[00:00:00] Speaker 02:
2336, Arista versus Cisco.

[00:00:02] Speaker 02:
I guess you get to stay in the same place, which we will do throughout.

[00:00:06] Speaker 02:
Mr. Powers.

[00:00:15] Speaker 03:
In this portion of the argument, I will only address Claim 2 of the 577 patent.

[00:00:20] Speaker 03:
That is the subject we're appealing.

[00:00:23] Speaker 03:
And they are cross-appealing on all the other claims.

[00:00:25] Speaker 03:
And I will handle that in rebuttal.

[00:00:27] Speaker 03:
With regard to Claim 2, it's a single issue, and that's whether Huey discloses this parallel using a pipeline system in parallel.

[00:00:38] Speaker 03:
There is no debate.

[00:00:39] Speaker 02:
Is this... Okay.

[00:00:41] Speaker 02:
So that issue is still alive.

[00:00:43] Speaker 03:
It's not been... It is still alive.

[00:00:45] Speaker 03:
Okay.

[00:00:46] Speaker 03:
Claim 2 is one that adds the parallel pipeline limitation.

[00:00:50] Speaker 03:
There is no dispute that Huey discloses in Figure 4 a pipeline system where it moves from left to right

[00:00:56] Speaker 03:
from the various stages of going from the address handler to the cell router to the cell policer.

[00:01:04] Speaker 03:
There is no debate that that's a pipeline system.

[00:01:06] Speaker 03:
The only question below is whether that was operating in parallel.

[00:01:10] Speaker 03:
The board did not find that it was not operating in parallel.

[00:01:13] Speaker 03:
The board simply said, we don't see the words in parallel or anything of that effect.

[00:01:16] Speaker 03:
We don't see an explicit disclosure, which doesn't answer the question.

[00:01:20] Speaker 03:
The question is whether one skilled in the art reading Huey as a whole would understand that it

[00:01:25] Speaker 03:
that that pipeline system was operating in parallel, i.e.

[00:01:29] Speaker 03:
that when packet one moved from stage one to stage two, that the system was operating on packet one in stage two at the same time as the system was operating on packet two in stage one.

[00:01:42] Speaker 03:
That's the question, and that question is answered definitively by Huey, albeit not in words that the board understood to be in parallel, when it said that at column nine,

[00:01:55] Speaker 03:
that that comparison being done in stage one is done, is done, quote, in real time, meaning those packets are flowing through, not going through one packet and then waiting until full processing and then sending the other packet, that they are flowing in real time.

[00:02:10] Speaker 03:
And that question is also answered by Huey column three, selling the, telling you that there's a stream of cells going through at the cell input rate.

[00:02:20] Speaker 03:
And that answers the question as to whether, in fact, those cells are progressing through that pipeline

[00:02:25] Speaker 03:
and being operated on in parallel.

[00:02:27] Speaker 00:
Was there an expert declaration commenting on Columns 3 and 9 to the effect that you're telling us now?

[00:02:33] Speaker 03:
The expert testimony is Dr. Chao, Arista's expert.

[00:02:37] Speaker 03:
It was in his deposition.

[00:02:38] Speaker 03:
There was not a report on that question, but he was asked about this exact issue.

[00:02:42] Speaker 03:
That's in the record at Appendix 2404.

[00:02:46] Speaker 03:
That's where we cited it.

[00:02:47] Speaker 03:
His testimony is at 1788 to 89.

[00:02:50] Speaker 03:
That's unrebutted testimony, where he says that is in parallel.

[00:02:56] Speaker 03:
Uh, and that's the interpretation of one skilled in the art of how something is operating in real time, meaning those things are flowing through.

[00:03:04] Speaker 03:
Beyond that, did you say 2404?

[00:03:05] Speaker 03:
2404 is where we cited it.

[00:03:08] Speaker 03:
The testimony itself is at 1788 to 89.

[00:03:14] Speaker 03:
Beyond that, we cited the prior, the long reference called Kogi, K-O-G-G-E, which is at 3298.

[00:03:20] Speaker 03:
That was a 1981 treatise.

[00:03:24] Speaker 03:
establishing that was well recognized in the art long before this time, that in fact the flow through those stages in a pipeline system would, as Kogi said, just be limited to the rate at which it's fed into the input.

[00:03:38] Speaker 03:
So it's flowing through, operating on stage two, at the same time you're operating on a different packet in stage one.

[00:03:44] Speaker 03:
That is in parallel.

[00:03:45] Speaker 03:
So the board's refusal to find that Huey satisfied the parallel limitation simply because it didn't use the words was a failure of the board

[00:03:54] Speaker 03:
to recognize the import of the Chow testimony and the Kogi reference.

[00:03:57] Speaker 03:
The Kogi reference we're not using as a combination.

[00:03:59] Speaker 03:
We're merely using it to show how one skilled in the art would understand the operation of a pipeline system.

[00:04:06] Speaker 03:
And the board's failure to recognize that simply relying on the absence of language was error.

[00:04:19] Speaker 02:
You weren't planning to re-discuss the Assigner-Stoppel issue.

[00:04:23] Speaker 01:
I wasn't planning to.

[00:04:24] Speaker 01:
Thank you, Chief Judge Prost.

[00:04:26] Speaker 01:
May it please the Court.

[00:04:27] Speaker 01:
And as Your Honor noted, for the reasons I just discussed, the Board should have applied the doctrine of Assigner-Stoppel, but I'm not going to reiterate those arguments other than to say, if ever there was a case for Assigner-Stoppel, this is it.

[00:04:40] Speaker 01:
It's a patent that Dr. Sheridan and Mr. Bechtelsheim assigned to Cisco.

[00:04:44] Speaker 01:
They then founded Arista.

[00:04:45] Speaker 01:
And then Arista now wants to invalidate the patent.

[00:04:48] Speaker 00:
Can you tell me what the nature of the assignment was?

[00:04:50] Speaker 00:
Was it your standard employment agreement when you join up with a company and then for a salary and benefits, the company gets to retain the fruits of your labor, including anything that you come up with if there are any patents that are gained.

[00:05:10] Speaker 00:
They belong to the company.

[00:05:12] Speaker 00:
They don't belong to the employee.

[00:05:14] Speaker 01:
Judge Shin, let me just put it this way.

[00:05:16] Speaker 01:
Both of these individuals were very handsomely compensated by Cisco.

[00:05:19] Speaker 01:
That is detailed in the record.

[00:05:23] Speaker 01:
Not that I think anything turns on whether or not this is just a normal employment assignment, because I think this court's case law is very clear that that is sufficient.

[00:05:32] Speaker 01:
I suppose it could be characterized this way, but I think that there was more going on there.

[00:05:37] Speaker 01:
But again, I think that if you look at the points in the record that we've articulated

[00:05:41] Speaker 01:
You'll see that they were very compensated by Cisco for the inventions that they assigned to Cisco.

[00:05:48] Speaker 01:
And then they went off and founded Arista and now seek to profit from invalidating these patents.

[00:05:53] Speaker 01:
Now, separate apart from Assigner-Restapel, there are three reasons that this case should be reversed or at least remanded.

[00:06:03] Speaker 01:
First, it's the board's treatment of the in response to limitation as merely requiring a sequential rather than a causal relationship.

[00:06:11] Speaker 01:
Second, the board's failure to require that the alleged access control patterns actually be used for access control.

[00:06:19] Speaker 01:
And third, its failure to require hardware enforcement of access control.

[00:06:24] Speaker 01:
Any one of these provides an independent basis for reversal or remand, and any one of those would also address claim two of the patent.

[00:06:31] Speaker 01:
So I want to take each in turn.

[00:06:34] Speaker 01:
Now, the claims require generating an access result in response to a match between a packet label

[00:06:41] Speaker 01:
and the access control pattern.

[00:06:42] Speaker 01:
And you didn't see construction of the term.

[00:06:45] Speaker 01:
Well, Chief Judge Crost, you won't see us asking for one under the heading of claim construction.

[00:06:53] Speaker 01:
Please construe this term the following way.

[00:06:55] Speaker 01:
But I think this is exactly like what this court recently addressed in HTC versus cellular communications, 877 F3 at 1367, which came down in December.

[00:07:07] Speaker 01:
And it's just like the issue that was in

[00:07:09] Speaker 01:
Vizio versus ITC, which is where you have the board engaging in an implied construction.

[00:07:16] Speaker 01:
Yes, it may have been teed up in the context of arguing non-infringement, and that's exactly what was at issue in Vizio.

[00:07:22] Speaker 01:
And the majority there over the dissent said, no, this still presents an issue of claim construction.

[00:07:28] Speaker 01:
And to be clear, there's no doubt that this issue was teed up and that the board understood that the issue went to what the scope of the claims were.

[00:07:35] Speaker 01:
The board...

[00:07:37] Speaker 02:
I don't want to use your time, but assuming we would agree with you that in response to requires causation, why would that change the result?

[00:07:47] Speaker 01:
So for fundamental reasons, Chief Judge Prost, because if you agree that it requires causation, there is no causation here

[00:08:00] Speaker 01:
between what the board is treating as the access control match and then the access determination.

[00:08:11] Speaker 01:
And perhaps the best way to see this, the board treats the VCVP header match as being... Do you want to refer us to the page in the board's opinion?

[00:08:22] Speaker 02:
Is that 35?

[00:08:24] Speaker 01:
I believe that is right, Chief Judge Prost.

[00:08:31] Speaker 01:
If you look at appendix 26 and appendix 27, it is where the board is addressing this issue.

[00:08:42] Speaker 01:
And the alleged access control pattern are the VCVP addresses.

[00:08:58] Speaker 01:
And the board then says, and this is appendix 27, the VCVP addresses used to make a match are not required to also generate the access result.

[00:09:10] Speaker 01:
Now what that means, and this is important, the VCVP addresses are what are being treated as the access control pattern.

[00:09:17] Speaker 01:
It is then this cell policer in Huey being combined with the ATM uni specification,

[00:09:24] Speaker 01:
That is what is making the access result decision, the alleged access result decision.

[00:09:33] Speaker 01:
And what is happening is that packets with the exact same VCVP headers, that is, whether they have the exact same match, those packets will be treated differently in the cell policer, some being discarded and some not, based on a traffic contract.

[00:09:52] Speaker 01:
that has nothing to do with whether or not there is a VC-VP match.

[00:09:57] Speaker 01:
To put it differently, what's happening here, just to give an example, it's as if I go to a movie theater, I buy a ticket, they look at the ticket, they see it's a match for the theater, I go into the theater, halfway through the theater, I start talking, they ask me to leave.

[00:10:15] Speaker 01:
The reason I'm being asked to leave

[00:10:18] Speaker 01:
is not in response to me buying a ticket and the ticket being a match for me to be in the theater.

[00:10:24] Speaker 01:
The reason I'm being asked to leave is because I'm talking.

[00:10:28] Speaker 00:
And I guess the other side or maybe the board's rationale is you never get filtered out of the movie theater.

[00:10:37] Speaker 00:
unless you first buy a ticket to get into the movie theater.

[00:10:42] Speaker 00:
And so, in a larger sense, maybe not in the specific sense that you're talking about when you're talking in the theater, but in a larger sense, you are getting removed from the theater in a way that's dependent on you getting into the theater in the first place.

[00:11:02] Speaker 00:
Well, Judge Chen.

[00:11:03] Speaker 00:
Is that kind of what the board is trying to say?

[00:11:06] Speaker 01:
I think that is the way that the board is approaching this.

[00:11:10] Speaker 01:
And to be sure, buying the ticket and being admitted to the theater necessarily precedes the decision.

[00:11:18] Speaker 01:
But the decision that is then being made, it does not turn on.

[00:11:23] Speaker 01:
And yes, I wouldn't be there if I hadn't bought the ticket.

[00:11:27] Speaker 01:
But the decision being made to kick me out of the theater

[00:11:31] Speaker 01:
It's just like what the cell policer is doing.

[00:11:33] Speaker 01:
It is making a decision that in no way, shape, or form turns on the match.

[00:11:37] Speaker 01:
And again, the best way to understand that is that you can have packets that would have an absolutely identical match, and they are being treated differently for reasons that have nothing to do with the match.

[00:11:49] Speaker 01:
And that is entirely what the decision is.

[00:11:51] Speaker 01:
And that feeds into my second argument, which is that the alleged access control pattern here, the VCVP heading,

[00:12:01] Speaker 01:
it's not being used for access control.

[00:12:05] Speaker 01:
And in fact, Aristo's response on appeal isn't to defend the board's rationale other than to say that the term access control pattern encompasses, quote, any string of bits stored in the CAM.

[00:12:17] Speaker 01:
That's at page 27 of the yellow brief.

[00:12:19] Speaker 01:
That makes that limitation absolutely meaningless.

[00:12:22] Speaker 01:
And that is exactly what we argued below.

[00:12:24] Speaker 01:
It could not be.

[00:12:27] Speaker 01:
to 1570.

[00:12:28] Speaker 01:
That's like reading the thumb out of thumb switch in Enri man-machine interface.

[00:12:34] Speaker 00:
I guess the idea is these access control patterns, they're basically codes that you match up with whatever the packet label has.

[00:12:44] Speaker 01:
That's exactly right.

[00:12:45] Speaker 01:
I mean, it is the priority here.

[00:12:48] Speaker 00:
And if you do have a match, then a few other steps happen before you decide whether

[00:12:57] Speaker 00:
that packet gets filtered out?

[00:12:58] Speaker 01:
As opposed to the fact that you have the match or not being what decides if the packet's being filtered out.

[00:13:04] Speaker 01:
And that's the problem.

[00:13:06] Speaker 01:
What the patent contemplates, and what's very clear is the decision, and if you look at column two, lines 33 to 35, the match is from the information in the packet header.

[00:13:18] Speaker 01:
That's what's used to enforce access control.

[00:13:21] Speaker 01:
If there's a match

[00:13:23] Speaker 01:
And it can be set up either way.

[00:13:24] Speaker 01:
A match could be to include, a match to exclude.

[00:13:26] Speaker 01:
But let's just use it for inclusion for a moment.

[00:13:29] Speaker 01:
If there is a match, then you should be admitted.

[00:13:32] Speaker 01:
You shouldn't be discarded.

[00:13:34] Speaker 01:
End of decision.

[00:13:35] Speaker 01:
But that's not what happens in the asserted prior art.

[00:13:38] Speaker 01:
You have a match, and then there is a separate discard decision that is in no way, shape, or form related to and doesn't turn on what the match decision was.

[00:13:48] Speaker 01:
That's the fundamental problem here.

[00:13:49] Speaker 01:
Now again, there is a mismatch in trying to use

[00:13:53] Speaker 01:
this ATM technology in order to try to read on this patent, and I think that's part of why you end up with the mismatch here, because the VCVP header isn't the type of thing that's being used for access control.

[00:14:05] Speaker 01:
It is literally just being used for a traditional type of routing decision, and if the header, if there is an inconsistency of the header, it doesn't know where to go.

[00:14:14] Speaker 01:
going on well only in the sense that the system if you don't match the VC VP heading the system doesn't know what to do with you it doesn't know how to route you and so therefore it gets before it gets discarded because you're not on the virtual path that you're supposed to be on that's that's what's going on but that's not what they are relying on for the access decision the access decision they rely on the cell policer

[00:14:38] Speaker 01:
The cell police relies on a traffic contract, i.e., have you exceeded your bandwidth?

[00:14:43] Speaker 01:
Have you sent too much data?

[00:14:45] Speaker 01:
And so the exact same match gets treated differently.

[00:14:49] Speaker 01:
So for those reasons, the court should absolutely reverse and remand this issue, which was squarely teed up before the board, again, if you look at appendix 3722.

[00:15:00] Speaker 01:
Now, our final argument on this is that the board's failure to require the relevant steps of the 577 patent

[00:15:07] Speaker 01:
to be conducted in hardware cannot be reconciled with the patent itself.

[00:15:11] Speaker 01:
This is analogous to the decision this court had before it in Vernetics.

[00:15:15] Speaker 01:
You have the title, the abstract, you have every embodiment that requires, quote, hardware processing of ACLs and hardware enforcement of access control.

[00:15:25] Speaker 01:
And you have repeated disparagement of any use of software for these features.

[00:15:30] Speaker 01:
And those are exactly the kind of indices that this court in- Any use or just the matching?

[00:15:35] Speaker 01:
Well, actually, there's an important point there, Judge Chen.

[00:15:39] Speaker 01:
The argument that Arista made below was that the only thing that had to be done in hardware was the maintaining, not even the matching.

[00:15:49] Speaker 01:
And the board relied on that as well.

[00:15:51] Speaker 01:
That's at Appendix 25.

[00:15:53] Speaker 01:
You can see that.

[00:15:54] Speaker 01:
And Arista's argument below was at 23.95.

[00:15:57] Speaker 01:
Now, on appeal, they've conceded that, well, the matching should be done in hardware.

[00:16:02] Speaker 01:
I think that concession alone should result in a remand here.

[00:16:06] Speaker 01:
That's at their yellow brief at page 16, because of course the claims don't specifically refer to the matching in hardware any more than any of the other steps, yet they recognize implicitly that it would make no sense and the invention wouldn't work if the matching wasn't being done in the hardware.

[00:16:26] Speaker 01:
But that was a point that they contested below and that the board

[00:16:30] Speaker 01:
went a different way on below.

[00:16:33] Speaker 01:
And that alone should give you a remand.

[00:16:35] Speaker 01:
I think the other point here is that the patent specifically and repeatedly disparages any use of software.

[00:16:41] Speaker 01:
And that's both for the processing and the enforcement.

[00:16:44] Speaker 01:
And if you look at column two in particular, there are three different passages, one at line three, one at line 14, and one at line 47.

[00:16:54] Speaker 01:
And each one of those talks about

[00:16:56] Speaker 01:
the prior art still having the drawback of using software or requiring at least some software enforcement, and that the invention permits or denies access without the need for software processing.

[00:17:08] Speaker 01:
And these kinds of repeated distinctions and disparagement are exactly what this court in pacing technologies and recently in Honeywell made clear.

[00:17:18] Speaker 00:
You're eating into your rebuttal.

[00:17:19] Speaker 00:
Do you want to say something about claim two?

[00:17:21] Speaker 00:
Well, Judge Chen.

[00:17:22] Speaker 00:
Why isn't the discussion of real time

[00:17:25] Speaker 00:
Telling us that there are packets being operated on in different places in the pipeline.

[00:17:29] Speaker 01:
Sure.

[00:17:30] Speaker 01:
A couple of points.

[00:17:31] Speaker 01:
I mean, this is entirely a substantial evidence issue.

[00:17:34] Speaker 01:
And the arguments that they make about their expert are points where the board actually rejected the factual underpinnings of Arista's expert evidence.

[00:17:44] Speaker 01:
And the board addresses this at appendix 32 to 33, rejecting the argument that column 8 discloses parallel actions.

[00:17:52] Speaker 01:
Given that I'm into my rebuttal time and given that the issues that I'm arguing would supersede and address claim to as well, I think I'd rather save the rest of my time for rebuttal.

[00:18:02] Speaker 01:
Thank you.

[00:18:33] Speaker 03:
I'd like to begin with the hardware issue, if I may.

[00:18:37] Speaker 03:
The hardware issue is their position now is that all steps of the method and all parts of the apparatus must be performed in hardware.

[00:18:46] Speaker 03:
That was not their position below.

[00:18:48] Speaker 03:
It is also not what the claim says.

[00:18:50] Speaker 03:
The claim specifically says that one part has to be done in hardware.

[00:18:54] Speaker 03:
That's that maintaining and matching step.

[00:18:56] Speaker 03:
That's the cam.

[00:18:57] Speaker 03:
That's where the ACLs look up and matches are done.

[00:19:01] Speaker 03:
That is in hardware in the claim.

[00:19:03] Speaker 03:
Nowhere does the claim say anything else has to be in hardware, and that omission I propose should be dispositive, where the claim says X is in hardware but is silent.

[00:19:12] Speaker 00:
That's never the end of a claim construction exercise.

[00:19:15] Speaker 03:
It never is the end of the exercise.

[00:19:17] Speaker 03:
It ought to be the ultimate end of the answer, in my view.

[00:19:19] Speaker 03:
But obviously you have to look at the specifications, and here they're arguing for a specification disclaimer, as they must.

[00:19:28] Speaker 00:
doing this processing in hardware.

[00:19:30] Speaker 00:
The abstract leads with we're going to do hardware processing.

[00:19:33] Speaker 03:
There is no question that a central purpose of this claim was that you're doing that matching step in hardware.

[00:19:39] Speaker 03:
That is the whole point of the invention.

[00:19:41] Speaker 03:
That's what they said was their invention.

[00:19:43] Speaker 03:
It was not that the entire process is done in hardware.

[00:19:46] Speaker 03:
The whole point when you look at every step where they're criticizing software, and if you look at, he cites columns 3, 14, and 47, lines 3, 14, and 47 in column 2,

[00:19:56] Speaker 03:
Every single one of those is talking about the ACL lookup matching process, ACL processing.

[00:20:02] Speaker 03:
That's the CAM.

[00:20:04] Speaker 03:
So there is no question at all that the central focus of this invention was on doing that step in hardware, because the CAM allows you to match up thousands of entries quickly, where software does not.

[00:20:17] Speaker 03:
But a step of taking the output of that and making a go no go decision, that's not the same thing as the very complicated nature that a CAM is required for.

[00:20:26] Speaker 03:
And the specification nowhere says so.

[00:20:29] Speaker 03:
And so the specification disclaimer argument A was not made below.

[00:20:34] Speaker 03:
Let's just be very clear about that.

[00:20:36] Speaker 03:
The board, in fact, said so.

[00:20:37] Speaker 03:
It was not made below.

[00:20:39] Speaker 03:
Now, they're trying to make it here.

[00:20:41] Speaker 03:
It has a very, very high burden.

[00:20:43] Speaker 03:
If you allow them to make it here, the question is, does the specification read as a whole amount to such a disclaimer?

[00:20:52] Speaker 03:
Every time they're criticizing software, it's about the ACL matching step.

[00:20:56] Speaker 03:
the ACL processing step.

[00:20:58] Speaker 03:
That's the part they're criticizing.

[00:20:59] Speaker 03:
And that's the part they put in the claim.

[00:21:02] Speaker 03:
So those are consistent.

[00:21:03] Speaker 03:
They do not support a broader disclaimer that other steps also have to be done in hardware.

[00:21:09] Speaker 03:
And if there's any question about this, what they're trying to say is there are portions of the specification that say we can eliminate the need for software processing.

[00:21:18] Speaker 03:
Undeniably true.

[00:21:20] Speaker 03:
Software processing of those ACLs in step two, which is matching.

[00:21:25] Speaker 03:
If there's any question that they're wrong, that that amounts to a disclaimer about all steps, that is answered by column five, lines one to six.

[00:21:34] Speaker 03:
Column five, lines one to six, expressly provides for software processing in step three, where you're doing the access result.

[00:21:43] Speaker 03:
And it's set up by the bottom of column four.

[00:21:45] Speaker 03:
And the bottom of column four says, in some cases, you may want to have additional software processing.

[00:21:51] Speaker 03:
And it says, the higher level processor

[00:21:54] Speaker 03:
specifies whether the packet 130 after processing by the higher level processor is forwarded to a selected output interface or is dropped.

[00:22:02] Speaker 03:
That's the excess result in step three.

[00:22:05] Speaker 03:
So the patent certainly cannot clearly support patents.

[00:22:10] Speaker 00:
Can we get to the more mind-bending question of whether Huey teaches the... In response to?

[00:22:18] Speaker 03:
I'm happy to.

[00:22:19] Speaker 03:
So in response to, the issue really then comes down to

[00:22:23] Speaker 03:
does in response to require that it be specifically generating the result.

[00:22:34] Speaker 03:
And that's not what the claim says.

[00:22:35] Speaker 03:
The claim has, and the board nailed this one with really specific reasoning.

[00:22:41] Speaker 03:
They made exactly the same argument below that they're making here, that it had to be, that the matching, that the

[00:22:47] Speaker 03:
address or the pattern had to be used as part of that process.

[00:22:52] Speaker 00:
The QE's policer doesn't care what the VP address is, right?

[00:22:57] Speaker 00:
When it makes a choice to filter out various packets, it does not care.

[00:23:03] Speaker 03:
It does not care.

[00:23:04] Speaker 03:
But that is also true of the patent.

[00:23:06] Speaker 03:
If you look at the bottom of column four, what is sent to their policer, if you will, is not the IP address, which is their analog.

[00:23:16] Speaker 03:
What is sent is, quote, an indicator.

[00:23:19] Speaker 03:
So it's not the address.

[00:23:21] Speaker 03:
They're sending exactly the same thing that Huey is sending, something that is the result of the match, but not the match itself.

[00:23:28] Speaker 03:
The argument that is being made to you here is that Huey doesn't satisfy the claim because the VPVCI address does not then be forwarded to the policer who uses that address to make a decision based on the content of that address.

[00:23:43] Speaker 03:
That's the argument that's being made.

[00:23:44] Speaker 03:
That is not what the claim says.

[00:23:46] Speaker 03:
If that's what they wanted to say, it was easy to write that.

[00:23:49] Speaker 03:
They chose broader language, and they're stuck with that broader language.

[00:23:53] Speaker 03:
And the broader language is merely in response to.

[00:23:55] Speaker 00:
And their argument... Let me get this straight.

[00:23:58] Speaker 00:
The claim, they find a match, or multiple matches.

[00:24:02] Speaker 00:
Yes.

[00:24:03] Speaker 00:
And then each match has some priority value attached to it.

[00:24:09] Speaker 00:
If we're talking about the patent, yes.

[00:24:11] Speaker 00:
Yes, about the patent claim.

[00:24:13] Speaker 00:
And so based on whatever priority value that is, which is representative, I assume, of what the access control pattern is, then you make the access result decision.

[00:24:29] Speaker 03:
In response to that selective match.

[00:24:31] Speaker 00:
So I guess you could argue there's a stronger connection between that access result

[00:24:39] Speaker 00:
and the identity of the access control pattern in the can compared to Huey's policer and the choices it's making when it is filtering out data packets.

[00:24:52] Speaker 00:
Is that fair to say?

[00:24:53] Speaker 03:
I don't think it's fair in the light of column four, because what column four tells you is you're not sending that address.

[00:24:58] Speaker 00:
You're not sending the address, but you're sending something else that's representative of the address.

[00:25:03] Speaker 03:
True.

[00:25:04] Speaker 00:
It's an indicator.

[00:25:05] Speaker 03:
It's an indicator.

[00:25:06] Speaker 03:
It's an indicator, and it says what the indicator

[00:25:09] Speaker 03:
The indicator has either specifies either the packet is forwarded or the packet is dropped or the packet is forwarded to the processor for further processing.

[00:25:19] Speaker 03:
That's, it doesn't send the address.

[00:25:20] Speaker 03:
And their argument to you here was that address has to be used.

[00:25:25] Speaker 03:
That's the argument they made to the board.

[00:25:28] Speaker 03:
And that, that address has to be used in step three as a, to generate the result.

[00:25:34] Speaker 03:
You look up the address, you say, ah, this result means X. That's not what column four says.

[00:25:40] Speaker 03:
that is squarely inconsistent with column four.

[00:25:43] Speaker 03:
And Huey is very consistent with column four.

[00:25:47] Speaker 03:
The address itself is not sent.

[00:25:49] Speaker 03:
And in response to is clearly true.

[00:25:51] Speaker 03:
It is in response to the fact of the match.

[00:25:54] Speaker 03:
And the board's reasoning here I think is perfect.

[00:25:56] Speaker 03:
If that match had not been made, that decision would not be made.

[00:26:01] Speaker 03:
It is clearly relevant in a but four sense.

[00:26:04] Speaker 03:
But it's also not just a random but four sense.

[00:26:08] Speaker 03:
It's its purposeful architecture.

[00:26:10] Speaker 03:
That match determines whether it's sent to the policer or not.

[00:26:14] Speaker 03:
That's the purpose of sending that match, is to allow the policer then to make the decision, which is exactly the same purpose in the claim.

[00:26:22] Speaker 00:
What factors does the policer use to decide which packets it's going to filter out?

[00:26:28] Speaker 03:
In Huey?

[00:26:29] Speaker 03:
In Huey.

[00:26:31] Speaker 03:
In Huey, it's the basis of what information is sent in its indicator, if you will.

[00:26:40] Speaker 00:
I don't understand.

[00:26:42] Speaker 03:
What's sent is not the address.

[00:26:43] Speaker 03:
What's sent is an indicator as a result of the match.

[00:26:47] Speaker 00:
Representatives of a whole bunch of data packets.

[00:26:49] Speaker 03:
Representative of various data packets.

[00:26:52] Speaker 03:
And how Huey makes that decision of which ones to drop or send, that I don't know.

[00:26:57] Speaker 03:
But I don't think that is relevant to the question.

[00:26:59] Speaker 03:
The question is, does it make that decision in response to the fact that it's sent to it?

[00:27:05] Speaker 03:
The whole point, it's not just

[00:27:08] Speaker 03:
It's not a random effect.

[00:27:10] Speaker 03:
It's not an accidental effect.

[00:27:12] Speaker 03:
It's a designed architected effect.

[00:27:15] Speaker 03:
The purpose of that matching step is to send it to the Huey cell policer for the purpose of making that decision.

[00:27:23] Speaker 03:
That is its purpose.

[00:27:26] Speaker 03:
So it's not, as I say, accidental and it's not unconnected.

[00:27:31] Speaker 03:
That's the structural design.

[00:27:33] Speaker 03:
And I think that's important to understand because

[00:27:37] Speaker 03:
That takes it away from their argument that it's just a happenstance.

[00:27:41] Speaker 03:
It has no connection to it.

[00:27:43] Speaker 03:
That's clearly false under Huey.

[00:27:46] Speaker 03:
So with regard to, I do want to make one additional point, that there is substantial evidence to support the board's finding, even under Cisco's new construction.

[00:27:56] Speaker 03:
And that's because there's no dispute that Huey teaches that a cam performs maintaining a matching step.

[00:28:02] Speaker 03:
So we're done with that.

[00:28:04] Speaker 03:
So the question is really to sell Policer.

[00:28:06] Speaker 03:
And the question is, is the cell police or hardware or not?

[00:28:12] Speaker 03:
So this goes to the hardware issue.

[00:28:14] Speaker 03:
Huey teaches as part of the port.

[00:28:16] Speaker 03:
That's hardware.

[00:28:18] Speaker 03:
Both experts say we can't tell whether it would be done in hardware or software from Huey itself.

[00:28:23] Speaker 03:
It could be either.

[00:28:24] Speaker 03:
Both experts acknowledge that.

[00:28:27] Speaker 03:
And there were specific teachings in timely prior IEEE articles that we cited to the board saying,

[00:28:36] Speaker 03:
you would want to do hardware because it's fast, cheap, and simple.

[00:28:41] Speaker 03:
Exactly the reasons they said hardware is better.

[00:28:44] Speaker 03:
That was known in 1991 in these IEEE articles.

[00:28:47] Speaker 03:
So the only thing that's missing from Huey is a specific disclosure of whether that process of the cell polisher is done in hardware software with the implication that it's part of hardware because it's part of the port.

[00:28:58] Speaker 03:
IEEE clearly teaches that one skill of the art would have, it would have been obvious to them to do it in hardware.

[00:29:04] Speaker 03:
The board did not say it because it disagreed with the hardware limitation.

[00:29:07] Speaker 03:
And at the time, they weren't even making this same hardware limitation.

[00:29:11] Speaker 03:
What they were arguing about the cell policer was that it wasn't done in the camp.

[00:29:15] Speaker 03:
They weren't arguing that the cell policer had to be done as part of hardware, separately.

[00:29:20] Speaker 03:
That's a new argument here.

[00:29:23] Speaker 03:
The argument there was quite different.

[00:29:24] Speaker 03:
So it's not one of these hidden claim construction issues.

[00:29:27] Speaker 03:
It's a new argument.

[00:29:30] Speaker 02:
Okay, we're essentially out of time, thank you.

[00:29:33] Speaker 02:
You have one other point to make?

[00:29:35] Speaker 03:
I want to make one quick point on Asinore-Stoppel, if I may.

[00:29:38] Speaker 03:
It has nothing to do with what I've argued before.

[00:29:40] Speaker 03:
It has to do with the factual predicates of Asinore-Stoppel, which is an argument that my friend made with respect to this patent.

[00:29:48] Speaker 03:
Asinore-Stoppel has significant factual predicates that this record is lacking.

[00:29:53] Speaker 03:
The specific role that the Asinores had in designing the accused instrumentalities.

[00:29:59] Speaker 03:
That evidence is scant.

[00:30:01] Speaker 03:
In fact, that evidence is limited to Dr. Cheriton had substantial involvement in software.

[00:30:06] Speaker 03:
Well, that's an initial matter.

[00:30:08] Speaker 03:
That's inconsistent with their hardware software argument.

[00:30:10] Speaker 03:
But it has nothing to do with the accused instrumentality, specifically.

[00:30:14] Speaker 03:
So they have failed miserably in establishing the record that would be required to support a sign or estoppel.

[00:30:20] Speaker 03:
And as your honors know, we have, and I'll leave that for the briefs, we believe that a sign or estoppel should not be applicable, that this court should shit on Bonk in order to

[00:30:29] Speaker 03:
say that a sign or stopper is no longer applicable post-lear.

[00:30:33] Speaker 03:
And in the facts of this case, even under this court's current law, a sign or stopper should not be applicable.

[00:30:39] Speaker 03:
But the record is not built out on that question.

[00:30:40] Speaker 03:
Thank you.

[00:30:41] Speaker 03:
Thank you.

[00:31:00] Speaker 01:
Thank you, Chief Judge Prost.

[00:31:01] Speaker 01:
On the last point that my colleague was making, just to be clear, if you look at this court's decision in mentor graphics at 50 F, 3rd, 1376, it rejects the notion that the assigner had to be involved in developing the allegedly infringing features in order for assigner estoppel to apply.

[00:31:18] Speaker 01:
And there's no currency to that argument.

[00:31:20] Speaker 01:
Now, turning to the hardware arguments,

[00:31:25] Speaker 01:
First, there are no alternative grounds for affirmance that could be done here.

[00:31:28] Speaker 01:
The issues that my colleague was talking about were disputed at appendix 1667 to 1670.

[00:31:34] Speaker 01:
Our expert addressed that issue at appendix 2508.

[00:31:37] Speaker 01:
Now, addressing the hardware issues that Mr. Power started with, first, there's no waiver here.

[00:31:45] Speaker 01:
I mean, let's be clear.

[00:31:46] Speaker 01:
In the reply below, they understood what our argument was.

[00:31:50] Speaker 01:
If you look at appendix 2395, they argued, quote, contrary to Cisco's contention,

[00:31:55] Speaker 01:
Not every step of the challenge claims must be performed in hardware rather than software.

[00:32:00] Speaker 01:
That argument was clearly teed up, and the board understood it at Appendix 24.

[00:32:04] Speaker 01:
It said, we argued that, quote, the claims are limited to performing access control in hardware versus software.

[00:32:11] Speaker 01:
Second, so much of the argument today that Mr. Powers has made is about the matching step taking place in hardware.

[00:32:18] Speaker 01:
Well, that's not what they argued below.

[00:32:20] Speaker 01:
Below, they argued that the only thing that had to be in hardware was the maintaining step.

[00:32:24] Speaker 01:
And I'd encourage you to look at appendix 2395, where they said, quote, the only hardware element required for any step is the associative memory that is used for maintaining a set of access control patterns.

[00:32:36] Speaker 01:
The board relied on this.

[00:32:37] Speaker 01:
And if you look at appendix 25, the board didn't require that the matching take place in hardware.

[00:32:42] Speaker 01:
What we're hearing today, and rightly so, that at least the matching has to take place in hardware,

[00:32:47] Speaker 01:
And given that, that's a reason at a minimum to remand.

[00:32:51] Speaker 01:
Now, in terms of new arguments, we heard a very new argument today based on column five of the patent.

[00:32:57] Speaker 01:
That's not an argument that Arista made in its brief.

[00:33:00] Speaker 01:
The column five discloses the processing take place in software.

[00:33:04] Speaker 01:
And that's not what's going on.

[00:33:06] Speaker 01:
If you look at the passage at the top of column five, this is talking about additional processing that could be done after

[00:33:13] Speaker 01:
after the elements of the claim are satisfied.

[00:33:17] Speaker 01:
This is saying there are more things that you could potentially do that are being disclosed here.

[00:33:21] Speaker 01:
Nothing that's here at the top of column five goes to the claimed elements themselves.

[00:33:28] Speaker 01:
If I may, Chief Judge, press one last point.

[00:33:30] Speaker 01:
And my last point is in response to, to be clear, Huey's policer doesn't care whether or not there was a match.

[00:33:40] Speaker 01:
In the patent, the access result depends on the fact that there has been a match.

[00:33:45] Speaker 01:
And the fact there's been a match is what tells you what to do.

[00:33:49] Speaker 01:
That is not what Huey Cell Policer does.

[00:33:52] Speaker 01:
It doesn't take its action based on the fact that there's been a match.

[00:33:57] Speaker 01:
Thank you.

[00:33:57] Speaker 01:
Thank you, Chief Judge Price.

[00:33:58] Speaker 02:
The case is submitted.