] Speaker 03:
And the board even noted in its final written decision, I believe, appendix 9, or sorry, appendix 11 and 12,

[00:25:34] Speaker 03:
by the authoring APJ, who was the same APJ who authored the decision on the 148 patent, confirmed that the petitioner was correct, that the board, quote, previously determined that the combination of Rose, Huma, and Hader renders obvious cashing entire packets while the rule sets are being swapped, close quote.

[00:25:57] Speaker 03:
So with respect to this limitation disputed, the motivation to combine issue was identical, and centripetal should be collaterally stopped.

[00:26:06] Speaker 03:
I would also respectfully submit they should be collaterally stopped on the core invalidity issue of whether those references teach that limitation, given the board's finding that it previously determined the exact same issue.

[00:26:19] Speaker 03:
I know there's cases from this court saying that that shouldn't be the result if the prior decision

[00:26:26] Speaker 03:
was decided under broadest reasonable interpretation.

[00:26:29] Speaker 03:
But those cases don't involve the fact here where the panel of the second opinion explicitly notes that they decided the exact same issue irrespective of the difference in claim construction standards.

[00:26:45] Speaker 03:
With my remaining time, if your honors would like, I'll briefly address the cashing entire packets argument.

[00:26:55] Speaker 03:
The board based its determination that HATER renders obvious cashing entire packets on two independent grounds, both of which are grounded in substantial evidence and either of which is sufficient to support affirmance.

[00:27:07] Speaker 03:
That's Appendix 23 through 26.

[00:27:12] Speaker 03:
And I'll start with the board's determination on the obviousness of cashing entire packets that include a data payload.

[00:27:19] Speaker 03:
The board noted HATER's relevant disclosures at Appendix 21, 23, and 24.

[00:27:25] Speaker 03:
These include hater's disclosure at appendix 430, column 6, that, quote, it may be desirable for the header of a packet to be stored in the L2 cache, closed quote, and that, quote, it may be desirable to store a portion of the data pack payload of a packet in the L2 cache as well, closed quote.

[00:27:43] Speaker 03:
And the board also relied on Hader's disclosure at appendix 439, columns 24, lines 34 through 36, that quote, the entirety of the header and at least one byte of data payload of the packet stored in the cache, end quote.

[00:27:59] Speaker 03:
And furthermore, the board relied on the testimony of Keysight's expert, Dr. Jacobson, who explained, based on these disclosures, the hater places no upper limit on the number of bytes of data payload that may be stored in the cache.

[00:28:11] Speaker 03:
And that would have been obvious to person of skill in the art to store in the cache entire headers and at least one byte of data payload and up to the entire data payload.

[00:28:21] Speaker 03:
That's Appendix 303 to 305 and Appendix 1505.

[00:28:29] Speaker 03:
And just quickly moving on to the independent grade basis for packets that do not include a data payload.

[00:28:40] Speaker 03:
The board addresses appendix 25 and 26, again reviewing hater's relevant disclosures and finding that hater explicitly disclosed caching the entirety of such packets.

[00:28:53] Speaker 03:
For example, the board reviewed hater's disclosure at column 11,

[00:28:57] Speaker 03:
through 3 and found that there are packets such as TCP acknowledgment packets that include only a header and lack of data payload.

[00:29:05] Speaker 03:
And that's consistent with hater's disclosure at column 11, lines 14 and 15 that, quote, figure 3 illustrates all of the header and the L2 cache, as well as the disclosure at column 24, lines 34 through 36 that the entirety of the header is stored in the cache.

[00:29:26] Speaker 03:
Based on these disclosures for packets that comprise only a header and lack a data payload, the board found that HATER clearly teaches caching the entirety of those packets.

[00:29:35] Speaker 03:
That's Appendix 26.

[00:29:36] Speaker 03:
I'd just like to note, as a final point, Your Honors, that before the board's intrepidal did not dispute HATER's disclosure of storing the entirety of packets that lacked

[00:29:46] Speaker 03:
a data payload in a cache instead regarding those packets, centripetal only argued that, quote, the claims require caching full versions of packets, i.e.

[00:29:56] Speaker 03:
the header and the payload data, close quote.

[00:29:59] Speaker 03:
That's found at appendix 1269 and also discussed at appendix 1488.

[00:30:05] Speaker 03:
The board addressed this argument that was actually presented to it

[00:30:10] Speaker 03:
at appendix 26, correctly rejecting it because nothing in the claim language or the specification advances in intent to exclude packets that lack a data payload from the meaning of packet.

[00:30:21] Speaker 03:
Here, some triples presented a number of new arguments on this issue in appeal, but they were not presented before the board in their patent owner response and therefore should be forfeited.

[00:30:30] Speaker 03:
But they're also meritless.

[00:30:32] Speaker 03:
Unless Your Honor would like to discuss that further or any other questions, I will see the remainder of my time.

[00:30:39] Speaker 03:
We thank you for your argument.

[00:30:43] Speaker 02:
Thank you.

[00:30:46] Speaker 02:
Two main points, Your Honors.

[00:30:49] Speaker 02:
First of all, as I said earlier, and as counsel went through during his argument, the entirety of the motivation to combine discussion from the board's point of view is in that one paragraph at page 22.

[00:31:06] Speaker 02:
If that becomes the standard for analysis under motivation to combine, the PTAB and any accused infringer can always invalidate the patent.

[00:31:17] Speaker 02:
But that is not this court's precedent.

[00:31:21] Speaker 02:
This court has been clear.

[00:31:22] Speaker 02:
I'll give you an example.

[00:31:24] Speaker 02:
The case in Renuvasive, 842, F3rd, 1376 at 1381.

[00:31:32] Speaker 02:
through 1383, a couple of key.

[00:31:34] Speaker 00:
Well, the final written decision may lack in some clarity and could have been written perhaps with more detail.

[00:31:44] Speaker 00:
Isn't it the case that it contains enough to show us the path to where the board's reasoning was headed and to finding that their reasoning was reasonable, supported by substantial evidence?

[00:31:56] Speaker 02:
I don't think so, Judge Reyna, because they don't grapple with the key fundamental questions of how these four disparate references can be combined.

[00:32:05] Speaker 02:
And this court made this clear, as I said, in the new invasive case.

[00:32:08] Speaker 02:
I'll give you a couple of quotes from that case.

[00:32:11] Speaker 02:
Quote, the factual inquiry, I'm sorry, although we review the factual finding for substantial evidence, the factual inquiry whether to combine references must be thorough and searching.

[00:32:23] Speaker 02:
And the need for specificity pervades our authority on the PTAP's findings on motivation to combine.

[00:32:30] Speaker 02:
And also,

[00:32:31] Speaker 02:
We have however identified some insufficient articulations of motivation combined.

[00:32:36] Speaker 02:
First, conclusory statements alone are insufficient and instead the finding must be supported by a reasoned explanation.

[00:32:44] Speaker 02:
Second, it is not adequate to summarize and reject arguments without explaining why the PTAB accepts the prevailing argument and so on.

[00:32:55] Speaker 02:
The last point on claim construction.

[00:32:58] Speaker 02:
The claim construction here is categorically wrong and it matters because if you reverse the claim construction, it becomes clear that no reference does caching and pausing during the rule swap.

[00:33:12] Speaker 02:
And here's the board's finding on this at appendix 8 through 9, but I'm going to read from the top of page 9.

[00:33:29] Speaker 02:
where it says, we are not persuaded that specifying when the caching occurs, i.e.

[00:33:37] Speaker 02:
while the rule sets are being swapped, is integral to the plain and ordinary meaning of caching the one on more packets.

[00:33:44] Speaker 02:
And a little bit above that says that it can be done quote, even if the swapping of rule sets has not yet occurred.

[00:33:53] Speaker 02:
That's absurd.

[00:33:54] Speaker 02:
The whole point of the patent is to pause in cash during the rule swap.

[00:33:59] Speaker 02:
The claim construction is clearly wrong.

[00:34:01] Speaker 02:
And there is timing limitations and timing sequencing words throughout the specification under this court's precedent that

[00:34:15] Speaker 02:
claim construction needs to be applied to these claims and as a result you should reverse.

[00:34:23] Speaker 02:
Thank you very much Your Honor.

[00:34:24] Speaker 00:
We thank the parties for their arguments this morning.

[00:34:27] Speaker 00:
This case will now be taken under submission.