`571-272-7822
`
`
`
`
`
`IPR2015-01872, Paper No. 28
`January 12, 2017
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`ERICSSON INC. And TELEFONAKTIEBOLAGET LM
`ERICSSON,
`Petitioner,
`
`v.
`
`INTELLECTUAL VENTURES II LLC,
`Patent Owner.
`____________
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`____________
`
`Held: December 8, 2016
`____________
`
`
`
`
`
`BEFORE: BRYAN F. MOORE, BRIAN J. McNAMARA, and
`DAVID C. McKONE, Administrative Patent Judges.
`
`The above-entitled matter came on for hearing on Thursday,
`December 8, 2016, commencing at 1:01 p.m., at the U.S. Patent
`and Trademark Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`
`
`
`
`
`
`
`
`
`ON BEHALF OF PATENT OWNER:
`
`J. ROBERT BROWN, Jr., ESQUIRE
`CHARLES J. ROGERS, ESQUIRE
`Conley Rose, P.C.
`5601 Granite Parkway
`Suite 500
`Plano, Texas 75024-6608
`
`BYRON L. PICKARD, ESQUIRE
`LORI A. GORDON, ESQUIRE
`Sterne, Kessler, Goldstein Fox
`1100 New York Avenue, N.W.
`Washington, D.C. 20005
`
`
`
`
`
` 2
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE McNAMARA: Good afternoon, everyone.
`This is the oral hearing in case IPR2015-01872, Ericsson and
`another Ericsson entity which I won't even try and pronounce
`versus Intellectual Ventures. Beginning with the petitioner,
`would the parties please introduce themselves.
`MR. ROGERS: Charles Rogers, backup counsel. I
`have with me here my colleague, Robert Brown, lead counsel,
`and in-house counsel, Brian Kearns.
`MR. PICKARD: Good afternoon. Byron Pickard on
`behalf of the patent owner, and I'm joined today by my partner,
`Lori Gordon.
`JUDGE McNAMARA: As you can see, Judge McKone
`is participating remotely. So I would ask everyone to please
`speak from the microphones, and in reference to any
`demonstrative or paper that you are displaying here, please refer
`to it so that -- where it is in the record so that Judge McKone can
`access it as well.
`Today the petitioner will open the hearing, present its
`arguments with respect to the challenged claims that we've
`instituted on. Patent owner will then respond and the petitioner
`can then exercise its right to rebuttal. Is there any amount of time
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
` 3
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`that the petitioner would like to reserve for rebuttal? There are
`45 minutes argument for both sides.
`MR. ROGERS: Petitioners would like to reserve ten
`minutes.
`JUDGE McNAMARA: Okay. If everybody is
`prepared, we can begin.
`MR. ROGERS: Good afternoon. May it please the
`Board, we are going to start out with slide 2. And side 2 lists the
`instituted claims, the majority of which are based upon a 103
`single-reference obviousness assertion, the single reference being
`the prior art Lu patent. And that is the majority of the claims.
`Then with the exception of claims 22 and 25, which bring in
`Pankaj as a combination of Lu and Pankaj as a 103 obviousness
`assertion.
`Moving to slide 3, slide 3 is the overview of the '994
`patent. The '994 patent generally relates to methods of processing
`queued packet datas that are described as a combination of
`hierarchical, which is tier-based, and weighted fair queueing.
`And the specification in the '994 patent refers to this combination
`as tier-based weighted fair queueing. And the spec of the '994
`patent says that this approach of combining these two prior art
`queueing schemes has not been considered in the past, but as
`we'll see going forward here, this combination is precisely what
`the Lu prior art reference discloses.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 4
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`Moving on to slide 4, there were no claim construction
`issues in the institution decision. There were no disputed claim
`construction issues. The only claim construction that was done in
`the institution decision was the Board properly construing claims
`11 through 19 and 24 as reciting means-plus-function terms. And
`the patent owner has not disputed the construction for these
`means-plus-function terms. There is one claim construction issue
`that's come about after the institution decision that arises from the
`patent owner's response regarding the term in claim 1 "queued
`packet data users." And we'll see that as we go forward here.
`Next slide, 5, the goal as stated in the specification for
`the '994 patent is to provide a fair and optimal allocation of
`limited communication resources to data packet users having
`different requirements or throughputs. And this is specifically
`described in column 2 as quoted here on slide 5 where the spec
`for the '994 patent says it is important to optimize use of the
`limited communication resource which becomes even more
`important when individual data packets have different
`requirements with respect to delay, bit error rate, et cetera, such
`as quality of service requirements.
`And slide 6, this is a reference to the background
`section of the '994 patent which begins with the description of the
`two prior art queueing schemes that it describes, first weighted
`fair queueing and then hierarchical queueing. And moving to
`slide 7, after the initial description of these two prior art queueing
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 5
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`schemes, the '994 spec describes the pros and the cons of each of
`these two prior art queueing schemes.
`JUDGE McNAMARA: Could you take just a minute
`and summarize for me the difference between weighted fair
`queueing and hierarchical round-robin.
`MR. ROGERS: So weighted fair queueing, that prior
`art scheme uses weights, and it's to provide proportions of the
`communication resource. So if you go through the prior art,
`beginning with first-in/first-out as queueing and then there's
`priority queueing, which will queue to one input over the other
`and won't even get to the other lower priority input to service that
`queue until the higher priority queue is idle or empty.
`Then there's fair queueing which will give an equal
`amount to each -- this is an example which is two inputs. It gives
`equal amounts to two different inputs. And then weighted fair
`queueing, actually instead of giving equal amounts to the
`example of two inputs will provide a percentage of the two.
`So that then is a great segue into hierarchical queueing
`and hierarchical weighted queueing. Hierarchical has -- you'll
`have two different levels. So you may provide one queueing
`scheme, whether it's 5-0, fair queueing, weighted fair queueing or
`priority queueing, you provide that at one level and then you have
`a hierarchy that you move down to the next level and you provide
`another queueing scheme.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 6
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`So the best example of the hierarchical weighted fair
`queueing would be where you have -- first you apply, let's say,
`75 percent and 25 percent at one level and then you work your
`way down a tree down to the next level. If you apply 75 percent,
`25 percent again, that far left-hand side root down to the leaf is
`the amount of the resource that's provided would be a product of
`the two. And that is specifically described in the Lu disclosure,
`that particular example.
`So moving forward, slide 7, pros and cons of each. And
`then moving to slide 8, you have the proposed solution in the '994
`specification to the cons, to the disadvantages of the prior art
`queueing schemes is described. The proposed solution is
`described as a packet data queueing algorithm based around a
`concept of employing different tiers of service to provide users
`with a commitment that a portion of the entire system bandwidth
`will be allocated to users operating on that particular tier. And
`the abstract of the '994 spec describes this method as including
`allocating a tier service for substantially each of a plurality of
`individual packet data queues in which a proportion of a total
`number of data packets is allocated to a number of the tiers of
`service.
`
`And if you go moving to slide 9, what slide 9 shows
`you is we'll get to the claims, the specific claims and how Lu
`discloses or renders obvious the claims of the '994 patent. But
`first I want to compare the description in Lu to the description of
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 7
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`what the '994 patent describes as its invention. The '994 patent at
`column 4 describes its approach of combining hierarchical and
`weighted fair queueing saying that it's never been considered in
`the past. Well, we know from the Lu disclosure that it has been
`considered. So precisely what the '994 patent claims describe, it's
`an invention to be specifically disclosed in the prior art of Lu. Lu
`at column 2 describes its invention as the class queueing system
`that uses a weighted-based scheduling scheme.
`And if you go to slide 10, further similarities between
`the Lu disclosure and the '994 spec. The '994 spec describing its
`data queueing algorithm as based around the concept of
`employing different tiers of service. The apparatus and method is
`summarized and described to provide a tier-based weighted fair
`queueing scheme in which packets are allocated in a round-robin
`fashion.
`Similarly in Lu --
`JUDGE McKONE: Mr. Rogers, you are welcome to
`proceed however you would like here, but given the limited
`amount of time, it may be more beneficial to concentrate on the
`specific elements of the claims that are really in dispute at this
`time.
`
`MR. ROGERS: Understood. So let's move forward in
`the slides to slide 12 where I'll give a summary of how we are
`going to go through these claims. We are going to first deal with
`claim 1 and then the corresponding claims 11 and 24. So 1 is a
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 8
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`method claim, 11 being corresponding apparatus and system
`claims. Then the next group of claims is 7 and 17, which we'll
`deal with together. And then 21, 22, and 25 is where we bring in
`Pankaj as a combination for the 103 obviousness.
`We do have slides on the Section 315(a) and (b)
`arguments but I'm reserving those as necessary for rebuttal
`because when we saw the patent owner's request for oral
`argument and the statutory requirement for listing issues that are
`to be argued, they are not listed. Nor do we see any 315(a) or (b)
`arguments in their slides. I understand that still is our burden of
`proof, but I did not plan on reserving time for making those
`arguments unless the Board has questions or allows the patent
`owner to make those type of arguments. So like I said, we do
`have slides on that, but we may not get to them.
`So moving, as suggested, to slide 13, claim 1, claim 1
`has a method claim that has five steps. We've got allocating a tier
`of service, determining a total number of data packets that can
`use in available communication resource, you've got two more
`allocating steps allocating different weights and then allocating a
`proportion of the data packets to the tiers, and then finally
`providing the communication resource to users of queued packet
`data users on a tier-by-tier basis. Those are the five method steps.
`There's only two in dispute at this point. So I want to limit the
`discussion to elements identified as B and E.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 9
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`So moving to slide 14, element B, this can be broken
`down into two different disputed areas. It's determining a total
`number of data packets that can use an available communication
`resource. The first aspect of this element B is determining the
`total number of data packets. So I'll handle that first.
`This determining a total number of data packets is
`disclosed in Lu as a weight-based scheduling scheme that may
`operate in cycles where you have for each cycle a number of data
`packets is selected for transfer. That's done based on weights.
`And the data packet volume for any period of time for each class
`may be determined.
`We go to slide 15, this is where we see --
`JUDGE McKONE: Before you move much further,
`just so I have the right context, what is it that you contend is the
`communication resource, the available communication resource?
`MR. ROGERS: Well, the communication resource, it's
`shown it can be a connection on a network, the bandwidth for
`transmitting data or it can be storing data. But in this case, the
`most prevalent example is communicating data and the
`communication resources defined as the bandwidth, the available
`bandwidth for the communication.
`JUDGE McKONE: So it would be the bandwidth, if we
`look at, you know, any of the figures, like Figures 4 or 5, that
`bandwidth of the line out, say, 209?
`MR. ROGERS: Yes, link 209.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 10
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`JUDGE McNAMARA: So it's the bandwidth of the
`length then after all of this processing has been done to figure out
`what should be transmitted?
`MR. ROGERS: Yes. I mean, in Lu you could also
`consider the final point where the data is queued also as a
`communication resource, but it's easier to think of it as the link
`209.
`
`JUDGE McNAMARA: Okay.
`MR. ROGERS: So it's the bandwidth of the link. So
`we'll see on slide 15 you have the '994 patent describing this
`determining of a total number of data packets with reference to a
`preferred embodiment where it assumes that allocation of
`resources can only be made at certain time intervals or rounds. In
`this particular example at column 7 it has each round 100 packets
`are allocated. Lu discloses the determining of the number of data
`packets in similar terms per round. It refers to cycles. And it
`uses a weight-based scheduling scheme that can operate in cycles
`where for each cycle a number of data packets is selected for
`transfer. So you are seeing precisely what the '994 patent
`describes is its determining step as disclosed in Lu.
`JUDGE McNAMARA: As I understand, though, the
`patent owner's position is that this weighting that goes on in Lu is
`not the same as finding the total number of packets. And you
`disagree with that; is that right?
`MR. ROGERS: Absolutely. In Lu --
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 11
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`JUDGE McNAMARA: That's probably a good issue to
`try to make really clear why you disagree with that.
`MR. ROGERS: Lu is done in two steps where you first
`determine the weights and then the weights determine the number
`of packets. That's what weighted fair queueing is. You apply
`weights so that you can proportion the bandwidth that's going to
`be provided to the various inputs, the various users. And so that's
`pretty straightforward that you apply weights and it's a portion of
`the bandwidth. You multiply and the product of that determines
`then the number of packets.
`JUDGE MOORE: When you say a portion of the
`bandwidth, are you saying that the number of packets, let's say
`the weights are 25 and 75, are the number of packets that's in that
`25 basket and the number of packets that are in that 75 basket
`determined by the total bandwidth available at the link or by the
`number of packets moving towards that link, if I'm saying that
`correctly? In other words, is the bandwidth the determiner of the
`exact number of packets in each bucket?
`MR. ROGERS: It is, because for weighted fair
`queueing you are giving a weight and multiplying that weight as a
`proportion of the available bandwidth, the bandwidth that has
`been selected. So if you look at Lu in column 4 that goes through
`column 5 -- this particular example we are getting to is in column
`6. You have a one-to-one correspondence between the weights
`and the number of packets. But that's really just a simplified
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 12
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`explanation of the way the selection of the packets is done. If
`you look at column 4 and column 5 in the Lu patent, it describes
`the minimum amounts assigned to the weight classes and
`applies -- you look at the minimum amounts that can be
`guaranteed and calculate your weights based upon the minimum
`balance and it becomes a proportion.
`And then as Lu describes, you have two different stages
`when you are selecting data. You have got a predetermined
`selection of data and then at the outset, before the process begins
`and then throughout the process as Lu describes, it's keeping track
`through soft and hard buffer thresholds. It's keeping track of the
`amount of data that is actually used and all the way to the point of
`knowing what the available bandwidth is because it keeps track
`and uses these thresholds to determine when you have an
`overcapacity. Lu describes that you use these soft and hard
`buffer thresholds to determine situations of overcapacity and to,
`along the way, adjust the weights and the capacity.
`So the system knows in Lu, it knows what the
`bandwidth is, it knows what the capacity is, and it provides a
`proportion of that through its weighting system so that the
`available resource is -- and use of the available resource is
`optimized.
`This example in 17 is really just a simplified example
`where your weights in determining the number of packets, there's
`a one-to-one correspondence. So in this example shown at slide
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 13
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`17, you have weights for classes, high, medium and low, that are
`assigned weights of 5, 2 and 1. And in this particular simplified
`example, when you have one-to-one correspondence, that
`determines that you have got five data packets selected for the
`high class, two for the medium and one for the low.
`Then you go -- so that's being transferred on that first
`
`cycle.
`
`JUDGE McKONE: But that's being transferred within
`the various buffers. Not transferred out to link 209, for example,
`correct?
`MR. ROGERS: Well, when it's weighted fair queueing,
`you work your way through the system. So you work your way
`through levels, and then transferred out is the final level. The
`answer to your question is yes and no because when you are
`talking about the lower levels, that's just the precursor to when
`you get to the ultimate higher level that is then transmitted out.
`So these are data packets selected along the way for each level.
`And as is done in hierarchical queueing, when you get to the top
`of the hierarchy, then it's output and queued to the
`communication resource.
`JUDGE McNAMARA: The question I have is that the
`claim element 1B says determining a total number of packets that
`can use an available communication resource. So as I understand
`this weighting that's going on in Lu, it's saying, well, I'm going to
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 14
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`assign five high, two medium, one low. How is it determined that
`it's eight that can use that resource?
`MR. ROGERS: With its buffer thresholds. This is the
`predetermined amount. With the soft and hard buffer thresholds
`it's monitoring the amount of data that's actually being used
`throughout the system. And it monitors the system and adjusts.
`As it says in Lu, it adjusts the weights and capacity accordingly.
`So it's monitoring the system. It's keeping track of overcapacity
`situations through its thresholds. And in some instances it talks
`about dropping data, but that's the hard threshold.
`In particular with respect to the soft ones, it's using
`those to make sure there's not data drop and it keeps track of what
`the capacity is. It can adjust the capacity as well. So it knows
`what the bandwidth is and the available communication resource,
`and it uses these weighting schemes and adjusts the weights so
`that it is determining the amount that can be used.
`JUDGE McKONE: Now, why do you conclude that Lu
`is determining a number of packets that can use the resource
`rather than other potential indirect measurements of bandwidth
`such as we are dropping packets, we probably have to decrease
`the input or it seems like there's a lot of extra space, maybe we
`increase the input. Why is it necessarily determining a number of
`packets that can use an output resource?
`MR. ROGERS: Because it's selecting the packets to be
`used and then it modifies and selects packets along the way. So
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 15
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`even if the predetermined selection of packets is interpreted to not
`fall within this determining step, it certainly is along the way in
`the process how Lu describes the actual number of packets being
`used and adjusting the weights so that it makes that selection of
`what is actually being used all the way to the point of fully taking
`advantage of the capacity of the communication resource.
`JUDGE McNAMARA: Again, it's determining the
`number of packets that can use a communication resource. So if
`the communication resource is a fiber optic cable versus another
`communication resource that is something of much lower
`bandwidth, can you point to something in Lu by column, line,
`whatever, that describes that Lu figures out how many packets
`that particular communication resource can use?
`MR. ROGERS: If you look at column 2, lines 46
`through 49, Lu discloses that based on a specific selection of
`class specification parameters, the resulting bandwidth/line,
`quality/error rate, et cetera, may be determined and applied as
`appropriate. And that application is the selection of the data
`packets that can use the available resource.
`JUDGE McNAMARA: And then is there some
`disclosure of changing that then to -- with that information,
`figuring out the number of packets that can be used?
`MR. ROGERS: Yes, because the weights determine the
`number of packets. So moving on to slide 18, I think we've
`covered all this, so I'm going to go to element E, slide 20.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 16
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`Element E at slide 20 is the final step of claim 1, providing said
`communication resource to queue packet data users on a
`tier-by-tier basis such that said communication resource is made
`available to a number of tiers. So I read that explicitly the way
`it's phrased, but this is where we have a claim construction
`dispute. The way the Board should read this, if you go back to
`slide 19 -- slide 20, if you look at slide 20, the reference to
`queued packet data users should be read as users of queued
`packet data. The patent owner wants this to be instead interpreted
`as users of packet data that are queued. They want you to
`narrowly read this claim to require queueing of users as opposed
`to the way this should be read.
`What this limitation is talking about is providing the
`communication resource to the users. There's no requirement in
`this limitation that says that the users are required to be queued.
`And so let's move on to slide 21, and you'll see the
`patent owner's argument. They say that queued packet data users
`mean users themselves are queued. Well, setting aside the
`physical impossibility of queueing users we know from the patent
`owner's expert that that's really not what they are arguing.
`JUDGE McKONE: But of course it's not what they are
`arguing. I don't think anybody is arguing that users are in a bread
`line or something or in a line at Disney World waiting for a ride.
`I think everybody is well aware that the argument is that an
`indication of some sort of the user is being queued. It seems like
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 17
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`the difference between your construction and theirs is that you are
`arguing that packets are queued. Whereas, patent owner has
`argued that an indication of the users are queued which is then
`correlated later to packets that are set. What support do you have
`that your construction is correct?
`MR. ROGERS: Because if you look at the claim, they
`are trying to read into this claim this requirement that users are
`queued. There's nothing in the claim that requires that users be
`queued or indication that users be queued. That's in claims that
`are not instituted. That's in claim 5. So if you are going to use
`claim differentiation as it should for interpreting these claims,
`you don't read requirements from a non-instituted claim into the
`independent claim.
`Claim 5 requires that there be an identification code that
`is queued. That's not in claim 1. Claim 1 is broader. And
`claim 1 should be read as what was understood at the time of the
`invention and even now that you are queueing data packets. And
`even if you read into a requirement that not only the data packets
`are queued, but some kind of identification is required, that's
`disclosed in Lu as well. If you look at slide --
`JUDGE McKONE: Before dealing with the claim
`construction argument, could you show me what your support is
`in the '994 patent for your construction of claim 1 that packets
`rather than indications of users are queued.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 18
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`And maybe I could be more focused on that. I
`understand patent owner's argument to be that Figure 3, for
`example, of the '994 patent shows queueing of user identifications
`which are later correlated to packets when a particular user
`identification is found in a queue. Is that incorrect, an incorrect
`characterization of Figure 3?
`MR. ROGERS: I think in that particular example in
`Figure 3 they are queueing -- you are correct, they are queueing
`the user identification, and the data itself is referenced and stored
`separate in memory in that particular example.
`JUDGE McKONE: Is there another example that you
`would like to point us to where the '994 patent describes queueing
`of packets as you would read claim 1?
`MR. ROGERS: Well, at column 7, lines 45 through 49
`it discusses allocation of data packets and it talks about tiers and
`these data packets are allocated to each tier. And under queueing
`schemes when you are talking about allocating a data packet to a
`tier or a queue, you are queueing the packet data. And --
`JUDGE McKONE: Isn't this in the context of a
`description of Figure 3 which you just agreed was queueing user
`IDs rather than packets themselves?
`MR. ROGERS: It's in the same area where they are
`discussing Figure 3, yes, but I don't read it as that limited. Under
`the broadest reasonable interpretation standard, I wouldn't think
`it's appropriate to limit these claims in that fashion. It may be
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 19
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`reading into the claim an embodiment. And then under claim
`differentiation, you look in that description of having a customer
`identification is limited to this dependent claim 5. So traditional
`claim construction constructs would have you not read that claim
`5 into claim 1. It would be redundant to require -- to read claim 1
`to be so narrow that it requires queueing of a user reference, user
`identification as opposed to queueing the data packets
`themselves.
`If you are intent on limiting the construction of even
`claim 1, then --
`JUDGE McKONE: I don't see this as an issue of
`limiting the construction. It seems like there's an either/or that we
`are presented with here. Either we are construing the claim to
`require queueing of packets or we are construing the claim to
`require queueing of user identifications.
`MR. ROGERS: Well, I view it as either/or is a way to
`characterize it. The other is and/or. The claim should be broad
`enough under the broadest reasonable interpretation to cover
`both. And if you are not going to have the claim interpreted, if
`you are going to have the claim interpreted to require queueing of
`not only the data packet, but of the user identification, you can
`find that disclosure in Lu as well. You'll see that Lu talks about
`queueing a data packet that has user identification in the header.
`And so even Lu itself you wouldn't interpret -- if you are going to
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
` 20
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`interpret the claim, claim 1 of the '994 to require queueing of user
`information --
`JUDGE McKONE: I assume you are talking about, for
`example, if we look at Figure 7 of Lu there is an identification
`field. That's what you are claiming is a user identification?
`MR. ROGERS: Yes. And that's one aspect.
`JUDGE McKONE: Do you have any evidence that
`that's a user identification?
`MR. ROGERS: Well, that's what our expert testified as
`to what one of ordinary skill in the art would understand. And I
`see it as well. When you have a data packet with the header and
`it has information, identification information, that's identifying
`who the user is. You are not going to send a data packet through
`a communication system willy-nilly just to go off into space.
`You’ve got to keep track of who the user is, and you do that
`through the header information.
`If you look in Lu --
`JUDGE McKONE: In the petition with regard to
`claim 5, it's my recollection that you argued that this
`identification field, it would have been obvious to make this
`identification field a user identification rather than the
`identification field actually disclosing a user identification field.
`Am I incorrect about your earlier arguments as to claim 5?
`MR. ROGERS: I think that's one way to look at it. But
`the other way is really --
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` 21
`
`
`
`
`
`
`
`Case IPR2015-01872
`Patent 7,385,994 B2
`
`
`JUDGE McKONE: I'm asking about the way you
`looked at it in the petition.
`MR. ROGERS: I didn't see it as that narrow. If you
`look at the way obviousness arguments are handled is, yes, we'd
`like to see Lu when it says identification information to
`specifically say customer ID. That would be better. But when it
`says identification information, that is -- as we approached it, is
`that makes it obvious to one of ordinary skill in the art that it
`knows that it's either there or it could be.
`But you know, the more important point is the system,
`all these systems have to keep track of the user and its association
`with the data. I'll point out another example. If you look at Lu on
`column 2, this talks about the state the art prior to the '994 patent
`and it talks about at column 2 at the very top, the class queueing
`system may use a weight -- hold on a second. I'm on the wrong
`one. Hold on.
`There is a discussion which I'll talk about as I look for
`the specific cite to it, but there's a discussion in Lu talking about
`dropping data packets and when it's voice data packets. And you
`want to drop -- if you are going to use dropping of data packets
`for voice, you have got -- you want to drop data packets in the
`same communication. An