[00:00:00] Speaker 01:
Our argument is 20-1768 Centripital networks versus Cisco.

[00:00:06] Speaker 01:
Mr. Hanna.

[00:00:08] Speaker 05:
Thank you, Your Honor.

[00:00:09] Speaker 05:
Again, may it please the Court.

[00:00:11] Speaker 05:
So with respect to this appeal, this is the 20-1768.

[00:00:15] Speaker 05:
We're talking about the 148 patent.

[00:00:17] Speaker 01:
And there's no claim construction issue.

[00:00:19] Speaker 01:
That was just for the first two, right?

[00:00:23] Speaker 05:
There's no claim construction issue with regard to dynamic security policy.

[00:00:26] Speaker 05:
That's correct.

[00:00:28] Speaker 05:
In terms of what the board did here is they looked at the word pre-processing and said that, I mean, they looked at post-processing and said that can mean pre-processing.

[00:00:43] Speaker 05:
And to want to skill in the art and to aid in lay persons, that simply does not make sense.

[00:00:51] Speaker 05:
Here we have two references.

[00:00:53] Speaker 05:
We have the Rose reference, and we have the Galnavi reference that is relevant to this pre-processing of rule sets.

[00:01:01] Speaker 05:
And here, the board did construe pre-processing of the rule sets.

[00:01:06] Speaker 05:
And they said that it's performing optimization operations on a first rule set and a second rule set prior to the rule sets being applied to packets.

[00:01:18] Speaker 05:
That's the important part here, is that it's prior to the rule sets being applied to packets.

[00:01:24] Speaker 01:
Yeah, but maybe to cut this to, I think, where you're going, so you can correct me if I'm wrong, the Rose reference says at any time.

[00:01:34] Speaker 01:
So at any time is really broad.

[00:01:37] Speaker 01:
And the board concluded, I think, that any time includes prior to applying rule sets to packets.

[00:01:43] Speaker 01:
So is that what you're taking issue with?

[00:01:46] Speaker 05:
Well, we got to be careful here because Rose says at any time you can select policies.

[00:01:53] Speaker 05:
It doesn't say that at any time you can optimize rule sets.

[00:01:58] Speaker 05:
And that's a big difference here.

[00:02:00] Speaker 05:
Just because a reference says the words at any time doesn't mean that you can use now Rose to apply to any prior art reference at any, you know, anywhere in the world and say, okay, now it says at any time, and so we can completely divorce

[00:02:15] Speaker 05:
the technology and say, okay, post-processing actually means pre-processing.

[00:02:20] Speaker 05:
I don't think that's correct at all.

[00:02:22] Speaker 05:
You have to look at the statements in context, and the statements here are that you can select a policy at any time.

[00:02:30] Speaker 05:
And ROSE is very specific.

[00:02:33] Speaker 05:
It's about selecting these policies.

[00:02:36] Speaker 05:
It has nothing to do with optimization.

[00:02:39] Speaker 05:
And that's why Cisco had to point to Gullnabi.

[00:02:42] Speaker 03:
And then when you... Of course, Councilor, your standard review is pretty steep on the question of what a reference teaches.

[00:02:51] Speaker 05:
This is undisputed.

[00:02:53] Speaker 05:
It's undisputed that Rose teaches the selection of policies.

[00:02:57] Speaker 05:
The board confirmed that again and again.

[00:03:00] Speaker 05:
So that's undisputed.

[00:03:04] Speaker 05:
Gnobby, it's also undisputed.

[00:03:07] Speaker 05:
The board specifically says that Gnobby teaches

[00:03:10] Speaker 05:
the post-optimization of rules.

[00:03:12] Speaker 05:
They specifically say that you look at the rules, you apply the rules to packets, and then, quote, during packet processing, you're going to try to optimize those rules.

[00:03:27] Speaker 05:
So that's also undisputed.

[00:03:30] Speaker 05:
It really comes down to this notion that if you look at the post-processing of rules,

[00:03:38] Speaker 05:
that one of, and this is another quote, this is what the board ruled.

[00:03:43] Speaker 05:
We could not agree with patent owner that optimization of rules before the processing of packets and after processing of packets are such different considerations that a person would not have been considered such processes useful for optimization at any period.

[00:03:59] Speaker 05:
That simply makes no sense.

[00:04:01] Speaker 05:
That's why what they're trying to do is fill this limitation in with common sense.

[00:04:07] Speaker 05:
They're trying to say, if I had post-processing that I'd normally, then I would just automatically apply that to pre-processing.

[00:04:16] Speaker 05:
And so there's actually, the board is in agreement that Rose discloses selecting apologies, not optimization, and that's where the anytime language comes in.

[00:04:27] Speaker 05:
It's undisputed that Gnobby teaches post-processing or post-optimization, and it doesn't teach pre-optimization or optimization before applying to packets.

[00:04:37] Speaker 05:
And so the question really comes down to is, can the board fill in that limitation with common sense?

[00:04:45] Speaker 05:
And the answer is simply no.

[00:04:46] Speaker 05:
The case law is very clear.

[00:04:48] Speaker 05:
Arendi is very clear here that common sense cannot be used to fill in an element.

[00:04:54] Speaker 05:
In addition, that element would have to be simple.

[00:04:58] Speaker 05:
Security software is complex.

[00:05:00] Speaker 05:
And where the processing happens matters.

[00:05:03] Speaker 05:
I mean, if you think about it.

[00:05:06] Speaker 05:
Using the Rose and Denabi system combined, what you're going to have is a system that traffic comes into your network, and it's potentially malicious, and it can infect your network.

[00:05:20] Speaker 05:
And then you adjust the rules based on the traffic that comes in to starve off any future attacks.

[00:05:28] Speaker 05:
But what the pattern is about is not allowing that traffic in in the first place.

[00:05:33] Speaker 05:
That's why you have this pre-processing, this optimization before you actually apply the rules so that I don't get infected at all.

[00:05:42] Speaker 05:
I can optimize these rules.

[00:05:44] Speaker 05:
I can prevent malicious traffic from entering my network before it even hits it.

[00:05:52] Speaker 05:
And that's why this patent and this pre-processing, this pre-optimization is so important and how it's just completely contrary to how Gnobby works.

[00:06:02] Speaker 05:
I mean, Gnobby would not work in a pre-processing context because, as your Honors knows, Gnobby specifically says you have to log, you have to mine this log data.

[00:06:15] Speaker 05:
That log data is these packets that are being applied to these rules.

[00:06:19] Speaker 05:
There's no way that you could actually have that log data if it wasn't applied to the rules.

[00:06:26] Speaker 05:
And then, to your Honors point, you say that, well, Rose says that it can be at any time.

[00:06:29] Speaker 05:
Well, if I'm one to skill in the art,

[00:06:31] Speaker 05:
And I'm looking at Rose, and it says at any time.

[00:06:34] Speaker 05:
And then I look at Ganaubi.

[00:06:36] Speaker 05:
Ganaubi says post-processing, that's what I'm thinking.

[00:06:40] Speaker 05:
I'm not thinking, I'm thinking what the reference discloses.

[00:06:43] Speaker 05:
I'm thinking, okay, I can do this at any, I can do, even if I could do it at any time, but Rose says selecting policies, but I could say, okay, the words at any time.

[00:06:52] Speaker 05:
I look at Ganaubi and says, okay, at any time, you know when you should do it, you should do it after you apply the rules, because that's,

[00:06:58] Speaker 05:
what this system is.

[00:06:59] Speaker 05:
That's how they think.

[00:07:01] Speaker 01:
Well, that was the anticipation for Gnobby.

[00:07:03] Speaker 01:
I mean, I don't understand.

[00:07:04] Speaker 01:
I mean, when you combine references, you take part of one and part of the other.

[00:07:10] Speaker 01:
So if you've got Rose saying at any time, can't you transfer that and combine it with Gnobby, right?

[00:07:17] Speaker 01:
No.

[00:07:19] Speaker 05:
Well, Your Honor, it's at any time I could select a policy.

[00:07:23] Speaker 05:
And then at any time I can select a policy, but then the policy can't be optimized, those rules can't be optimized until after they're applied to packets.

[00:07:31] Speaker 05:
That's the combination.

[00:07:32] Speaker 05:
That's what the at any time means.

[00:07:34] Speaker 05:
It doesn't mean at any time I can do optimization.

[00:07:36] Speaker 05:
ROSE is completely silent as to optimization.

[00:07:39] Speaker 05:
And Gnobby never says that you can do processing, pre-processing of those rules because it wouldn't work.

[00:07:46] Speaker 05:
It just simply would not, it just wouldn't work.

[00:07:49] Speaker 05:
So at most you'd have a system that says,

[00:07:52] Speaker 05:
At any time, I can select a policy and then optimize my rules, I have to apply it to package.

[00:07:59] Speaker 04:
Okay.

[00:08:00] Speaker 04:
Council, I think the board found that Rose teaches that a policy, and this is a quote, may be pre-installed, updated, reinstalled, revised, or otherwise changed when and as desired.

[00:08:16] Speaker 05:
Correct, but there's nothing about optimization there, Your Honor.

[00:08:19] Speaker 05:
That doesn't talk about optimization.

[00:08:22] Speaker 05:
So it says that it can be done as desired.

[00:08:25] Speaker 05:
And also, the words updating and so forth imply that there's already a problem.

[00:08:34] Speaker 04:
Reinstall, revised, or otherwise changed when and as desired.

[00:08:40] Speaker 05:
Right.

[00:08:40] Speaker 05:
OK, so first of all, that has nothing to do about optimization.

[00:08:44] Speaker 04:
Well, I think Cisco's expert, Dr. Reddy, he said that those teachings, the one I just pointed out, would inform a procedure that pre-processing could be performed before the rules are applied to data packets.

[00:08:57] Speaker 05:
Processing?

[00:08:58] Speaker 04:
That was Dr. Reddy's testimony.

[00:09:02] Speaker 04:
And again, we have a substantial evidence standard here.

[00:09:09] Speaker 04:
That makes it a pretty hard role for you to hope.

[00:09:13] Speaker 05:
But Your Honor, look at that language.

[00:09:15] Speaker 05:
That language is processing, not optimization.

[00:09:20] Speaker 05:
You have to look at optimization.

[00:09:22] Speaker 05:
He never ever says, and he can't say, that Rose discloses optimization.

[00:09:27] Speaker 05:
He talks about processing every single time.

[00:09:30] Speaker 05:
Because the board construed that the claims require optimization of the first and second rule set before applying it to packets.

[00:09:38] Speaker 05:
And so, I mean, processing can be a number of things, but optimization is going to be actually being able to optimize that rule set, and that's why they had to point to GUNABI.

[00:09:49] Speaker 05:
So if you look at that language every single time, and I check, every single time he says it, he always says processing, he never says optimization because he can't say that.

[00:10:00] Speaker 05:
And so this optimization, you have to look to GUNABI in order to see

[00:10:06] Speaker 05:
when this is going to occur, and every single time Gnobby talks about it, it's actually after it's being applied to packets.

[00:10:14] Speaker 05:
And so again, it comes back to this, can the board fill this in, fill in this limitation with common sense?

[00:10:22] Speaker 05:
And Arendi and all the cases that we cited specifically say that you can't do that.

[00:10:27] Speaker 05:
You can't fill that in.

[00:10:30] Speaker 05:
And in fact,

[00:10:32] Speaker 04:
Here we're talking about motivation to combine, so why is it wrong to use Gnobby's teaching on the optimization point?

[00:10:42] Speaker 05:
Because the optimization has to be prior to being applied to packets.

[00:10:49] Speaker 05:
That's the construction the board used.

[00:10:51] Speaker 05:
Gnobby does not teach prior to being applied to packets.

[00:10:54] Speaker 04:
Well, when you combine, though, that with Rose, I mean, you get exactly that.

[00:11:01] Speaker 05:
You don't, Your Honor, because

[00:11:02] Speaker 05:
If you combine them, you have at any time you can select a policy.

[00:11:06] Speaker 05:
You can select one.

[00:11:08] Speaker 05:
Nothing about optimization.

[00:11:09] Speaker 05:
So at any time you can select a policy.

[00:11:12] Speaker 05:
But that policy is not optimized until it's applied to packets.

[00:11:18] Speaker 05:
So this is the leak that the board had to make, is that they couldn't point to anything in the references.

[00:11:27] Speaker 05:
The motivation and all that doesn't matter because they can't point to anything there.

[00:11:31] Speaker 05:
They couldn't point to anything to meet that element.

[00:11:34] Speaker 05:
And so that's why that quote that I read is that they had to specifically rely on common sense to say, well, when you're talking about post-optimization, one of skill in the art would understand that you could do pre-optimization.

[00:11:50] Speaker 05:
So they never, they can't point to anything because it actually teaches away from that.

[00:11:59] Speaker 05:
So the other point that we have is this hater reference.

[00:12:02] Speaker 05:
Now, the claims also require that you need to cache packets.

[00:12:08] Speaker 05:
Hater specifically discloses, it's even in the abstract of hater, that only cache parts of the packet.

[00:12:16] Speaker 05:
You do not cache the entire packet.

[00:12:19] Speaker 05:
And nowhere in that reference does it talk about caching the packets as required by the claim.

[00:12:26] Speaker 05:
So for this additional reason,

[00:12:29] Speaker 04:
The cashing limitation was not construed, correct?

[00:12:35] Speaker 05:
Cashing was not construed.

[00:12:36] Speaker 05:
That is correct, Your Honor.

[00:12:37] Speaker ?:
Okay.

[00:12:38] Speaker 04:
All right.

[00:12:38] Speaker 05:
Thank you.

[00:12:39] Speaker 05:
All right.

[00:12:40] Speaker 05:
I just heard the buzzer, so I'd like to save the rest of my time for a rebuttal unless there's questions that I could answer, Your Honor.

[00:12:45] Speaker 01:
No.

[00:12:45] Speaker 01:
That's fine.

[00:12:46] Speaker 01:
Thank you.

[00:12:47] Speaker 01:
Mr. Foster?

[00:12:50] Speaker 02:
Thank you.

[00:12:51] Speaker 02:
May it please the court, Theodore Foster, on behalf of Appalachian Cisco Systems

[00:12:56] Speaker 02:
The evidence cited by the board supports its conclusion of unpatentability.

[00:13:01] Speaker 02:
As the board stated at Appendix 16, as the court noted in the earlier argument, the Rose Prior Art describes adjusting rule sets at any time, which includes prior to applying rules to packets.

[00:13:15] Speaker 02:
And it reiterates this in the quote one of the judges referenced earlier, that rule sets may be, quote, revised or otherwise changed

[00:13:24] Speaker 02:
when and as desired."

[00:13:28] Speaker 02:
I would further note that the Rose reference emphasizes the need for efficiency in its paragraph 18, and that ties in with the Gulnabi reference, which improves the efficiency of rule sets using the same merging, splitting, and reordering techniques discussed in the 148 patent.

[00:13:46] Speaker 02:
The board also noted at Appendix 17 that Centripetal's expert, Dr. Orso, conceded that those of skill in the art knew how to optimize rules in a pre-processing fashion.

[00:13:58] Speaker 02:
The combined teachings of the prior art, the combination of Rose's teaching of adjusting rules at any time and Goulmabi's specific techniques and disclosure of optimizing rules using merging, splitting, and reordering,

[00:14:13] Speaker 02:
together render obvious the claimed pre-processing step.

[00:14:22] Speaker 02:
Regarding the caching limitation that Council was beginning to discuss, I would note that Hader discloses at Appendix 2096 that caching, its caching disclosure includes caching the entirety of the header.

[00:14:38] Speaker 02:
and at least one byte of the payload, which is an open-ended description that also encompasses caching the entirety of the payload, as Centripetal is now arguing its claims require.

[00:14:56] Speaker 02:
Court does not have any questions for me on these two issues.

[00:15:00] Speaker 02:
I would cede the remainder of my time.

[00:15:02] Speaker 01:
Well, let me ask you just one quick thing.

[00:15:05] Speaker 01:
We haven't discussed, but you offered up an alternative, which was disagreeing with the board's claim construction.

[00:15:14] Speaker 01:
And in red, I think the answer was, in gray, that that was waived.

[00:15:21] Speaker 01:
So you haven't had a chance to respond to that waiver point.

[00:15:24] Speaker 01:
So take a minute and do that.

[00:15:26] Speaker 02:
Sure, I don't believe that argument was waived.

[00:15:30] Speaker 02:
While claim construction of the pre-processing term was an issue and the board did issue a construction, I think there were further latent issues that remained unresolved by the board's construction.

[00:15:46] Speaker 02:
I would point the court to appendix page 340, which is a page from the petitioner's replies and the middle paragraph there

[00:15:56] Speaker 02:
we cited to Dr. Reddy's testimony, his further testimony offered with the reply that the 148 patent does not exclude processing packets with rule sets while logging data and then using that logging data to merge split or reorder rule sets, then with those newly updated, revised, optimized and improved rule sets, processing further packets.

[00:16:25] Speaker 02:
And that further explains

[00:16:27] Speaker 02:
you know, by Dr. Reddy looking at Appendix Page 2867, which is cited from Appendix 340, Paragraph 28, he gives a specific example of what he means and what we mean by this idea that the claim simply requires before processing of some packets, not

[00:16:51] Speaker 02:
before processing any packets ever.

[00:16:53] Speaker 01:
And I would also note that in the... I'm sorry, Judge... No, I was just going to say that your argument was your answer was sufficient.

[00:17:03] Speaker 01:
Okay.

[00:17:03] Speaker 01:
I mean, but if you want to add something, go ahead.

[00:17:08] Speaker 02:
My only other comment was that in the gray brief, centripetal feet seemed to pull back

[00:17:14] Speaker 02:
from their sometimes statements that the pre-processing had to be before processing any packets ever.

[00:17:23] Speaker 02:
And if it's not before processing any packets ever, then even the specific techniques in Golnavi, which use data logging, could be used to optimize rules and then subsequently have those rules be applied to further future packets and that that would be pre-processing.

[00:17:41] Speaker 01:
Okay.

[00:17:42] Speaker 01:
Thank you.

[00:17:43] Speaker 02:
Thank you.

[00:17:44] Speaker 01:
Mr. Hanna, final words?

[00:17:49] Speaker 05:
Yes, Your Honors, and thank you.

[00:17:50] Speaker 05:
Thank you for staying with me all day today.

[00:17:54] Speaker 05:
But what I didn't hear in counsel's argument is any rebuttal to the fact that the board had to apply common sense in order to meet this development.

[00:18:03] Speaker 05:
And again, I point to Appendix 17 in which the board specifically stated that they could not find that there was this pre-optimization of rules

[00:18:14] Speaker 05:
this optimization rules before being applied to packets, and so they had to rely on common sense there.

[00:18:20] Speaker 05:
And just to drive it home, this optimization is narrower than this processing.

[00:18:28] Speaker 05:
Processing can mean a number of things.

[00:18:29] Speaker 05:
This optimization has to be on this particular rule set, and so a disclosure of

[00:18:36] Speaker 05:
the processing happening is not, doesn't meet the claims limitation or the claims instruction here because it specifically requires the optimization of this first and second rule set.

[00:18:49] Speaker 05:
So any disclosure in rows about any time selecting these policies or even adjusting, there's nothing talking about optimization.

[00:18:58] Speaker 05:
And so one is skill in the art, looking at that and they have to look to when are you gonna optimize

[00:19:04] Speaker 05:
can now be only said to optimize after you apply it to the packets.

[00:19:10] Speaker 05:
With respect to this claim construction issue that Your Honor has raised, you know, I note that it's completely contrary to the construction that they actually offered in the underlying proceedings.

[00:19:23] Speaker 05:
During, on appeal, they say it's processing to occur before some packets are processed, but not necessarily before a device processes any packets.

[00:19:32] Speaker 05:
In the lower proceedings, they actually argued optimizing by network device a first rule set and a second rule set before the network device processes packets.

[00:19:41] Speaker 05:
And so it's contrary to the positions that they took lower in the lower court.

[00:19:47] Speaker 05:
And again, with regard to the hater disclosing caching of packets, again, if we just look at the abstract of hater, there's nothing that says that you're actually going to cache the entire packet.

[00:20:00] Speaker 05:
It's simply not disclosed.

[00:20:03] Speaker 05:
There was no articulated reasons in the board's decision for cashing the entire packet.

[00:20:09] Speaker 05:
And so we're left with this court having nothing to analyze or review.

[00:20:19] Speaker 05:
With that, Your Honors, unless you have any, you know, I'm happy to add to any further questions you have about any of these references to these decisions.

[00:20:26] Speaker 01:
Hearing none, thank you.

[00:20:28] Speaker 01:
We thank both sides and the case is submitted.