[00:00:00] Speaker 01:
Intellectual Adventures versus T-Mobile.

[00:00:44] Speaker 01:
Okay, please proceed Mr. Black.

[00:00:57] Speaker 05:
Thank you your honor.

[00:00:58] Speaker 05:
May it please the court.

[00:00:59] Speaker 05:
The district court's grant of summary judgment rests on two issues of claim construction.

[00:01:04] Speaker 05:
The construction of application aware resource allocator and the court's construction of the means plus function element of claim 20.

[00:01:12] Speaker 05:
including the term optimized, which the court found to be indefinite.

[00:01:15] Speaker 05:
Because the district court erred in its claim construction, the judgment should be reversed.

[00:01:20] Speaker 05:
The application where a resource allocator is defined in the specification.

[00:01:25] Speaker 01:
I'm just trying to figure out what the issue is, really.

[00:01:29] Speaker 01:
I'm not landing any place.

[00:01:31] Speaker 01:
The difference between your view and what the district court decided was he seemed to say there were two separate things going on.

[00:01:40] Speaker 01:
There's the application.

[00:01:41] Speaker 01:
information on the knowledge of the type of the application, and then information from the application.

[00:01:48] Speaker 01:
And he sort of seemed to say those were two different steps.

[00:01:53] Speaker 01:
Is that the dispute?

[00:01:54] Speaker 01:
I'm just trying to really understand what the dispute is here.

[00:01:58] Speaker 01:
What are the two alternatives?

[00:02:00] Speaker 05:
Certainly, Your Honor.

[00:02:00] Speaker 05:
The record is quite confusing here.

[00:02:03] Speaker 05:
And I think the best way to explain it is that the way software layers work,

[00:02:08] Speaker 05:
They don't actually exist in the real world.

[00:02:11] Speaker 05:
They are an abstraction.

[00:02:12] Speaker 05:
And it's a way that computer scientists discuss software layers.

[00:02:16] Speaker 05:
So when you look on your computer and you see an image, what you're seeing is application data, which is being used to display an image.

[00:02:24] Speaker 05:
In wireless systems, there's something down at the bottom of the stack called the MAC layer.

[00:02:30] Speaker 05:
And that is responsible for sending out packets over the airwaves.

[00:02:34] Speaker 05:
It's only concerned about making sure that the packets get where they need to go.

[00:02:39] Speaker 05:
And by design, the MAC layer has no information about layer 7.

[00:02:44] Speaker 05:
It's a bit like if you write a letter, a personal letter, and you put it in an envelope, and then you send it in, you put it in a FedEx package, and it goes to the recipient.

[00:02:53] Speaker 05:
The mailman doesn't have to know anything about the contents of the envelope.

[00:02:58] Speaker 05:
The mailman only needs to know what the routing information is on the cover.

[00:03:02] Speaker 05:
And that's a tremendous advantage.

[00:03:04] Speaker 05:
And that provides what's called layer isolation in the software world.

[00:03:09] Speaker 01:
And is that information used?

[00:03:12] Speaker 01:
Is the whole purpose of that information so that the MAC layer, they can determine what bandwidth is necessary for the variety?

[00:03:18] Speaker 05:
The MAC layer's sole job, yes.

[00:03:20] Speaker 05:
In the ordinary system, the MAC layer's sole job is to just determine how the packets get modulated and sent over the airways.

[00:03:28] Speaker 05:
But the MAC layer has absolutely no idea what kind of packet it's dealing with.

[00:03:32] Speaker 05:
So in the world of yesterday,

[00:03:39] Speaker 05:
The packets could have all types of different information in it, and the MAC layer doesn't know anything about it.

[00:03:43] Speaker 05:
Its sole job is to deliver the message.

[00:03:45] Speaker 05:
It has no idea.

[00:03:46] Speaker 05:
It makes no discrimination.

[00:03:48] Speaker 05:
It's not aware of what the packet contents are, and therefore, it doesn't do anything special.

[00:03:54] Speaker 05:
Now what happened as voice over IP, which puts voice packets into traditional IP packets, and the internet converged, what happened is the telecommunications industry realized that you could put

[00:04:08] Speaker 05:
Both the voice packets and the internet packets together, put them in the same pipe and send them out over the airways.

[00:04:14] Speaker 05:
And that's a tremendous advantage because it allows you to stuff all these packets together and put them in the same bucket or same pipe as they say.

[00:04:23] Speaker 05:
The problem is voice packets have to arrive on time, every 20 milliseconds, or you're going to have poor call quality.

[00:04:31] Speaker 05:
The other 18 milliseconds, they might not be doing anything.

[00:04:33] Speaker 05:
But on the 20th millisecond, the packet has to be there.

[00:04:37] Speaker 05:
Internet packets are different.

[00:04:38] Speaker 05:
They all have to arrive, but they don't even actually have to arrive in order.

[00:04:42] Speaker 05:
They can come out of order and then be put in order at the recipient.

[00:04:45] Speaker 05:
So you have different quality of service requirements.

[00:04:49] Speaker 05:
The voice packets must arrive on time.

[00:04:51] Speaker 05:
The traditional internet packets, not so important.

[00:04:55] Speaker 05:
So what they did, what the invention is about, is saying that the MAC layer, which normally would have no idea what kind of packet it's dealing with, the MAC layer is now aware.

[00:05:05] Speaker 05:
aware in the sense of artificial intelligence.

[00:05:07] Speaker 05:
This is a construct.

[00:05:09] Speaker 05:
But it's aware, has some information when it's making the decisions about what packets to send when over the airwaves.

[00:05:16] Speaker 05:
It has some information about what's happening up here at the application layer.

[00:05:20] Speaker 05:
And it can therefore discriminate among packets and make the system work efficiently.

[00:05:25] Speaker 05:
Now, the application aware resource allocator is a tool.

[00:05:29] Speaker 05:
It's described in the first sentence of the abstract and the first sentence of the specification when it's defined there.

[00:05:35] Speaker 05:
as an application where a resource allocator, where the resource allocator allocates bandwidth resource based on an application type.

[00:05:45] Speaker 01:
Oh, I'm sorry, what are you referring to?

[00:05:46] Speaker 05:
It's in the, it's in the abstract.

[00:05:49] Speaker 05:
It's the first sentence of the abstract and the first sentence of the, of the summary of invention.

[00:05:55] Speaker 05:
That also happens to be the claim construction that we requested.

[00:05:59] Speaker 05:
Now, claim two of the patent refers to

[00:06:03] Speaker 05:
the inputs to the resource problem.

[00:06:06] Speaker 02:
You're really dense in the technology and it's probably helpful, but I don't actually need it right now.

[00:06:13] Speaker 02:
So isn't the real problem, the district courts, I read the district courts construction, I read your entire pattern, I read his construction, I'm like, nothing wrong with it.

[00:06:22] Speaker 02:
It wasn't until I got to the motion to strike when I'm like, uh, time out, it has to come from layer seven.

[00:06:27] Speaker 02:
The information has to come, has to be taken from layer seven.

[00:06:29] Speaker 02:
It can't be a packet header in three or four.

[00:06:31] Speaker 02:
I understand a port is not good enough.

[00:06:33] Speaker 02:
Port's not good enough.

[00:06:34] Speaker 02:
It doesn't give you enough information to identify application type, but packet headers do and can.

[00:06:39] Speaker 02:
Why can't, so, so it's like, it's morphed.

[00:06:43] Speaker 02:
When, when he articulated the construction, I don't have a problem with the words in the construction.

[00:06:49] Speaker 02:
It wasn't until he then applied it in the motion to strike where he suddenly said, oh, no, no, no, you're wrong.

[00:06:56] Speaker 02:
You have to actually take the information in order to allocate bandwidth from the seventh player.

[00:07:06] Speaker 02:
And that was the thing that sort of surprised me because his construction on its face didn't seem problematic.

[00:07:11] Speaker 02:
It wasn't until he effectively modified it or narrowed it in the application vis-a-vis the motion to strike.

[00:07:17] Speaker 02:
And so for me, what it came down to was, is the district court right?

[00:07:22] Speaker 02:
Because despite what he says the construction is in his application, he has clearly ruled out the possibility that the information about application type can come from layer three or four, and that that is sufficient for the QOS to then do its functionality at the MAC layer appropriately.

[00:07:38] Speaker 02:
It doesn't have to actually pull anything from seven itself.

[00:07:42] Speaker 02:
And so that's, it seems to me, the heart

[00:07:45] Speaker 02:
of where, if he erred, he erred was in saying coming from three or four isn't enough, am I right?

[00:07:51] Speaker 02:
Is that right?

[00:07:52] Speaker 02:
Do I understand it right?

[00:07:53] Speaker 05:
That is correct, Your Honor.

[00:07:54] Speaker 05:
We don't believe that the words of the construction that we proposed, which we think are correct, are much different from the actual words that were used by the defendants for their construction.

[00:08:05] Speaker 05:
What was going on here, and it's a bit of a problem in the district courts now, is if an infringer wants to

[00:08:12] Speaker 05:
argue non-infringement, but doesn't have a great argument based on the literal wording of the claim.

[00:08:18] Speaker 05:
Of course, you looked at a prosecution history to see if there's some way that you can shift the construction.

[00:08:24] Speaker 05:
This case was a tough one for the defendants, because there was nothing they could point to in the prosecution history that would get them all the way to non-infringement.

[00:08:31] Speaker 05:
So they were doing what I'm going to call

[00:08:33] Speaker 02:
Well, not only that, I found at least three, and I'm sure more given the 80-some columns in this patent, of examples where the MAC-level QoS requirements are obtained entirely from packet headers found in three and four.

[00:08:50] Speaker 02:
And then what's more, the claims in this case, we're not talking about one dependent claim that would be eviscerated by this.

[00:08:58] Speaker 02:
We're talking about many, right?

[00:09:01] Speaker 02:
Let's now throw out claim 2, claim 11, claim 19.

[00:09:07] Speaker 02:
All of these claims would be

[00:09:11] Speaker 02:
inappropriate under this construction.

[00:09:13] Speaker 02:
None of them would have any meaning.

[00:09:15] Speaker 02:
They would be sort of mooted entirely as though they don't exist.

[00:09:19] Speaker 05:
I agree with you completely, Your Honor.

[00:09:20] Speaker 02:
And the examiner who allowed all these claims is the same examiner that participated in this prosecution.

[00:09:25] Speaker 02:
So obviously, that examiner did not have the impression that you had given away all of that, or he wouldn't have then allowed you to have all these defendant claims.

[00:09:35] Speaker 05:
That's our position.

[00:09:36] Speaker 05:
And the last page of the relevant page of the prosecution history, which is at A3500,

[00:09:42] Speaker 05:
just before the notice of allowance has repeated references to IP packet headers, packet headers.

[00:09:49] Speaker 01:
Can I just take you back though to where I started, which is you were getting there, but taking quite a while, which is I'm just trying to understand.

[00:09:57] Speaker 01:
I don't want to hear global issues about district courts doing what

[00:10:02] Speaker 01:
lawyers do.

[00:10:03] Speaker 01:
I just want to know what the district court did here.

[00:10:06] Speaker 01:
And you're right.

[00:10:07] Speaker 01:
I understand, I think, one, if you could articulate, but I think I understand the difference between his original construction and the one on reconsideration.

[00:10:15] Speaker 01:
But one, articulate what the difference was.

[00:10:18] Speaker 01:
And two, tell us what led him there, which I assume is what you think is a misreading of the prosecution history.

[00:10:24] Speaker 01:
But if you could just articulate that in English, it would be helpful.

[00:10:26] Speaker 05:
Yes.

[00:10:27] Speaker 05:
I think that the resource allocator

[00:10:31] Speaker 05:
allocates the resources.

[00:10:33] Speaker 05:
That's all the claim one requires.

[00:10:35] Speaker 05:
Claim two says, what does the resource allocator rely on?

[00:10:40] Speaker 05:
And the answer is packet headers or a direct communication from the software layer.

[00:10:46] Speaker 05:
So there's two basic types of embodiments.

[00:10:49] Speaker 05:
In one embodiment, there's a layer seven communication all the way up here that goes all the way down to layer two.

[00:10:55] Speaker 01:
And that direct communication between 7 and that layer was baked into his original construction, to the original construction, right?

[00:11:03] Speaker 05:
Well, baked in.

[00:11:04] Speaker 05:
It certainly was included.

[00:11:05] Speaker 05:
We don't deny that that's an embodiment.

[00:11:07] Speaker 05:
But he excluded many other embodiments of the patent, including reviewing and allocating based on the contents of packet headers at other levels.

[00:11:17] Speaker 05:
And what the patent does is it says... You're talking about the second construction.

[00:11:22] Speaker 05:
It's very hard here because he didn't give us much of an opinion on this.

[00:11:26] Speaker 05:
But the bottom line is that there are two basic ways to get the information to the resource allocator.

[00:11:31] Speaker 05:
A direct communication from layer seven and effectively an indirect communication where the allocator infers what's happening at the application layer based on reading, say, layer three or four or other packet headers.

[00:11:43] Speaker 05:
Both of them get the information to the allocator.

[00:11:46] Speaker 05:
Both of them are covered extensively in the patent.

[00:11:49] Speaker 05:
It appears that what the district court did is limit us to the direct from layer seven embodiment, reading out a preferred embodiment that's clearly claimed.

[00:11:58] Speaker 05:
And he did so really based solely on footnote seven of his opinion, which refers to two office actions, one of which is the one I referred to before at page A 3500.

[00:12:10] Speaker 05:
Clearly stated.

[00:12:11] Speaker 02:
You're taking too long to answer the Chief's question.

[00:12:14] Speaker 02:
Shouldn't it be, yeah, your best guess is he found disclaimer in the prosecution history.

[00:12:20] Speaker 02:
You've got nothing.

[00:12:21] Speaker 02:
You don't have a lot to go on, but that's what you think he did.

[00:12:23] Speaker 02:
And here are the pages they cite and why they don't apply.

[00:12:26] Speaker 02:
So how about you take that structure?

[00:12:28] Speaker 05:
Certainly, Your Honor.

[00:12:29] Speaker 05:
They cited to the district court at a footnote, and he cited to two office actions.

[00:12:37] Speaker 05:
He appears to have relied the language that they asked for came out of an office action of April 2nd, 2003, and the language that they asked for says that the allocator should be able to take into account when allocating bandwidth information about applications at OSI's OSI application layer 7.

[00:12:58] Speaker 05:
And we read that and we said, okay, it says information.

[00:13:01] Speaker 02:
So you're on page 3499 and 3500 of the appendix, correct?

[00:13:04] Speaker 05:
That is correct.

[00:13:04] Speaker 05:
That is correct.

[00:13:05] Speaker 05:
And at 3499, the defendant said, well, we've got to read into the claim language information about applications at Layer 7.

[00:13:13] Speaker 05:
And we said, well, we don't have a problem with that.

[00:13:16] Speaker 05:
You're going to use the information about applications at Layer 7.

[00:13:20] Speaker 05:
But we said it's not just directly receiving the information.

[00:13:24] Speaker 05:
The allocator has the information.

[00:13:25] Speaker 05:
Where it gets it from is irrelevant.

[00:13:27] Speaker 05:
On page 3500, the next page, there are repeated references to

[00:13:33] Speaker 05:
column 18, column 53, and column 63 of the patent.

[00:13:37] Speaker 02:
In fact, on page 3500, doesn't that's the paragraph, the relevant paragraph to begin with, in the present invention?

[00:13:43] Speaker 02:
Yes.

[00:13:44] Speaker 02:
And then it goes through and talks extensively about how packet headers are one of the examples of where this exact information can come from, and aren't those packet headers contained in layer three and four?

[00:13:58] Speaker 05:
Yes, Your Honor.

[00:13:59] Speaker 05:
So let me just

[00:14:01] Speaker 05:
amplify on that a bit.

[00:14:02] Speaker 05:
The key page is 3500, and it says, in the present invention, words that have hung many a patentee, but here actually explain clearly what is met.

[00:14:14] Speaker 05:
In the present invention, scarce wireless bandwidth can be dynamically allocated using the present invention to provide, et cetera.

[00:14:22] Speaker 05:
And then if you look down a little bit further, it says differentiation, and that's differentiating between the types of packets, quote,

[00:14:29] Speaker 05:
is done on the basis of some identifiable information contained in packet content, such as EG packet headers, site to column 18 of the spec.

[00:14:39] Speaker 05:
One exemplary method includes analyzing several items in an IP packet header or payload.

[00:14:45] Speaker 05:
IP packet header or payload, that's layer three.

[00:14:47] Speaker 05:
There's no ambiguity about it.

[00:14:49] Speaker 05:
The next thing that happened is a notice of allowance.

[00:14:52] Speaker 05:
I also want to refer to the Court's optimized construction, but I'm almost out of time.

[00:14:58] Speaker 05:
The Court

[00:14:58] Speaker 05:
should have been looking at optimization of a software application, which is something that experts understand and could have provided an infringement opinion on.

[00:15:08] Speaker 05:
The court did not look at the specification.

[00:15:10] Speaker 05:
He simply concluded.

[00:15:11] Speaker 01:
What is it in the specification that clarifies or tells us what the parameters are of optimization?

[00:15:18] Speaker 05:
So the optimization here is, in effect, customization based on the type of software application.

[00:15:24] Speaker 05:
There are numerous references throughout the specification about

[00:15:28] Speaker 05:
for instance, prioritizing voice over internet packets.

[00:15:32] Speaker 05:
That's optimizing software applications in the system when the prior art was complete inability to differentiate.

[00:15:47] Speaker ?:
Thank you.

[00:15:54] Speaker 04:
May it please the court?

[00:15:56] Speaker 04:
My name is Doug Cabale.

[00:15:57] Speaker 04:
I represent the defendants.

[00:15:59] Speaker 02:
Why don't we, I mean, he didn't so much time on it, but why don't you start on the optimization?

[00:16:03] Speaker 02:
I am, it seems like an awfully amorphous word, and it looks to me to be akin to the aesthetically pleasing language in datamize versus plum tree, but I'm, why don't you start there?

[00:16:16] Speaker 02:
I know he didn't get to spend much time on it, but if you don't mind, start with a response to what he did say about indefiniteness.

[00:16:24] Speaker 04:
Yes, Your Honor.

[00:16:25] Speaker 04:
You're correct that this is very similar to the datamized case.

[00:16:28] Speaker 04:
It's similar to the interval licensing case.

[00:16:31] Speaker 04:
We've got a term optimized that the specification describes in terms of the specification and the claims described in terms of quality of service.

[00:16:40] Speaker 02:
You're optimizing.

[00:16:40] Speaker 02:
If we agree that this is indefinite, does that end the case or is the claim construction issue directed to other claims?

[00:16:48] Speaker 04:
The optimized term is in claim 20, but it's not in claim 1.

[00:16:52] Speaker 04:
There is a... So we have to decide both issues?

[00:16:55] Speaker 04:
We have to decide both.

[00:16:56] Speaker 04:
Okay, so the patent tells us that this quality of service, this thing that's being optimized, is a relative term.

[00:17:02] Speaker 04:
It has different meanings to different users.

[00:17:05] Speaker 01:
It can be... Why don't you cite us the line that didn't specify the difference?

[00:17:08] Speaker 04:
Yes, that's column 12, line 51.

[00:17:15] Speaker 04:
Then on column 12, line 62, it explains that perhaps it's best to understand QOS as a continuum defined by what net... Hold on, hold on.

[00:17:21] Speaker 02:
You're going too fast.

[00:17:22] Speaker 02:
I can't... Column 12, line 51.

[00:17:24] Speaker 04:
Give us a page number, Odie, that you're looking at.

[00:17:27] Speaker 04:
Sure.

[00:17:28] Speaker 04:
Do you have an appendix?

[00:17:28] Speaker 01:
It's 13597.

[00:17:30] Speaker 04:
Thank you, Judge Prost.

[00:17:35] Speaker 04:
So we first looked at 13597, column 12, line 51.

[00:17:45] Speaker 04:
And I directed your attention to a line.

[00:17:50] Speaker 02:
So different users may have different definitions, that kind of thing.

[00:17:54] Speaker 04:
Yes, Your Honor.

[00:17:55] Speaker 04:
It could be understood as a continuum defined by what network performance characteristic is most important to a particular user.

[00:18:07] Speaker 02:
Where is the continuum language?

[00:18:09] Speaker 04:
That's at line 62.

[00:18:15] Speaker 01:
And then column 13.

[00:18:16] Speaker 01:
Is that what you're going to take us to?

[00:18:19] Speaker 04:
Yeah.

[00:18:20] Speaker 04:
I was going to touch on column 14, line 39, where it explains that ultimately the end user experience is the final arbiter of QoS.

[00:18:30] Speaker 04:
So we see that QoS is an end user concept.

[00:18:34] Speaker 01:
Well, I thought what was pretty probative was the language in column 13.

[00:18:39] Speaker 01:
Ultimately, however, it is desirable that the QoS mechanism operate in a manner that provides the user

[00:18:45] Speaker 01:
with optimal service in whatever manner the user defines it.

[00:18:48] Speaker 04:
You're exactly right, Judge Gross.

[00:18:51] Speaker 04:
That sentence appears after the patent has explained that things change in the network.

[00:18:56] Speaker 04:
Traffic patterns change.

[00:18:57] Speaker 04:
Demands of users change.

[00:18:59] Speaker 04:
And so this patent is saying that really to look at optimization, you have to look at how this user has defined what it determines to be optimal.

[00:19:09] Speaker 04:
And so this is an end user concept.

[00:19:12] Speaker 04:
The claim does recite end user application, IPQOS.

[00:19:18] Speaker 04:
The end users use applications.

[00:19:20] Speaker 02:
So how do you figure out infringement?

[00:19:21] Speaker 02:
You have to survey the end users to figure out if they were happy with it?

[00:19:26] Speaker 04:
I don't think you can determine infringement in this case because it is subject to the vagaries of the user's subjective opinion.

[00:19:38] Speaker 01:
All right, you want to move back to the point you spent most of the time with your friend on.

[00:19:43] Speaker 01:
Tell me what in the specifications supports your, and I guess the ultimately the district court's, final unit in construction here.

[00:19:53] Speaker 04:
So, Your Honor, if I could start with figure four of the specifications that's reproduced at page 12 of our red brief.

[00:20:04] Speaker 04:
And that shows these layers that we've been talking about.

[00:20:08] Speaker 04:
And we've talked about how the top layer, which is shown in red, is the application layer seven.

[00:20:15] Speaker 04:
This dark green middle layer is the MAC layer.

[00:20:19] Speaker 04:
That's this application-aware MAC layer in the patent.

[00:20:23] Speaker 04:
And in between is a yellow layer.

[00:20:25] Speaker 04:
That's layer four, the transport layer.

[00:20:29] Speaker 04:
And then there's a network layer that's shaded in lighter green.

[00:20:34] Speaker 04:
So traditionally,

[00:20:37] Speaker 04:
The idea was, and I think everyone agrees, that these layers were supposed to be isolated from one another.

[00:20:42] Speaker 04:
The MAC layer was not aware of, could not communicate with, for example, the yellow layer, layer four, the transport layer.

[00:20:51] Speaker 04:
It was not aware of, it could not communicate with

[00:20:53] Speaker 04:
the application layer.

[00:20:55] Speaker 02:
I didn't think that this invention was so much about couldn't communicate.

[00:21:00] Speaker 02:
First off, they could have easily based on technology.

[00:21:03] Speaker 02:
It was more about it wasn't the way it was done.

[00:21:06] Speaker 02:
You didn't use application information to allocate bandwidth.

[00:21:11] Speaker 02:
It makes total sense though.

[00:21:12] Speaker 02:
Application information is useful.

[00:21:14] Speaker 02:
If I'm sending a voice message, I just want it to get there quickly.

[00:21:17] Speaker 02:
If I'm sending an Excel spreadsheet, I need it to be accurate.

[00:21:20] Speaker 02:
So I've got to differentiate in the bandwidth in terms of the type of file I'm sending because different parameters matter with regard to this type of file.

[00:21:30] Speaker 02:
That's what I thought was the crux of this invention.

[00:21:32] Speaker 02:
Am I mistaken?

[00:21:33] Speaker 02:
I thought it was using information about the application to allocate bandwidth.

[00:21:38] Speaker 04:
It's more than that.

[00:21:40] Speaker 04:
That's part of it, but there's more than that.

[00:21:43] Speaker 04:
What we see from 3499 is that there's two aspects of application aware.

[00:21:49] Speaker 04:
The first is knowledge of the application, and Judge Moore, you've pointed that out.

[00:21:53] Speaker 04:
And what the patent explains that you use the knowledge of the application for is a way to segregate different packets, to sort the packets based on their application type.

[00:22:07] Speaker 04:
But just knowing the application type doesn't tell you

[00:22:12] Speaker 04:
what are the particular IP QOS requirements that are supposed to be used for this application type, and the patent tells us where we get those.

[00:22:21] Speaker 04:
If you look at column four, lines 10 to 11, which is a reproduction of the abstract, it's also at the end of the abstract, it tells us that the priority class information comes in an application layer communication, in a communication from layer seven to the MAC layer.

[00:22:40] Speaker 04:
It's correct that originally... I'm sorry, I'm on column four.

[00:22:44] Speaker 02:
What line number?

[00:22:46] Speaker 04:
10 and 11.

[00:23:03] Speaker 04:
Just above that, it's explained that there's an application communication.

[00:23:07] Speaker 04:
It comes from the application.

[00:23:08] Speaker 04:
It goes to the MAC layer.

[00:23:10] Speaker 04:
this application communication includes a priority class of the IP flow.

[00:23:15] Speaker 04:
So that's more than just knowing I know the application.

[00:23:17] Speaker 02:
There are at least two embodiments.

[00:23:20] Speaker 02:
Actually, there's many embodiments disclosed in this 80 plus column patent.

[00:23:26] Speaker 02:
You're telling me that I have to read into every claim, none of which currently say this, that there has to be an application communication that includes a priority class,

[00:23:38] Speaker 02:
into every claim.

[00:23:39] Speaker 02:
Not one claim contains that limitation, but you're now saying I should treat this as if it says the President's invention in all embodiments includes this particular transfer of information.

[00:23:49] Speaker 04:
There are two ways that you can get Layer 7 information.

[00:23:53] Speaker 04:
It can be a direct communication from the application, which is as described here, or the MAC layer can read the packet header in Layer 7.

[00:24:02] Speaker 04:
the back layer can read layer seven packet headers, it can read layer four packet headers, it can read layer three packet headers.

[00:24:09] Speaker 04:
Through any of those mechanisms, it can determine the application type, that is true.

[00:24:14] Speaker 04:
But the only place that's described in this pattern of actually getting the IPQOS requirements, which is beyond just the application type, is getting it from reading layer seven information.

[00:24:27] Speaker 01:
Well that's what, okay, that seems,

[00:24:31] Speaker 01:
to be what the district court concluded.

[00:24:33] Speaker 01:
I'm not seeing where anything in the specification really supports that.

[00:24:38] Speaker 04:
Let me help you.

[00:24:39] Speaker 04:
May I respond to that, Your Honor?

[00:24:42] Speaker 04:
Yeah.

[00:24:43] Speaker 04:
So what we know here is that originally, at least an idea was you'd have segregation between the layers.

[00:24:49] Speaker 04:
And we know that the patent's idea was, I'm going to break that model.

[00:24:53] Speaker 04:
I'm going to let the MAC layer be aware of and communicate with

[00:24:59] Speaker 04:
higher layers.

[00:25:00] Speaker 04:
And we see at column 76 of the patent, line 31, a discussion of the fact that because in this invention, the MAC layer communicates with higher layers, communicates with layer 3, layer 4, and layer 7.

[00:25:17] Speaker 04:
For that reason, it's application aware, because it's communicating with layer 7.

[00:25:23] Speaker 04:
It's transport aware, because it's communicating with layer 4.

[00:25:27] Speaker 04:
and it's network aware because it's communicating with layer three.

[00:25:31] Speaker 04:
Those are three different things.

[00:25:32] Speaker 04:
Application aware is different than transport aware.

[00:25:36] Speaker 04:
And the patent tells us the difference is which layer are you communicating with.

[00:25:40] Speaker 04:
If you're communicating with layer four, that makes you transport aware.

[00:25:45] Speaker 04:
We see at the appendix on 3515, the applicant was facing an enablement rejection.

[00:25:54] Speaker 04:
And it said to the patent office that

[00:25:57] Speaker 04:
A person of ordinary skill in the art would understand that, quote, application aware refers to the resource allocator's knowledge of information from application layer

[00:26:13] Speaker 01:
So you're saying, and this is what I'm trying, I'm just trying to understand what the district court was doing here, that there are two separate kind of boxes or inquiries for information, for application type.

[00:26:25] Speaker 01:
You either go through seven directly or you can go through three, four, whatever.

[00:26:29] Speaker 01:
But for application knowledge, that's a whole different thing.

[00:26:34] Speaker 01:
And in application knowledge, it requires

[00:26:38] Speaker 01:
separate and apart from application type inquiry.

[00:26:41] Speaker 01:
Application knowledge requires direct from layer seven.

[00:26:45] Speaker 01:
I just want to know if that's what the district court was saying.

[00:26:48] Speaker 04:
Yeah, I think the district court was saying that knowledge of the application type, of what kind of an application it is, can come from three, four, or seven.

[00:26:56] Speaker 04:
But the second requirement, using information to allocate bandwidth coming from layer seven, or using information at layer seven,

[00:27:06] Speaker 04:
is a separate requirement.

[00:27:08] Speaker 01:
But see, I'm not seeing it in the spec, because the spec uses information type.

[00:27:14] Speaker 01:
So the two words have come out as we say the district court had two separate boxes.

[00:27:18] Speaker 01:
One was information.

[00:27:20] Speaker 01:
It was application type, and the other was information.

[00:27:24] Speaker 01:
And I think the discussion of bandwidth, starting with the first sentence of the abstract, talks about application awareness, where the resource allocator bandwidth resource

[00:27:36] Speaker 01:
to an application based on application type.

[00:27:40] Speaker 01:
So I think everything I read in the spec talks about allocating bandwidth in connection with application type, which is one box of the district court.

[00:27:52] Speaker 04:
And that's the way the claim was originally written, that the application aware allocator allocated bandwidth based on application type.

[00:28:01] Speaker 04:
That claim was amended.

[00:28:03] Speaker 04:
Based on application type,

[00:28:05] Speaker 04:
became part of applicants' argument about what the term application aware itself meant.

[00:28:11] Speaker 04:
And what they did is they then separately recited in the claim.

[00:28:14] Speaker 01:
OK, so let's look at the, you're talking about the amendment that's on page 20 of your brief?

[00:28:20] Speaker 04:
It's probably most helpful, if you want to see the original claim, you'd look at 3527.

[00:28:25] Speaker 04:
OK.

[00:28:37] Speaker 01:
Okay.

[00:28:40] Speaker 04:
So on 3527, you can see that it was claimed that the application aware resource allocator allocated bandwidth resources according to the, or based on the application type.

[00:28:52] Speaker 04:
That was amended.

[00:28:53] Speaker 04:
It was amended to see what you see before you in the briefing as the claim that's being asserted.

[00:28:59] Speaker 04:
Now it says that the application aware resource allocator allocates resources

[00:29:04] Speaker 04:
based on the IPQOS requirements of the application.

[00:29:08] Speaker 02:
Just to be clear, while you say claim one was modified, why wouldn't we assume it was simply broadened as opposed to eliminating the possibility that the resource allocator can do it on application time?

[00:29:21] Speaker 04:
Sure.

[00:29:22] Speaker 04:
If you look at when the amendment was made and the arguments that were made,

[00:29:27] Speaker 02:
Yeah, but why don't you instead look at Claim 10 and Claim 11, which say it can be done by application type.

[00:29:33] Speaker 02:
So it seems to me that Claim 10 is exactly old Claim 1, and Claim 10 is dependent directly on Claim 1.

[00:29:41] Speaker 02:
So why in the world would I read Claim 1 as not having been sort of broadened and still include application type when defending Claim 10 expressly says it's not on the basis of application type?

[00:29:52] Speaker 04:
Well, because the

[00:29:54] Speaker 04:
The applicant has always maintained that part of the definition of application aware is using information from layer 7.

[00:30:02] Speaker 04:
Originally, it tried to use, in addition to that, doing it based on the application type.

[00:30:11] Speaker 02:
in multiple places in this specification, which does use information that it obtains directly from layer 7.

[00:30:18] Speaker 02:
But there are many embodiments that use information that is obtained alternatively from layer 3 or 4 related to packet header.

[00:30:25] Speaker 02:
And that's what's claimed in all of these claims.

[00:30:28] Speaker 02:
How many claims says you can get this information on application type from either one of the following?

[00:30:34] Speaker 04:
Judge Moore, if you look at the specification and the parts that talk about using layer 4 or layer 3 information,

[00:30:41] Speaker 04:
Those are talking about segregating packets.

[00:30:44] Speaker 04:
So you can use that information to determine an application type to put a packet into this bucket or to that bucket.

[00:30:50] Speaker 04:
But what I don't have yet, I don't have the IPQOS requirements that I have to use to actually do my bandwidth allocation on.

[00:30:58] Speaker 04:
The patent describes that those come from layer seven.

[00:31:02] Speaker 04:
The discussion about determining application type deals with segregating packets.

[00:31:07] Speaker 02:
How about column 22, where it is line 30,

[00:31:11] Speaker 02:
Information about IP streams for use in configuring the appropriate QoS mechanism parameters can be extracted from packet headers.

[00:31:20] Speaker 02:
It then says, this is the information that's used for bandwidth reservation and application switching purposes.

[00:31:25] Speaker 02:
Well, packet headers, that clearly tracks back to three and four.

[00:31:30] Speaker 02:
So I... Can I respond?

[00:31:32] Speaker 02:
Yes, of course.

[00:31:33] Speaker 04:
So just above that, it has said that you need to get the appropriate QoS mechanism parameters

[00:31:40] Speaker 01:
You've got to focus me on them.

[00:31:43] Speaker 01:
We're claim 22, so what are you referring to where the line is?

[00:31:46] Speaker 04:
So I believe the judge referred to around line 36.

[00:31:53] Speaker 04:
And I'm pointing above line 36 to say line 33.

[00:31:56] Speaker 04:
The information about the IP streams is communicated vertically in the protocol stack model from the application layer

[00:32:05] Speaker 02:
layer 7 to the MAC layer.

[00:32:10] Speaker 04:
From layer 7 to the MAC layer for bandwidth reservation purposes.

[00:32:18] Speaker 04:
So that's not talking about layer 3 or 4 information, that's talking about using layer 7 information to do your bandwidth allocation.

[00:32:25] Speaker 04:
The discussions about layer 3 and 4 are used to simply identify the application type

[00:32:31] Speaker 04:
so that you can characterize the packet and put it into one bucket or another.

[00:32:36] Speaker 04:
But you still need the second step, which is what the applicant was talking about on 3499.

[00:32:41] Speaker 04:
There's two steps.

[00:32:42] Speaker 04:
You have to figure out what the application type is.

[00:32:45] Speaker 04:
I can put it into my bucket based on that.

[00:32:47] Speaker 04:
But beyond that, you also have to use information from layer seven to do your bandwidth allocation.

[00:32:53] Speaker 02:
Why don't we turn to column 50 and show me where it says the information has to come from layer seven.

[00:33:01] Speaker 02:
Column 50, line 20.

[00:33:03] Speaker 04:
That says that many different priority classes are possible depending on the QoS requirements of the end users.

[00:33:17] Speaker 04:
So depending on the QoS requirements.

[00:33:19] Speaker 02:
And the characterization is based on packet header information to determine the QoS requirements for IP data flow.

[00:33:26] Speaker 02:
Why isn't that exactly contrary to what you're saying?

[00:33:30] Speaker 04:
The packet headers are used to characterize the packets into buckets, but what you need is the QoS requirements.

[00:33:38] Speaker 02:
The packet header information is used to determine the QoS requirements for the IP data flow.

[00:33:45] Speaker 04:
Could you give me a line on that?

[00:33:46] Speaker 02:
I'm sorry.

[00:33:47] Speaker 02:
I'm reading line 22 on column 50.

[00:34:00] Speaker 02:
It's the packet header information, that's what comes from three and four, is used to determine the QoS requirements.

[00:34:11] Speaker 04:
And the sentence starts that that happens during characterization.

[00:34:14] Speaker 02:
I don't see any pulling information from seven.

[00:34:19] Speaker 04:
So what you're reading here is during characterization the packet comes in and you say what kind of packet is this?

[00:34:26] Speaker 04:
What kind of application is this?

[00:34:29] Speaker 04:
And then

[00:34:30] Speaker 02:
since it's a previously known... It's used to determine the QOS requirements.

[00:34:37] Speaker 02:
That's what we're talking about here, right?

[00:34:44] Speaker 02:
I just think there are multiple embodiments disclosed in this specification, as best as I can tell.

[00:34:49] Speaker 02:
And you convinced the district court to limit them to a single embodiment.

[00:34:54] Speaker 02:
And I don't understand why, because the claims, it seemed to me, cover multiple embodiments on their face pretty clearly.

[00:35:00] Speaker 04:
I guess just closing out the layer, the column 50 discussion there, what I see is a characterization of packets.

[00:35:10] Speaker 04:
I don't necessarily, I don't see using layer three or layer four to determine the quality of service itself.

[00:35:22] Speaker 04:
So, if, logically,

[00:35:27] Speaker 04:
if you look at the applicant's statements in prosecution history, logically that the applicant claimed an application aware resource allocator that allocated based on application type.

[00:35:41] Speaker 04:
And the Amed reference was used to reject that.

[00:35:44] Speaker 04:
And it came back with multiple arguments, one of which was Amed doesn't show the application type.

[00:35:50] Speaker 04:
But a separate argument was that Amed doesn't show allocating bandwidth

[00:35:56] Speaker 04:
based on information that comes from layer 7.

[00:35:59] Speaker 04:
If you look at 3518, that argument is made.

[00:36:02] Speaker 04:
They say applicants' invention claims a resource allocator that is application aware, i.e.

[00:36:15] Speaker 04:
it uses information about an application by looking above layer 4 of the OSI model.

[00:36:23] Speaker 04:
And everyone agrees here that above layer 4 of the OSI model in this patent is layer 7.

[00:36:29] Speaker 04:
So the applicant is saying application aware, that term, IE, means it uses information about an application.

[00:36:38] Speaker 04:
And it doesn't just use information about an application coming from anywhere.

[00:36:42] Speaker 04:
It does it in a specific way, using information about an application by looking above layer 4 of the OSI model.

[00:36:51] Speaker 04:
The context of why this limitation is put in makes complete sense when you look at what they were trying to get over from the prior art perspective.

[00:37:02] Speaker 04:
The prior art showed application type.

[00:37:04] Speaker 04:
There was a dispute about that, of course.

[00:37:06] Speaker 04:
But the argument was made, even if it shows application type, it doesn't show that you're using information from layer seven to make your bandwidth allocation decisions.

[00:37:17] Speaker 04:
That was made with respect to the Amed reference,

[00:37:21] Speaker 04:
It was made again at 3499 on the Haleakar reference.

[00:37:28] Speaker 04:
Haleakar was another reference that did its allocation based on the type of service or the type of allocation that was being analyzed.

[00:37:39] Speaker 04:
But what it didn't show was doing that based on information that came from Layer 7 or information at Layer 7.

[00:37:47] Speaker 04:
If you look at the applicant's language at 3499,

[00:37:51] Speaker 04:
and you consider the construction that Ivy is proposing here, they're saying that if you look at the applicant's language, which is where the court's construction came from, the resource allocator must be able to take into account when allocating bandwidth information about applications at OSI application layer seven.

[00:38:13] Speaker 04:
Their whole answer to this is that that's just modifying applications.

[00:38:18] Speaker 04:
It's not modifying where the information comes from.

[00:38:21] Speaker 04:
But everyone in this case agrees that all applications are at layer 7.

[00:38:27] Speaker 04:
There is no such thing as a layer 3 or a layer 4 application.

[00:38:31] Speaker 04:
So when you read the applicant's statement to say that you have to not only know about the type of application, but you have to be able to take into account information about applications at layer 7,

[00:38:45] Speaker 04:
If that just simply means you just have to take into account information about applications that are at layer seven, that would make no sense.

[00:38:53] Speaker 04:
Every application is at layer seven.

[00:38:55] Speaker 04:
What the prior art didn't show was taking into account information that came from layer seven, that you either read a packet header from at layer seven or you had a communication from the application from layer seven.

[00:39:08] Speaker 04:
That wasn't there in the prior art.

[00:39:10] Speaker 04:
They argued that multiple times to justify the meaning of application aware

[00:39:15] Speaker 04:
The patent itself distinguishes application aware from transport aware based on where you're communicating.

[00:39:24] Speaker 04:
If it's layer four, you're transport aware.

[00:39:26] Speaker 04:
If it's layer seven, you're application aware.

[00:39:29] Speaker 04:
That's part of the ordinary meaning of application aware.

[00:39:32] Speaker 04:
Always has been in this patent.

[00:39:34] Speaker 04:
And in the prosecution, they stress that.

[00:39:37] Speaker 01:
We are way beyond our time.

[00:39:42] Speaker 04:
Thank you for your time.

[00:39:49] Speaker 05:
We submit that with respect to the application aware term that the specification, the claim language, the prosecution history, including page 3500 of the appendix, all show that the applicant claimed, embraced the full scope of the application awareness and that the information necessary can come from any source, does not have to come directly from layer seven as the district court apparently concluded.

[00:40:16] Speaker 05:
On the optimized term,

[00:40:20] Speaker 05:
Council pointed to column 12, which says that it is best, therefore perhaps it is best to understand QoS as a continuum defined by what network performance characteristic is most important to a particular user in the user's SLA.

[00:40:38] Speaker 05:
He didn't actually quote the user's SLA.

[00:40:40] Speaker 05:
That's a service level agreement.

[00:40:42] Speaker 05:
And one of the embodiments of the patent is a user agreeing ahead of time which types of data it wants to get

[00:40:50] Speaker 05:
priority for.

[00:40:51] Speaker 05:
So the answer to the question is can we prove infringement if that were the construction?

[00:40:55] Speaker 05:
The answer is yes.

[00:40:56] Speaker 05:
That's an embodiment in the path.

[00:40:58] Speaker 05:
Our larger point here is that the optimized term here is optimizing the end user QoS of the software application.

[00:41:06] Speaker 05:
We provided an expert declaration which discussed how QoS, that term, is used by those who are still in the art and said this was not indefinite.

[00:41:14] Speaker 05:
The district court found it indefinite without referring to the expert declaration where we believe that there are underlying issues of fact for trial.

[00:41:22] Speaker 05:
Furthermore, if you read optimize, more in the context of customizing, you're customizing this system for the prioritized voice over internet packets.

[00:41:34] Speaker 05:
That is optimizing the system to those of skill in the art.

[00:41:37] Speaker 05:
That's what our expert would have testified to in trial if he were permitted to do so.

[00:41:42] Speaker 05:
Furthermore, this is a means plus function claim.

[00:41:45] Speaker 05:
And the court should have looked at the specification for the things in the specification which could be tied to the optimized term.

[00:41:54] Speaker 05:
And having done that, then we have the meets and bounds of the claim.

[00:41:57] Speaker 05:
So even if optimized generally might be too broad a term, when it's embedded in a means plus function claim like this, we look to the specification to determine the meets and bounds of the claim as is not optional but mandatory under 112.6.

[00:42:11] Speaker 05:
And that would give us what we need.

[00:42:12] Speaker 05:
The district court did not do that analysis.

[00:42:14] Speaker 05:
He truncated the analysis.

[00:42:16] Speaker 05:
And he therefore found the claim indefinite without hearing from experts who were doing a proper search of the specification.

[00:42:23] Speaker 05:
For those reasons, Your Honor, we respectfully request that the court reverse.

[00:42:27] Speaker 05:
And unless there are any further questions.

[00:42:30] Speaker 00:
Thank you.

[00:42:31] Speaker 00:
Thank you, Your Honor.

[00:42:32] Speaker 00:
That concludes our proceedings for this morning.