[00:00:00] Speaker 03:
Our next case is Centripetal Networks v. Cisco Systems 2024-2097. Mr. Rowan.

[00:00:11] Speaker 01:
Thank you, Your Honors, and may it please the Court. Matthew Rowan on behalf of Centripetal. I'm going to endeavor to reserve two minutes for rebuttal. This appeal involves three network security patents and two competing suites of technology. The asserted claims and the relevant devices are highly technical and complex. but the errors the district court made below are not. Instead, the district court made three straightforward errors of law, one for each patent. I'd like to start with the 193 patent, which is the filtering patent, then turn to the 806 patent, which is the rule-swapping patent, and then end briefly on the 176 patent, which is the correlation patent.

[00:00:50] Speaker 03:
What are you starting with?

[00:00:51] Speaker 01:
With the 193 patent. So first, on the 193 patent...

[00:00:57] Speaker 01:
The district court erred as a matter of law by misconstruing a key term in the asserted claims. So claims 18 and 19 of the 193 patent describe a system for preventing exfiltration by using rules to filter, quote, particular types of data transfers from one network to another, blocking some traffic while allowing other traffic to reach its intended destination. Now, that key term, the particular type of data transfer, is what Judge Haynes misconstrued.

[00:01:28] Speaker 01:
And the plain text of that key term makes clear that a particular type of data transfer refers generally to the subset of data transfers from a first network to a second network that meet criteria specified by one or more packet filtering rules.

[00:01:45] Speaker 03:
Well, so the question here is imprinted.

[00:01:49] Speaker 03:
And the district court found with respect to the 193 pattern that the evidence is overwhelmingly that particular function that we're talking about is not present in the accused technology. Overwhelming. And the infringement of the apply deference. Fairly erroneous.

[00:02:15] Speaker 01:
So two responses to that. So in her first opinion, that is correct, what Judge Haynes said. But she, in full candor, corrected her opinion on reconsideration. So her first opinion was based on an initial claim construction error, where she initially concluded that she construed the term network device in claims 18 and 19 to refer only to a single endpoint computer device. And then based on that error, which she corrected on reconsideration, said, well, the evidence is overwhelming that it's not the case that a particular type of data transfer is what the accused switches and routers prevent because what she said was all they prevent is all data transfers from one endpoint computer to another.

[00:03:03] Speaker 01:
And that's not a particular type. We pointed out in our Rule 52 motion that that misunderstands what the claims say.

[00:03:11] Speaker 02:
Okay, but by the time of the reconsideration decision, why didn't she get it right at that point? Because it seems to me that if you look at the specification and the prosecution history, which you don't discuss either in your opening or reply brief, that it's correct that there's two-stage filtering here. and that the Cisco filtering is just origin and destination filtering at the first stage.

[00:03:44] Speaker 02:
And there's substantial evidence to support that conclusion.

[00:03:49] Speaker 01:
Why is that wrong? So two responses to that, Your Honor. So first, I don't read Judge Haynes' opinion, either her initial opinion or her reconsideration opinion, to accept that. the two-stage framing that Cisco puts forward? Now, of course, the specification does say.

[00:04:05] Speaker 02:
Let's assume for the moment that in the reconsideration decision, she was accepting the two-stage thing. And the first stage is origin and destination filtering, and that that is all the Cisco device does with the quarantining. Why is that wrong?

[00:04:23] Speaker 01:
Well, so I think that's wrong for a couple reasons. I mean, first, I do think it still depends on misunderstanding what a particular type of data transfer means. Because what essentially she's saying is that a particular type of data transfer has to be a particular method associated with data transmission. And what she says is that the accused technology doesn't do that. Because what she says is applying an SGT tag to a packet is not equivalent to something like HTTP GET or HTTP POST that commands the device to do a certain thing.

[00:04:55] Speaker 01:
Okay, but wait, wait.

[00:04:56] Speaker 02:
At the first stage, there's this five-tuple filtering, right? And do you agree that the filtering that Cisco does, was it in the category of the five-tuple filtering?

[00:05:12] Speaker 01:
So the way that the SDT tag gets applied is, so it's based on the criteria in the five tuple, but that's not inconsistent with the asserted claims because what the asserted claims make clear, I think this goes exactly to the legal error that Judge Haynes made.

[00:05:30] Speaker 01:
If you read the specification in full, it makes clear that there are of course, can be embodiments where the particular type of data transfer is defined by the particular method associated with transmission.

[00:05:43] Speaker 02:
I'm sorry, I don't understand what you're saying. What I'm saying is there's a first stage with the five-tuple criteria, and that the Cisco filtering... falls within that five-tuple thing because it's origin and destination filtering. So it falls within the first stage and not the second stage.

[00:06:04] Speaker 01:
Why is that wrong? So I guess what I'm saying is that those aren't the correct two stages that the patent describes. So the first stage is the determination that there's a match. And the second stage is the application of an operator.

[00:06:20] Speaker 02:
That seems to be the heart of the problem because as I read the specification and the prosecution history, the first stage is five-tuple filter.

[00:06:28] Speaker 01:
So I respectfully disagree with that. Okay, I understand. And I think the best place to look for why that's inconsistent with, that isn't the position that we took in the prosecution history is page 1904, which is page four of Centripital's IPR opposition brief.

[00:06:46] Speaker 01:
where it says that the 193 quote discloses exemplary embodiments for preventing exfiltrations when packets match, one, a particular source and destination criteria, and or a particular transfer protocol, and, two, a particular type of data transfer.

[00:07:05] Speaker 02:
So the two stages... Okay, I understand what you're saying, but if I disagree with you and I say that the two-stage filtering is what I described to you, you lose then, right?

[00:07:16] Speaker 01:
So, I mean, I think I would lose, but I would urge your honor not to reach that conclusion on the 193 because I don't think that it makes sense of particularly claims 16 and 17. And, of course, this court can look to unasserted claims to determine the meaning of disputed terms that recur throughout the patent. And claims 16 and 17...

[00:07:35] Speaker 01:
Whereas, you know, the claims that the district court cited were the ones that, you know, they're depending claims on claim one that use methods like HTTP GET or HTTP POST. But claims 16 and 17 refer to the particular type of data transfer being data that contains a particular TLS value. And containing a, you know, packets with a TLS value are not consensually different from packets with an SGT tag.

[00:08:01] Speaker 01:
And essentially, if you read the two steps the way your honor is suggesting, then you're reading those two claims out of the patent. And I don't think that that would really make any sense because what Cisco argued to the PTAB, which was rejected, of course, was that, oh, you know, there are these two steps and we only, you know, discloses both, and what the PTEP said was not, you're right that patent 193 is sort of limited in this two-step way where the first stage is the five-tuple and the second stage is the method.

[00:08:43] Speaker 01:
It said, no, the prior art didn't disclose the operator, and that's why the prior art isn't covered, but there's no dispute, and Judge Hayes found very clearly, as a matter of fact, that the accused switches and routers do apply operators to drop packets or allow packets based on the particular type of data transfer. And the particular type of data transfer from network one to network two is data transfers that come with a quarantine SGT tag as opposed to data transfers that don't come with a quarantine SGT tag.

[00:09:16] Speaker 01:
And I just don't think that you can get to her affirming her infringement conclusion as to the 193 without agreeing with her that particular type of data transfer means something much narrower.

[00:09:29] Speaker 04:
Can I ask this? Is part of your argument that the paragraph on Colony, the two-sentence thing, It does say there are two stages, but it doesn't say that identifying a particular computer destination must be in stage one and can never be in stage two.

[00:09:55] Speaker 01:
That's correct.

[00:09:57] Speaker 04:
And as long as it's not dividing the world for particular computer destination designations, then it's not restrictive as to the second stage, let's call it the type, particular type of data transfer, put aside the method of transfer, which I don't think has any much to do with this at all, that it can also be the particular type of data transfer, a particular computer inside the network system.

[00:10:28] Speaker 04:
Right, the computer and the network.

[00:10:29] Speaker 01:
Exactly.

[00:10:30] Speaker 04:
And I think Judge Haynes recognized... And there's other things in the spec that seem to confirm that.

[00:10:35] Speaker 01:
Yes. So, I mean, I would point your honors to column 5, lines 25 to 42.

[00:10:39] Speaker 04:
It seems to me a lot of what we've been discussing is how much or even how to interpret this column 8... stage one, stage two language, and whether everything that is mentioned in the half of the sentence or the stage one thing can be there and only there, or it can also be in stage two.

[00:11:04] Speaker 01:
So I agree with you completely, and I think that it has to also be there in stage two. Otherwise, the unassertive claims, claims 16 and 17, don't make a whole lot of sense.

[00:11:14] Speaker 04:
The other rules that are... And am I right in thinking there's a kind of mismatch between type of data transfer as method of transfer and this kind of what feels like a separate point about the allocation between stage one and stage two of IP and TCP on the one hand and application addresses, that's not the right term, identification information, on the other hand.

[00:11:44] Speaker 01:
Right, so I think that's exactly right, and I think where Judge Haynes erred in focusing on column 8, or my friends on the other side erred in focusing on column 8, is not realizing that what column 8 says very specifically is it's talking about some embodiments. So it says, in some embodiments, that's the start of the sentence that she quoted from. And it goes on to make clear that while conceptually you might think of it as two stages where the first stage is packet level and the second stage is header level and the second stage is application level, such a method may, however, be used in other conceptual models.

[00:12:20] Speaker 01:
And what that's getting to is exactly the point that Your Honor is making, which is that sometimes the two-stage filtering sort of works as Judge Haynes found, but that's not the entirety of what the patent discloses. The patent discloses a much broader sort of technology that is – not just restricted to filtering based on the specific method associated with data transmission.

[00:12:47] Speaker 04:
Can you say a word about the next pattern?

[00:12:49] Speaker 02:
Sure. Could I ask you one other question before we move on? Yes, of course. And that is, do you agree that whatever the district court might have done in shifting her rationale from the original decision of the reconsideration of the CISCO argument about the two stages that you and I were talking about earlier was consistently made from the beginning.

[00:13:12] Speaker 01:
So Cisco certainly has made that argument all along. It made it to the PTAB. What we weren't on notice of was that the court was going to construe the claim for the first time. You were not on notice that the court was going to agree with them? So I think, I mean, you know, what I would say is, of course the district court's always allowed you to roll in claim constructions. But you have to give an opportunity to respond.

[00:13:36] Speaker 01:
And, you know, this case had been going on for a long time before the reconsideration.

[00:13:41] Speaker 02:
Why didn't you have the opportunity? You did respond, and you had the opportunity to respond to the argument that Cisco was making, which is the same argument which Cisco is defending here on appeal.

[00:13:51] Speaker 01:
So by the time of the Rule 52 motion, that specific issue was not teed up directly in the same way. So sure, we had briefed it earlier. We have evidence that actually shows, as we mentioned in our brief, that actually shows that we actually could satisfy that if we pointed her to it. But the reason that it's sort of unfair is when we spent the Rule 52 motion pointing out to Judge Haynes that her initial conclusion on the 193 depended on misconstruing a network device to meet an endpoint computer, this sort of one-stage, two-stage issue dropped to the side.

[00:14:27] Speaker 01:
I see I'm into my rebuttal time, which is briefly on the 806.

[00:14:31] Speaker 04:
Is either the gist or entirety of your... This is what, 806 is over on next?

[00:14:40] Speaker 01:
Yes, the 806 is the rule swapping.

[00:14:42] Speaker 04:
What has to be responsive to the signal is the...

[00:14:50] Speaker 04:
or the collection of five steps in the process, not any individual element like the ceasing, partly confirmed by the way in which the last step specifically adds the word step five responses to step four, and that's missing from the ceasing step. Is that?

[00:15:20] Speaker 01:
That's exactly right. And so what I would say is that the signal initiates in the sequence.

[00:15:25] Speaker 01:
And what is novel in the technology is that the sequence is responsive to the signal. And so I think it's just an error of law and logic to think that just because some of the constituent parts of the sequence may happen in normal packet processing.

[00:15:42] Speaker 04:
And in your view, that's the argument you made in your blue brief on this.

[00:15:46] Speaker 01:
That is our core argument.

[00:15:49] Speaker 04:
I will tell you, when I read the blue brief, I didn't quite see that. When I think Red came back and said, this is like saying, if I stand in front of a grandfather clock at 12.59 and 55 seconds and say, strike one, and it strikes that that's responsible to, and that's crazy.

[00:16:11] Speaker 04:
Then I went back and said, And then I guess I read your gray brief, and you said, no, we were always talking about the sequence. And I went back to the Blueberry and said, yes, you absolutely were always talking about the sequence. But there was something about it that I didn't quite get the first time around.

[00:16:27] Speaker 01:
Well, I apologize for the imprecision of our language in the Blueberry. But I mean, I do think, and I know I'm far out of time, but just in response to this, I mean, the standing ovation sort of analogy that we used in the blue brief, I think, you know, it's not only responsive, but it's... There were too many steps in it for me to be moved by it. But that is, in some way, sort of the point, is that there are so many steps in a standing ovation, but it's the sequence that matters. And I will return for... Thank you.

[00:16:55] Speaker 03:
We'll save two minutes for you. Thank you. Thank you.

[00:17:03] Speaker 00:
Good morning, and may it please the Court, Mark Fleming from WilmerHale, together with Heath Brooks and Nora Zhu on behalf of Cisco. Judge Haynes' factual findings are amply supported and certainly not clear error. I'll address the claim construction, but I think the easiest way through on the 193 patent is to recognize that the claim construction argument is actually beside the point, because however that term is construed, Judge Haynes made two factual findings that are unchallenged on appeals, that together defeat centripetal's infringement accusation.

[00:17:34] Speaker 00:
The first is Judge Haynes found, and this isn't disputed, that centripetal mapped this limitation, the packet filtering rule, to Cisco's quarantine rule and only the quarantine rule. She says that repeatedly. Their expert, Dr. Mitzenmacher, admitted it repeatedly. Second, Judge Haynes also found the quarantine rule only filters by one thing, destination of the packets and only destination of the packets. If they're headed for a forbidden destination, then they are dropped. If they're headed for a permitted destination, then they're allowed through.

[00:18:04] Speaker 00:
Centripital doesn't challenge that either.

[00:18:06] Speaker 04:
It's amply supported. Destination there includes IP address, TCP address, whatever.

[00:18:12] Speaker 00:
Only destination IP address. Where are they heading? Not what protocol do they use, nothing like that. There was no testimony whatsoever that the quarantine rule filters by port or protocol or anything like that. It's only by destination IP address. Again, that's not disputed. You won't see anything in their briefs suggesting clearly to the contrary and certainly no trial evidence. suggesting the quarantine rule does anything differently. So whatever, even if you think, arguendo, that Judge Haynes was wrong in how she interpreted a particular type of data transfer, it doesn't ultimately matter because she has found, in unchallenged, fully supported factual findings, that the accused quarantine rule, which is all they accused as a packet filtering rule, only filters by destination, and that's sufficient to affirm.

[00:19:04] Speaker 04:
If that can't be a particular type of data transfer.

[00:19:08] Speaker 00:
filtering by destination alone is not... I don't think they've ever argued that that's a particular type of data transfer because they recognize that that is only one of the stages of what the district court found... The claim doesn't use this stage thing, right?

[00:19:26] Speaker 04:
It's just language. The claim says from one network to another and then by particular type. So...

[00:19:34] Speaker 04:
The network doesn't have a single IP address, right? So if you pick an individual IP address within the first network or the second network, then you're not being duplicative of the first network, second network piece of the claim. So why can't an IP address be a particular type of data transfer?

[00:19:56] Speaker 00:
That is not something that there's any trial evidence on at all. The idea that going to a particular destination IP address is somehow both filtering by... Isn't this a claim construction point? No, it's not. It would have to be... First of all, there would have to be some kind of infringement evidence applying that kind of... I guess I'm starting from where...

[00:20:18] Speaker 04:
you were, which is to say, I know you're not supposed to do this on claim construction, but the only claim construction issue that would matter is something that would cover the IP address as the destination address, because that's the only thing the accused thing does, right? So why is an IP address...

[00:20:41] Speaker 04:
when put into the if part of the if-then rule, not specifying a particular type of data transfer. And here we're completely now putting aside the district court's view about method of data transfer, which seems somehow completely to ships passing in the night between what we're just talking about.

[00:21:04] Speaker 00:
Well, we're talking about an alternative. I mean, I'm more than happy to talk about the district court's claim construction, but before we get to that, let's be clear about what the patent actually says, because Mr. Rowan suggested that column 8 is only talking about an embodiment. In fact, what column 8 says, starting at line 39...

[00:21:22] Speaker 00:
is the filtering process described herein may be viewed as having two stages. So there's no limitation to an embodiment. It's the process described herein. And then going down to line 45, conceptually, the first stage may determine if the network policy allows any communications. any communications between the resources identified in the five-tuple rule. If so, the second stage may determine if the policy allows the specific method or type of communication, e.g. file read, file write, encrypted communication, etc., between the resources.

[00:21:55] Speaker 00:
And this is the exact language that they quoted to the PTAB on 19013 in defending these specific claims from invalidation. And again, it wasn't about an embodiment. These are their words on 19013. In other words, the independent claims of the 193 patent recite rules for preventing the transfer of a subset of packets out of the network. These rules involve implementing a two-stage process for filtering traffic. And they then block quote the language that I just read to your honor. And then on the following page, the bottom of 19.014, they say, once it is determined that a received packet matches the source destination criteria, Skipping down a bit, the second stage occurs wherein an operator is applied to drop packets associated with a particular type of data transfer, e.g.

[00:22:44] Speaker 00:
HTTP POST method. And so what this shows is that filtering by destination is the first stage. They called it that. That's how they got over the prior art. That's how their patent describes it. And that's why we say, on the one hand, it doesn't ultimately matter how you construe it because all that the Cisco quarantine rule, which is what they accused, does is filter by destination the first stage.

[00:23:08] Speaker 00:
And separately, we can talk about the claim construction itself. We think the claim construction is fully consistent with what the specification says, namely, once you've filtered by source and destination, when you're looking at whether it has a particular type of data transfer, you're going to be filtering by the various examples that the patent has, which are protocols, commands, HTTP post, FTP, SMTP. The quarantine rule does none of that, and they've never even suggested that it does. Their experts did not suggest that it does.

[00:23:37] Speaker 04:
Can you remind me, in the...

[00:23:43] Speaker 04:
the IPR document, that's a pattern of his preliminary response, right?

[00:23:48] Speaker 00:
Correct.

[00:23:48] Speaker 04:
And it was a successful preliminary response, so there was no institution, right? What was the point of distinction that you think Centripital was making there? I guess I have a general recollection, and you'll correct me, obviously, if I'm wrong, that this was not actually the point of the distinction.

[00:24:10] Speaker 00:
The second stage was precisely the point of the – well, maybe not precisely. It was a big part of their argument for distinguishing the prior art. And the prior art was Cisco prior art. It was a manual talking about the SourceFire product, which is a company that Cisco acquired. And the argument was that the SourceFire product filtered only at the first stage and didn't filter by particular type of data transfer. And so the meaning of that, what it meant to filter by particular type of data transfer in addition – to filtering at the first stage, was precisely the argument Centripital was making, and the public is entitled to rely on that as part of the intrinsic record, just as this court said in ALIS.

[00:24:49] Speaker 00:
Regarding this matter that they were not on notice, that this was our view of the ordinary meaning of the claim, I mean, as the questions have indicated, I mean, we've been clear from the beginning, and Judge Haynes finds this on page... Let me just ask you this.

[00:25:03] Speaker 04:
I mean, suppose I thought this is... not the ordinary meaning of the claim language, but that column 8 offers you rather striking support. What do I do?

[00:25:12] Speaker 00:
Well, I think you can simply affirm the claim language, the construction, because it was very clear... No, no, no.

[00:25:20] Speaker 04:
I'm sorry. Your point about an IP destination address cannot be the condition in an if-then rule, not the type of data transfer, which seems...

[00:25:32] Speaker 04:
Suppose, I think, that that's really far afield from any interpretation of the claim language. Just on this two-stage conception of the claim, where it looks like the two stages in the claim are not quite the two stages being described in column A.

[00:26:00] Speaker 04:
Because the network, right, the two stages in the claim are network.

[00:26:04] Speaker 00:
First network and the second network. That's right.

[00:26:05] Speaker 04:
And the eight is, even in your view, individual IP destination, particular device, i.e. not network, and something later in the rule. They're not identical.

[00:26:18] Speaker 00:
So that is not an argument that has ever been put to the district court, even in their motion for reconsideration. I'm not sure that's even in their briefing on appeal.

[00:26:27] Speaker 00:
So I don't think there's any basis to reverse what Judge Haynes has done on the basis of that kind of a reading of the claim. I mean, I think we're talking about it, I believe, for the first time this morning. I didn't hear Mr. Rowan make an argument like that. So I think you could say it's a very interesting possible conception had they made that argument, but it's not an obstacle to affirming what Judge Haynes did in this case. And they were certainly on notice of our understanding of the claim. Judge Haynes has a lengthy string site on page 95 of the appendix. We have more on page 43 and footnote 13 of our red brief.

[00:27:02] Speaker 00:
And Mr. Rohn didn't deny this is the conception of the claim that we've always asserted. As this court said in TQ Delta in Hamilton v. For Real Foods, that's plenty of notice that the court is at liberty to adopt that claim construction. It was discussed at length at the hearing in front of Judge Haynes. It starts at 21-233, where Judge Haynes asks Centripital's lawyer, what's the testimony regarding what that clause of words means or how it should be interpreted? And then there's a lengthy colloquy. They discuss the spec. They discuss column two.

[00:27:33] Speaker 00:
And then on 21237, counsel discusses Cisco's argument and talks about that for a while, and then in the end says, I think we've proven it both ways. This is 21245. So certainly Centripital understood and addressed our position about what the term meant, and nowhere suggested they needed to put in more evidence to counter it. In fact, they haven't even told this court. what difference it would make, what additional evidence they would have put in.

[00:27:59] Speaker 04:
Would you mind if I asked you to switch to the 806?

[00:28:02] Speaker 00:
Of course. Happy to, Your Honor.

[00:28:04] Speaker 00:
So centripetal calls this a claim construction argument. It's not. Centripetal did not seek a construction of responsive to being signaled, even in the reconsideration motion. So this ultimately is a fact issue, which is whether the products meet the ordinary meaning of the claims. They admit now, as the colloquy with Your Honor pointed out, that there has to be a causal relationship between the signal and what comes afterwards.

[00:28:31] Speaker 00:
and Cisco offered, as the district judge found, no clear expert opinion as to any causal relationship. That's on page 60. Judge Haynes credited our engineer's testimony.

[00:28:41] Speaker 04:
So I guess what I keep thinking is this. It seems to me almost self-evident that if you consider, what is it, the five steps after the colon as a single thing, that is obviously responsive because the signal triggers, you know, the sequence.

[00:28:59] Speaker 00:
You wouldn't swap a rule set unless you've gotten a signal to do it.

[00:29:01] Speaker 04:
Right. So, and it's, let's assume that the ceasing is not, in your view, is, you know, like my grandfather clock.

[00:29:12] Speaker 00:
Right. Absolutely.

[00:29:13] Speaker 04:
And it would have gone on anyway. So, There clearly is evidence of the causal connection if the claim is understood to refer to the five-step process as a whole. That's not a lack of evidence problem. This is just a claim construction problem.

[00:29:34] Speaker 00:
Well, it was never presented as such, even in their motion for reconsideration. Nobody said, Judge Haynes, you need to, or even Judge Morgan, you need to construe the meaning of responsive to. So at that point, we're just talking about comparing the undisputed findings now about how these things operate. Judge Haynes found there is no causal relationship to the ceasing or the caching. Remember, ceasing processing of packets is the very first thing after the responsive clause. So the very first thing you have to do, responsive to receiving the signal, is cease processing of the first portion of packets and cache them.

[00:30:09] Speaker 00:
And you have to do that throughout the reconfiguration process while the rules are being set up and the processors are being configured to use the second rule set. And whatever their argument might mean now, whether you call it claim construction or something else, surely the very first thing that comes after responsive has to be responsive to the signal. They could have written the claim differently. They could have simply... Under their conception, they presumably didn't need the word responsive to at all. They could simply have said...

[00:30:39] Speaker 00:
get a second rule set, cease processing the packets, cache the packets, reconfigure them, and then apply the second rule set. I don't think they could have gotten a claim that broad. Responsive to is something they had to add in order to get over the prior art. And so everything that comes after, and most certainly the first thing that comes after, needs to be responsive to the signal. And of course, there's a second problem with this patent, right? The second one that Judge Haynes found, which is the ceasing of packets in the caching It's not enough that it just happened during the brief idle time between two packets.

[00:31:10] Speaker 00:
It has to happen during the entire time that the processors are being reconfigured to use the second rule set. And our engineer, Mr. Shankar, testified that takes a few minutes. Now, we're fine with the packets continuing to be processed using the old rules. while the new rules and the processor is being configured to use the new ones. Their patent is not happy with that. Starting from column one, they say, the moment you get the new rule set, you want to stop processing packets using the old rule set, because otherwise you're not using the latest and greatest intelligence, and you might be letting something through that will be an attack or an exfiltration.

[00:31:46] Speaker 00:
So you want to stop, you want to cache everything, wait until the new rule set is ready, and then take everything out of the cache and process it using the new rule set. We don't do that, and Judge Haynes found that this was a requirement of the claims, that the ceasing of processing has to be entirely, it has to happen altogether until the new rule set is ready for use. And we simply don't do that, and there's plenty of evidence supporting it. That's their patent. They're entitled to it. They think there's a security reason to do it. We don't do it. We've made a different trade-off.

[00:32:17] Speaker 00:
We would rather the packets go through and that the communication continue, even if they're being processed using the old rule set. That's what their patent disparages and does not claim. There was no discussion of the 176, so I won't address it unless the Court has questions.

[00:32:33] Speaker 03:
Thank you, Mr. Fleming. Mr. Rowan has a couple of time, two minutes.

[00:32:38] Speaker 00:
Thank you.

[00:32:43] Speaker 01:
Thank you, Your Honors. I'll pick up on the 806.

[00:32:46] Speaker 01:
So it is a claim construction issue in the sense that we are just asking that the term responsive to be construed consistent with its claim meaning. That's not something for which you need to ask for claim construction because... Of course, the default assumption is that the terms will be construed. Responsive to doesn't suggest causation? Of course it suggests causation, and we've never argued that it doesn't. Our argument has been that the thing that has to be caused is the sequence. The signal causes the sequence.

[00:33:14] Speaker 04:
Did you present that below as a claim construction issue? That is not a question of claim construction of responsive to. It's a question of responsive to with some more words added. That whole phrase has to be construed.

[00:33:30] Speaker 01:
Off the top of my head, I don't have a citation to give you for putting it that way, but I know that the first time that it was put forward that the reason that we couldn't show infringement was that the elements were not responsive to was in Judge Gaines' opinion. So a little bit of timing of when... If we were not required to put a claim construction argument forward because we're saying it's ordinary meaning... then it's a little bit, you know, difficult to think of it.

[00:34:02] Speaker 01:
It sort of morphs into a factual finding.

[00:34:05] Speaker 01:
And, you know, my friend on the other side, you know, relied on the Schenker testimony, which Judge Haynes discusses at Appendix 66.

[00:34:13] Speaker 04:
I think it's important to note that... This is that it'll take a few minutes sometimes to actually configure. Correct. This is after all of the compiling is done and it's in. active memory. It still takes minutes, really, to figure.

[00:34:31] Speaker 01:
So that's their testimony, but Judge Haynes made clear, this is at footnote 5 of Appendix 103 in her reconsideration opinion, that that discussion formed no basis of her infringement finding, because of course it couldn't. The fact that it didn't have the intended outcome is not a basis to find non-infringement. And then just very briefly on the 193, I don't think there is a way to affirm without affirming Judge Haynes' claim construction with respect to a particular type of data transfer because she's very clear.

[00:35:02] Speaker 01:
The reason she found infringement is she said, this is Appendix 94, quote, the problem with Centrifugal's argument is that a quarantine SDT tag... I'm sorry. I'm sorry.

[00:35:11] Speaker 04:
To delay you one more minute, Mr. Fleming's argument was actually none of that matters because you know exactly what the...

[00:35:23] Speaker 04:
what's SGC or something, is doing. And the only question is, can that be in the if clause of the rule?

[00:35:33] Speaker 01:
And I think as a matter of plain English, of course it can, in the same way that... Now violating the two-stage idea of colony.

[00:35:42] Speaker 02:
Right, and your whole argument is that two stages are distinct. And they're presented as distinct in the specification and in the prosecution history, which you don't even discuss in your briefs.

[00:35:58] Speaker 01:
So I think our argument is that the two stages don't have to be distinct, that they can be.

[00:36:01] Speaker 02:
It's the evidence that they're not distinct. When you look at that prosecution history, it's described as two distinct stages.

[00:36:10] Speaker 01:
So I don't think that's correct for the quote that I gave you before from 19013, which is the preceding page to the page that Mr. Fleming quoted, where We made it very clear that the quote-unquote first stage can be based either on the protocol, the data protocol type, or the source and destination value. And so that just shows right there that those two things are not the two stages. Because in defending against IPR, we made very clear, and this is clear throughout the patent, that it can be the case that the quote-unquote first stage is based just on source or destination IP, which it doesn't have to be.

[00:36:51] Speaker 01:
And that's the problem with Judge Haynes' ultimate conclusion, is that she read a limitation into the claims based on some snippets from the specification that's ultimately inconsistent.