`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`LIMELIGHT NETWORKS, INC.,
`Petitioner,
`
`v.
`
`AKAMAI TECHNOLOGIES, INC.,
`Patent Owner.
`____________
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`____________
`
`Record of Oral Hearing
`Held: December 5, 2017
`____________
`
`
`
`
`
`
`Before GREGG I. ANDERSON, JENNIFER MEYER CHAGNON, and
`JASON W. MELVIN, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`BARRY J. SCHINDLER, ESQUIRE
`VIMAL M. KAPADIA, ESQUIRE
`Greenberg Traurig, LLP
`500 Campus Drive
`Suite 400
`Florham Park, New Jersey 07932
`
`ON BEHALF OF PATENT OWNER:
`MICHAEL J. SUMMERSGILL, ESQUIRE
`HEATH A. BROOKS, ESQUIRE
`DONALD R. STEINBERG, ESQUIRE
`Wilmer, Cutler, Pickering, Hale & Dorr, LLP
`60 State Street
`Boston, Massachusetts 02109
`
`
`
`
`
`
`The above-entitled matter came on for hearing on Tuesday,
`December 5, 2017, commencing at 2:10 p.m., at the U.S. Patent and
`Trademark Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` 2
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE MELVIN: So we'll proceed to the hearing in
`IPR2016-1894, also between Limelight Networks, petitioner, and
`Akamai Technologies, patent owner. This is reviewing patent number
`7,472,178. We will follow the same procedures as the prior hearing. I
`understand we have different counsel arguing for patent owner?
`MR. BROOKS: Yes, Your Honor, Heath Brooks on behalf of
`Akamai.
`MR. SCHINDLER: Your Honor, you still have me.
`JUDGE MELVIN: Excellent. And would you like to reserve
`ten minutes?
`MR. SCHINDLER: Yes, Your Honor, I would like to reserve
`ten minutes, if I may.
`JUDGE MELVIN: Okay. We are ready when you are.
`JUDGE ANDERSON: Can I get the hearing room staff to
`switch the camera off our judges -- there you go.
`MR. SCHINDLER: You wanted to see me, I know.
`JUDGE ANDERSON: I did.
`MR. SCHINDLER: So Your Honor, thank you again for this
`afternoon. If you would go to, we would like to start on slide 31 of
`petitioner's slide 31. The first claim term we would like to address is
`"upload." There appears to be a lot of discussion in patent owner's
`statement about upload, so we believe this is one of the terms we should
`start with. And we just put up the term in the context of the claim.
`
`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
`
`
`
`
`
` 3
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`If you would go to slide 32, essentially this is what patent
`owner is arguing, that we understand that the meaning of the term
`upload -- so we first take the position that the patent owner never asks for
`a claim construction of the term "upload," but it appears now that they
`are asking for that term or whatever -- however they are reading into it.
`And that's on 32. That's what we believe their definition of what they are
`now asking what upload means.
`If we go to slide 33, Judge Anderson can you hear me? Is it
`
`okay?
`
`JUDGE ANDERSON: (Nodding.)
`MR. SCHINDLER: Slide 33 is our response to this we believe
`it's basically importing limitations and it's narrowing. We believe what
`we've posited is slide 34, is basically we believe that the specification
`does not define upload. Now, I would like to spend a little time on this
`slide. So this slide is going to be in context with, so this is directly from
`the spec 6:61 through 7:10. And it's our position that this actually
`discloses -- because there appears to be argument that upload can only
`mean pushing and not pulling. And it's our position that because of the
`spec that appears to argue that upload can also be pulling, that we should
`not be reading a limitation of only pushing.
`Now, how do we get there? So this is talking about the figure.
`Let's put up the figure. I would like to in this respect put up Figure 7. So
`the reason why it's hard to read, and I apologize, but if you could look on
`your patent at Figure 7, what Figure 7 shows you -- and this is important,
`so if you look at the term on Figure 7, and that appears strangely out 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
`
`
`
`
`
` 4
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`order, it appears after Figure 4. So what I'm focusing on is the word
`"upload via FTP or tape." It's all the way on the left-hand side. You see
`content provider 704, 700, between it says upload via FTP or tape.
`And if you go back to the slide that I had before, 34, this is
`actually consistent with what's being described because it's says, first it
`talks about -- the first part it talks about FTP client. You see the upload
`may be accomplished with any industry standard FTP client. That's
`shown in that figure FTP. And then if you go a little bit further down, it
`talks about the above technique are merely representative and it talks
`about tape. And there's the word CD-R, CD-W [sic] tape. And then it
`says yet another alternative is to have a given one of the storage sides
`dynamically pulled.
`So it's our position that this is all within the context of upload.
`It's not another embodiment. It's talking about what's going on at the
`upload via FTP or tape or possibly pulled. So it's our position that if you
`were to go with patent owner's definition of limiting it to just pushing,
`you would read out an embodiment that's directly in the spec.
`If you would go on to 36 and 37, we gave a series of definitions
`on upload, upload directly from IEEE and directly stating that.
`Go to 38. It's our position that if you take our definition of
`what upload means, which means merely transfer, we believe there is no
`dispute that Speedera does do an upload, and as quoted in our petition
`and in our reply. If you go to slide 39, Chaudhri, we take the same
`position that if you take our definition, Chaudhri makes it.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 5
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`If you go to the next term, we would like to talk about
`asynchronous, which is slide 40. We have up in the claim operate
`asynchronously. And go to slide 41. This is what patent owner argues,
`that petitioner's argument fails because it ignores that web servers must
`also operate asynchronously. Not just autonomously or independently.
`Slide 42 is our reply. And basically one of our arguments, key arguments
`which is in 42 is basically that their expert did not say that it was not
`asynchronous. He wouldn't say basically anything. I don't recall
`Speedera saying that they were synchronous or any sort, whether
`synchronous or asynchronous. So their expert basically punted on this,
`wouldn't say whether it was synchronous or asynchronous.
`If you go to slide 43, their expert, we believe, reads into this
`thing called common goal. And their expert says, well, asynchronous
`must mean common goal. And that's if you see the question and answer,
`the answer is they cannot operate asynchronously from each other
`because they are not processing towards a common goal. And we don't
`really know what common goal means. We believe that's reading in a
`limitation that's not in the spec.
`JUDGE MELVIN: What is it you think that asynchronously
`adds beyond autonomously?
`MR. SCHINDLER: So autonomously and independently.
`JUDGE MELVIN: That's your word, right?
`MR. SCHINDLER: Independently is our word. It means these
`two machines are not joined together. They are independent of each
`other. So that's what we would say of autonomous is more than just
`
`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
`
`
`
`
`
` 6
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`autonomous. Asynchronous means that they both have to be not
`synchronized with each other. I hate to use the opposite but that's really
`what it is. We take the position that Speedera is -- they are not
`synchronizing with each other. So if they are not synchronizing with
`each other, then they have to be asynchronous.
`If you go to slides --
`JUDGE ANDERSON: So excuse me, what is your evidence
`for that in Speedera that it's --
`MR. SCHINDLER: Not coordinating? So if you go to, we
`have a nice picture of Speedera. So if you look at Speedera here, this is
`Figure 1 of Speedera. What you are dealing with is the web servers.
`And it's basically the colors are blue, blue and blue, that they are
`independent and they can't be synchronizing with each other because
`there's nothing in Speedera that says you have to make sure that when
`one responds, the other one responds the same time. It's basically the
`way our expert is and the way the deposition is. And we put up, for
`example, the Burger deposition that talked about this. So that's our
`position that the figures themselves show that they are independent from
`each other.
`JUDGE ANDERSON: So independence, if we look at the fact
`that they are different servers or content delivery sites, that's enough for
`us to conclude that they are asynchronous?
`MR. SCHINDLER: Well, because the fact of the matter is
`unless Speedera said that they were going to be synchronizing each other,
`what I would call a default measure is always asynchronous. You have
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`to be able to actively synchronize something because you have two
`different machines that are operating separately. They are independent
`from each other. Unless you have a program that is saying you need to --
`the two need to synchronize with each other, the natural state is always
`going to be asynchronous.
`JUDGE ANDERSON: So where is the evidence of that? It's
`your burden, so where is the evidence? That sounds good as argument.
`MR. SCHINDLER: So we cite to -- we have our expert who
`gave a number of -- that's Friedman. And if you give me a second, we
`cited -- so this is you are talking Speedera, correct?
`JUDGE ANDERSON: (Nodding.)
`MR. SCHINDLER: So in Speedera, for example, if you go to
`slide 38, and we cite both a reply in the petition and within that we cite
`within that is the Friedman declaration that cites that. I'm sorry, so this
`was upload. I think put up slide 48 again. So our evidence, Judge
`Anderson, this actually gets to one of the motions to exclude, a Speedera
`claim chart, Exhibit 1012 at page 41, which is the Friedman talking about
`Speedera. And it's in the context of independence. That's the evidence
`that we posit.
`JUDGE MELVIN: You are saying you have to rely on that
`claim chart?
`MR. SCHINDLER: For additional evidence. We don't believe
`we need to. But Judge Anderson said the additional evidence. This was
`our expert's at that claim chart.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 8
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`JUDGE MELVIN: The chart is part of what your expert -- the
`chart contains statements of your expert; is that accurate?
`MR. SCHINDLER: Correct.
`JUDGE MELVIN: Not just quotes from the prior art?
`MR. SCHINDLER: It is quotes from the prior art.
`JUDGE MELVIN: Solely?
`MR. SCHINDLER: No. It also provides his opinion.
`So if we go to the next term, the next term is "CDN storage
`sites." And that starts with slide 14. And it's our position -- so this is
`how patent owner again didn't ask for a -- go to slide 15. Patent owner
`didn't ask for a claim construction, but it appears that patent owner is now
`putting together a CDN storage site that is a narrower construction. This
`is the Court's construction of what a CDN storage site is on slide 15. And
`this is the institution decision, which is what we believe is a correct
`decision on a storage site.
`If you go to slide -- really just moving to along, if you really go
`to slide 20, this is our position with regard to Speedera. And the
`Speedera does have storage sites and does talk about storage sites
`dynamically pulling the content from the content provider. So we believe
`that Speedera does have CDN storage sites.
`If you look again at the picture on slide 23, slide 23 is really
`what appears to be what patent owner is arguing, as Speedera explains
`the web server and the web cache operate collectively as a single content
`delivery machine. Our response to that is at slide 24, that if -- so this is
`that Speedera explains that the web server and the web cache operate
`
`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
`
`
`
`
`
` 9
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`collectively as a single machine, but the answer is if you look down
`below, this is content delivery issues include the POP. So basically in
`this picture, you can see that they are actually independent machines.
`They are not a single content delivery machine, because if you look at the
`color, the web server is red, the web cache is blue, and they are separate
`machines.
`If you go to slide 25, the patent owner expert, we believe, is
`consistent with our definition because he answered the question that said
`they could be different software modules running on a single server or
`they could be multiple machines, separate machines.
`JUDGE MELVIN: Is that the same? Could be on different
`machines, it's the same as disclosing that they are on different machines?
`MR. SCHINDLER: The answer is absolutely. This is their
`expert. We agree that you are right, Your Honor, you could argue that
`that means that he didn't positively say they are on separate machines.
`But it's our position through our expert that they are on separate
`machines.
`And if you go to slide 26, they appear to put in a limitation
`about hierarchal places with regard to this idea of CDN storage site. But
`however, if you map it in the way we mapped it, they perfectly match.
`This is matching Speedera to claim 1. They perfectly match.
`And finally, if you go to page 28, this is our position. This is in
`our reply that web server CDN, because they are distinct from content
`server web cache provide storage for the CDN and are located on
`different physical machines.
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`The next term I would like to do is "identifying," which we go
`to slide 50. I want to start off with, this is the term again identifying one
`or more CDN content servers. And if we go to slide 52, we believe this
`is what patent owner is arguing, that the petition fails to show that name
`server identifies the alleged CDN content server. And this is where we
`like to talk about -- the slide that we believe is critical is if you go to slide
`56. So slide 56 is really the idea of -- so you'll see in a lot of patent
`owner's presentations they stay within slide -- Figure 3C. They never
`really go to Figure 4. But Figure 4 supplements Figure 3C because it
`says 550. And this is what happens within 550 that there's an identifying
`that's occurring within Figure 4. And then it goes back because you see
`then step 570 goes back. So this is again a blown-up version of what
`occurs in 550, 560, 570.
`JUDGE MELVIN: So the passage in the specification, that
`same -- the patent owner points to column 10, lines 16 through 22, I think
`it is.
`
`MR. SCHINDLER: Of Speedera.
`JUDGE MELVIN: That's right. And it says that the user
`receives the IP address of POP server 30 which is the larger green box in
`your annotated version.
`MR. SCHINDLER: Right.
`JUDGE MELVIN: And then there's the POP server that routes
`the requests to web cache, right? So it's not exactly the same as
`identifying the web cache server, right?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 11
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`MR. SCHINDLER: Right. But what's going on here is that's
`why Figure 4 supplements what's going on here. The IP address of
`Figures 5C and step 550 is for the web cache server because step 600 is
`described as the initial step. It's exactly what's happening is you can't just
`read Figure 3C. You have to read Figure 3C in context of what's
`occurring in 550, 560 and 570. And when it says requested content
`stored in cache, that's identifying.
`To support that, again, if you go to slide 57, slide 57 is
`Friedman that's supporting exactly what we are talking about.
`And then finally, if you go to Chaudhri, which is slide 59,
`Chaudhri, we take the argument that the patent owner is essentially
`arguing what they call additional steps. And because Chaudhri puts
`additional steps in that therefore, Chaudhri doesn't make it. And it's our
`position that since it's a comprising claim that you can have additional
`steps.
`
`And unless you have further questions, I am good.
`JUDGE MELVIN: Okay. We'll reserve the rest of your time
`for rebuttal.
`MR. BROOKS: If it's all right with the Board, I would like to
`begin by spending two or three minutes on some background on the '178
`patent and then move into the particular dispute starting with the upload
`dispute.
`Could I get slide 3. Here we have column 1 of the patent which
`describes the problem in the prior art. There were two ways for content
`providers to store their content, third-party storage service providers and
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`in-house storage solutions. And the '178 patent notes that both of these
`two structures had significant technical and business-related drawbacks.
`Slide 4, please. So the '178 patent solved these problems with
`the distributed storage system that was integrated into the CDN. The
`system provided persistent storage of network -- of content provider
`content on the CDN thus, allowing the content provider to offload its
`origin server or origin source to the CDN storage system.
`Now, the distributed storage system of the '178 patent has three
`key steps. Go to slide 5, please. The first step is that the content
`provider uploads to the CDN storage system the content that it desires to
`have served by the CDN storage system.
`Go to slide 6. The second step is that the content provider's
`content that has been uploaded to one storage site is then replicated to
`other storage sites within the distributed storage system. If we can go to
`slide 7. And then the last of the key steps is that a CDN content server
`which is also called an entity or an edge server in the '178 patent will
`download content from the storage site in order to service a user request.
`Go to slide 8. This system provides numerous benefits,
`including a scalable solution, highly available system and importantly, it
`reduces the load on the content provider's web servers by allowing the
`content provider to offload its content, offload its origin responsibilities
`to the CDN.
`If we go to slide 19, so I would like to begin with the upload
`limitation. You'll see here on the slide that the claims explicitly require
`receiving via upload. Via upload is a specific term in the claim and it has
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`to be given a meaning. It's not just any transfer of content. The claim
`explicitly says the content is transferred via upload. If upload meant any
`transfer of content, then the claim would just be written receiving and
`storing. But the patentee put in via upload, and we have to give that term
`a meaning.
`So let's look at the specification. Could we go to slide 20,
`please. The specification makes clear that the upload of the claims is a
`content transfer at the instigation of the content provider. The
`specification states that the upload by the content provider is the first step
`of the process. That is, it's not in response to any request. It is the first
`step. The uploaded content is what the content provider desires to
`offload to the CDN storage system. No other entity makes the decision
`to pull the content. It is what the content provider decides to upload to
`the system.
`Can we go to slide 22, please. Now, the specification
`repeatedly contrasts uploading to the storage site by a content provider,
`i.e., pushing content, with downloading from the storage site by a content
`server, i.e., pulling content. And we see this distinction in the abstract
`which is on slide 22. And then on slide 23 we see the same distinction in
`the summary of the invention. And then on slide 24 we also see the
`distinction in other portions of the specification. And in particular there
`on the bottom right of the slide, we see a portion of the specification that
`actually defines download as retrieve. So again, that's consistent with
`download being a pulling of content.
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`Now, in the petition, the petitioner sort of ignored the
`distinction between upload and download or at least didn't address it.
`Then we made our argument in our opposition and our patent owner
`response. And then in their reply, they came back and argued that, well,
`based on one sentence in the specification, upload and download mean
`the same thing. In other words, that the upload language has no
`independent meaning.
`And so I would like to say a few things about the portion of the
`spec they are relying on. Could we go to slide 25, please. So a few
`things about this citation at column 7, lines 6 to 10 of the patent. First,
`the sentence that they are relying on does not itself use the words
`"upload" or "download" at all. The word "upload" is used a few
`sentences above, but it's not actually used in the sentence they are relying
`on.
`
`Secondly, the sentence describes its feature as an alternative to
`what is described above. So it is an alternative to the uploading that is
`described above. And in other words, it's an alternative that applies, as it
`says, when the -- if the content were not otherwise already available to
`the distributed storage system. So what it's saying is in the event that the
`content has not been uploaded by the content provider, then as a last
`resort, the system will try to go back to the content provider and
`download that content from the content provider.
`JUDGE MELVIN: What is the first alternative? This says
`another alternative, right? So what is the first alternative?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`
`
` 15
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`
`MR. BROOKS: The first alternative is the uploading. The first
`choice or the first alternative that is disclosed is the uploading. And then
`another alternative, if the upload doesn't happen, another alternative for
`getting the content is to do this downloading, this dynamic pulling.
`Now, the third thing to note that sort of confirms this
`interpretation is if we look at column 8, lines 1 through 3 of the patent,
`blow that up, please. So this is in the next column and it says, In an
`alternative embodiment, once the redirect threshold is met, the missing
`file may be dynamically downloaded from the origin server. So this is
`essentially describing the same sort of functionality we see in column 7,
`lines 6 to 10. But instead of dynamically pull, it says dynamically
`download. So I think if you read these in combination, it's confirming
`that the patent sees pulling as download.
`And so fourth, it should also be noted that the claims don't
`actually cover -- if we could go back to slide 25, please. They don't
`actually cover the embodiment that petitioner is pointing to. Under the
`claims, the content has to be uploaded to the storage site. And when the
`content server goes to the storage site to try to get that content, the
`content is there. That is what the claim says. It retrieves the storage
`server -- the content server retrieves the content from the storage site. In
`this alternative the storage site doesn't have the content. So this is an
`alternative presented in the patent, but it's an unclaimed alternative.
`And the last thing I'll note is that the petitioner put in some
`dictionary definitions, and one of those -- you know, one of those
`dictionary definitions, Exhibit 1021 at 5, actually says it defines
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`download as, quote, the opposite of upload. So their attempt to read the
`spec as saying download and upload mean the same thing is inconsistent
`with their own dictionaries that they put in.
`I would like to address those dictionaries --
`JUDGE ANDERSON: Wait a minute, counsel. So are you
`saying we need to construe upload or did you construe upload? It doesn't
`appear you did. Did you?
`MR. BROOKS: We did not do like a formal claim
`construction --
`JUDGE ANDERSON: Doesn't that seem important? Your
`whole argument seems to be based on it, so isn't that important?
`MR. BROOKS: Yes. Deciding what upload means is
`important, yes, Your Honor. So --
`JUDGE ANDERSON: So what does it mean? What is your
`proposed construction of upload?
`MR. BROOKS: Upload is a transfer of content at the
`instigation of the content provider.
`JUDGE ANDERSON: And that's based on what?
`MR. BROOKS: That's based on the specification evidence that
`I have gone through and also the testimony of Dr. Burger, our expert.
`JUDGE ANDERSON: I'm looking at the claim. Can you look
`at the claim and tell me why it says that the -- why it supports that upload
`goes from the content provider to the -- let me get the term right here. So
`the claim, we are looking at sub element B, right, and C, I guess, to the
`extent it's the same. The first participating content provider is what we
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`are talking about. And then we go down and we are looking at part III,
`right?
`
`MR. BROOKS: Yes, Your Honor.
`JUDGE ANDERSON: So receiving via upload and storing in
`the first content storage directory at the storage site.
`MR. BROOKS: Yes, Your Honor.
`JUDGE ANDERSON: Where is that upload coming from,
`according to the claim?
`MR. BROOKS: So the claim uses the term "upload" and
`indicates that the receipt has to be by upload. And based on the
`specification and its description of uploading, it's clear that the upload is
`happening from an entity to another entity. And the entity that the upload
`is happening from is the content provider. That's clear from the
`specification. That's our position.
`JUDGE ANDERSON: Well, is it clear from the claim?
`MR. BROOKS: Well, the claim uses the term "upload." So we
`have to determine what upload means. And it has to mean -- presumably
`means something different from download. So our position is based on
`the specification, it means a transfer at the instigation of the content
`provider.
`JUDGE ANDERSON: Okay. So why doesn't Gupta or
`Speedera show uploading?
`MR. BROOKS: So with Gupta, can we go to slide 27, so
`applying our definition, Gupta doesn't show uploading because it's clear
`that when Gupta doesn't have the content cached on its web server,
`
`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
`
`
`
`
`
` 18
`
`
`
`Case IPR2016-01894
`(Patent 7,472,178 B2)
`
`Gupta's web server requests the content from the customer's origin server
`and the origin server responds to that request by sending the content. So
`the content is coming at the instigation of the web server or what they
`call the CDN. Not the content provider. And that's supported by our
`expert at paragraph 68 of his declaration. And then as to Chaudhri --
`JUDGE MELVIN: That's only if the content is not already
`
`there.
`
`MR. BROOKS: Yes, Your Honor. If the content is already
`cached there, then it's -- there's no retrieval. There's no upload or
`download. It's already there.
`JUDGE MELVIN: How did it get there?
`MR. BROOKS: Well, there's no disclosure that it' been
`uploaded in Gupta. And they don't point to any actual disclosure in
`Gupta that it's been uploaded. I should take that back. I think in their
`reply, not in their petition, but in their reply I think they point to a vague
`statement in Gupta that the content provider provides the content. But
`there's no indication that the content provider is doing that at its own
`instigation rather than what is actually disclosed for sure is that the
`content provider provides content in response to a request in Gupta.
`Just to follow on my answer to Judge Anderson, you asked
`about Chaudhri as well. Can we go to slide 29, please. So Chaudhri is
`retargetters, which are what they call the storage sites. They also receive
`content via download rather than upload. And this is shown by the fact
`that when Chaudhri's retargetter does not have content, it, quote, requests
`the content from the publisher's servers and, quote, retrieves the content
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`from the publisher's servers. And that's disclosed at paragraphs 51 and
`80 of Chaudhri. So this content retrieval is at the instigation of the
`retargetter. Not the content provider. This is Dr. Burger makes this the
`point in paragraph 97 of his declaration. And the '178 patent as well
`makes clear at column 8, lines 10 to 12 that retrieving content is a
`download. Not an upload. And I don't believe petitioner has any expert
`testimony saying anything to the contrary.
`If I could go back for a moment just to talk about the dictionary
`definitions they put in, just a few words on those dictionaries. First of
`all, their dictionary definitions conflict with their argument from the spec
`that the terms mean the same thing. Their dictionaries say, well, an
`upload is from small computer to large computer and a download is from
`large computer to small computer. Whereas, their spec argument is that
`the terms mean the same thing.
`Secondly, their argument actually conflicts with the
`specification. Their dictionary argument actually conflicts with the
`specification. And what I mean by that is, if we could go back to column
`8, lines 1 to 3, so here the patent is describing a case in which the content
`is not found on the storage sites and therefore, there's an attempt to
`download the content from the origin server. And in this circumstance,
`the content transfer would be from the content provider to the storage
`site. And under their definition, at least as I think they are arguing, that
`would be a transfer from a small computer to a large computer. And yet,
`here in the '178 patent itself, it's describing such a transfer as a download.
`
`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 IPR2016-01894
`(Patent 7,472,178 B2)
`
`So that's directly contrary to the dictionary definitions that they have put
`in with their reply.
`JUDGE ANDERSON: Doesn't this claim contemplate that
`somewhere someone is requesting content from the content delivery -- I
`got to get the claim terms right here. When you read part B2 of claim 1
`which says, Providing the first participating content provider with a host
`name and a provider portion and a second portion, so that's identifying
`some kind of content at that -- that that content provider has. And if you
`look at the end of the claim where it says that -- where it describes that
`second portion is common to each of the first and second CDN-specific
`host claims, it appears to me that the whole reason for this is that