Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 1 of 14 PageID #: 47850
`
`ACCELERATION BAY LLC,
`
`
`
`
`
`ELECTRONIC ARTS INC.,
`
`
`
`
`
`
`
`
`
`
`
`v.
`
`
`
`IN THE UNITED STATES DISTRICT COURT
`FOR THE DISTRICT OF DELAWARE
`
`)
`)
`)
`)
`)
`)
`)
`)
`)
`
`
`Plaintiff,
`
`Defendant.
`
`
`
`
`
`C.A. No. 16-454 (RGA)
`
`
`
`
`
`
`
`
`
`DEFENDANT’S REPLY BRIEF IN SUPPORT OF ITS MOTION
`FOR SUMMARY JUDGMENT OF NON-INFRINGEMENT
`BY COLLATERAL ESTOPPEL
`
`MORRIS, NICHOLS, ARSHT & TUNNELL LLP
`Jack B. Blumenfeld (#1014)
`Cameron P. Clark (#6647)
`1201 North Market Street
`P.O. Box 1347
`Wilmington, DE 19899
`(302) 658-9200
`jblumenfeld@morrisnichols.com
`cclark@morrisnichols.com
`
`Attorneys for Defendant
`
`
`
`
`
`
`OF COUNSEL:
`
`David P. Enzminger
`Michael A. Tomasulo
`Gino Cheng
`Joe S. Netikosol
`WINSTON & STRAWN LLP
`333 South Grand Avenue, 38th Floor
`Los Angeles, CA 90071
`(213) 615-1700
`
`Louis L. Campbell
`WINSTON & STRAWN LLP
`255 Shoreline Drive, Suite 520
`Redwood City, CA 94065
`(650) 858-6500
`
`Joseph C. Masullo
`WINSTON & STRAWN LLP
`1700 K Street, N.W.
`Washington, DC 20006
`(202) 282-5000
`
`January 7, 2022
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 2 of 14 PageID #: 47851
`
`
`
`I.
`
`II.
`
`
`
`
`TABLE OF CONTENTS
`
`EA Has Demonstrated That Summary Judgment of Collateral Estoppel of Non-
`Infringement Based on Take-Two Is Warranted ................................................................. 1
`
`Page
`
`
`A.
`
`B.
`
`C.
`
`EA Satisfies the Collateral Estoppel Standard ........................................................ 1
`
`Acceleration’s Supposed “Application Layer Overlay” Network Does Not
`Infringe .................................................................................................................... 4
`
`Any Distinctions Between EA’s Voice Squelching and Sub-Channels Features
`and the Take Two Games Are Immaterial .............................................................. 5
`
`1.
`
`2.
`
`Voice Squelching Feature Is an Immaterial Distinction ............................. 5
`
`PvZ Voice Sub-Channel Feature Is an Immaterial Distinction .................. 7
`
`In Addition to Summary Judgment of Collateral Estoppel, the Court May Also Grant
`Summary Judgment of Non-Infringement in the First Instance ......................................... 9
`
`i
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 3 of 14 PageID #: 47852
`
`
`
`Cases
`
`TABLE OF AUTHORITIES
`
`Page(s)
`
`Acceleration Bay LLC v. Take-Two Interactive Software, Inc.,
`No. 16-455-RGA, 2020 WL 1333131 (D. Del. Mar. 23, 2020) ...................................... passim
`
`Medtronic, Inc. v. Mirowski Fam. Ventures, LLC,
`571 U.S. 191 (2014) ...................................................................................................................9
`
`
`
`
`ii
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 4 of 14 PageID #: 47853
`
`
`
`I.
`
`EA Has Demonstrated That Summary Judgment of Collateral Estoppel of Non-
`Infringement Based on Take-Two Is Warranted
`
`Acceleration’s opposition never addresses the central issue raised in EA’s motion:
`
`collateral estoppel of non-infringement is required because, both in Take-Two and here,
`
`Acceleration does not dispute that the accused network is a client-server network. Instead,
`
`Acceleration offers the same theories for why Acceleration’s supposed “application layer overlay
`
`network” is an m-regular and incomplete network.1 Those theories failed in Take-Two, and they
`
`also fail here for the same reasons. Acceleration now offers three main counterarguments for why
`
`those theories are not barred by collateral estoppel: 1) proving collateral estoppel is a heavy burden,
`
`2) EA focuses on the wrong layer, 3) EA’s games are different. These arguments all fail.
`
`A.
`
`EA Satisfies the Collateral-Estoppel Standard
`
`It does not matter that collateral estoppel imposes a high bar because EA has cleared it.
`
`The accused Take-Two NBA 2K game and all accused EA games use a client-server network, and
`
`for each, Acceleration contends that there is an “application layer overlay” network overlaid on
`
`the client-server network that is purportedly m-regular and incomplete. It is not true that there is
`
`an overlay network in either case, but that is beside the point. The “application overlay network”
`
`Acceleration conjures up did not meet the claim limitations in Take Two, and it does not here for
`
`indistinguishable reasons. In Take-Two, Acceleration’s expert said the relay server was a
`
`participant in the accused network. Acceleration Bay LLC v. Take-Two Interactive Software, Inc.,
`
`No. 16-455-RGA, 2020 WL 1333131 (D. Del. Mar. 23, 2020) (“Take-Two”) at *9
`
`(“Dr. Mitzenmacher, Plaintiff’s own expert, wrote that the relay servers ‘are participants’”). Here
`
`
`1 Despite its quibbling at the last status conference over whether the Court had granted summary
`judgment on the ’497 patent, Acceleration confirms it is not moving forward with that patent.
`D.I. 582 at 2, n1. Either way, the ’497 patent is out of this case. All that remains are Acceleration’s
`theories that EA infringed the ’344 and ’966 patents by internal testing of the NHL and PvZ games
`and that the accused games infringe the ’147 patent under the doctrine of equivalents.
`
`1
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 5 of 14 PageID #: 47854
`
`
`
`too, Acceleration’s expert has said that the server is a participant in the accused network for the
`
`FIFA, NHL and PvZ games:
`
`
`
`
`
`
`
`
`
`
`
`
`
`D.I. 428, Ex. A-1 (Medvidović opening report) ¶ 2 (highlighting added).
`
`D.I. 428, Ex. A-1 (Medvidović opening report) ¶ 79 (highlighting added).
`
`D.I. 428, Ex. A-1 (Medvidović opening report) ¶ 86 (highlighting added).
`
`D.I. 428, Ex. A-1 (Medvidović opening report) ¶ 93 (highlighting added).
`
`Q. So I’d like to go to Paragraph 79.
`A. Of which exhibit?
`Q. Of your opening report. Sorry.
`A. Thank you.
`Q. So we’ve looked at most of this already, but in the sixth line of
`that paragraph—actually, I’ll start at the fifth line so it’s a
`
`2
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 6 of 14 PageID #: 47855
`
`
`
`complete thought:
`“The DirtyCast server is part of the underlying network and is also
`a participant in the application overlay FIFA Network when it
`applies game logics and selectively forwards data as I explain
`further below.”
`By using the word “when,” are you trying to say that the
`DirtyCast server is sometimes a participant and sometimes not?
`A. No. It’s always a participant, but the way it participates is by
`applying the game logics and selectively forwarding data.
`
`D.I. 434, Ex. E-9 (Medvidović deposition tr.) at 93:3-23.
`
`Accepting the word of Acceleration’s expert that the DirtyCast server is a participant, the
`
`network cannot be m-regular because, just like in Take-Two, one participant (the server) connects
`
`to all the participants while the other participants are each connected to fewer participants. See
`
`Take-Two at *9 (“the network is not m-regular because one participant (the server) is connected to
`
`a different number of neighbors than the other participants are.”). To avoid this outcome,
`
`Acceleration’s attorneys simply contradict Acceleration’s experts to argue that the server is not a
`
`participant. Take-Two at *9 (“Plaintiff counters that the server is not a participant in the game.”);
`
`D.I. 582 at 12 (“The DirtyCast server is not a participant in the application layer network in the
`
`relevant sense”). That did not work in Take-Two and it does not work here.
`
`Acceleration tries to brush away the disagreement between its attorneys and its experts as
`
`a “semantic” argument. D.I. 582 at 12. But it’s not. This argument goes to the very crux of
`
`Acceleration’s case: what are the participants and how do they form an m-regular network? For
`
`Acceleration to win on infringement at trial, it would need to: 1) take a position on whether the
`
`server is a participant, 2) that is supported by its experts, 3) that makes the network m-regular and
`
`4) that also makes the network non-complete. No matter what position Acceleration’s lawyers take
`
`on whether the server is a participant, it cannot satisfy all three of the other points. If the server is
`
`a participant, Acceleration will have the support of its experts, but cannot show an m-regular
`
`network because the server is connected to all players and therefore necessarily has more
`
`3
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 7 of 14 PageID #: 47856
`
`
`
`connections than the player participants do and cannot meet the requirement of an m-regular,
`
`incomplete network. If the server is not a participant, Acceleration will violate the claim
`
`construction order on the meaning of “connection,” would have no expert support, and cannot
`
`show an incomplete network. Acceleration’s infringement allegation is a four-legged table that can
`
`never sit flush on the floor. Acceleration faced this exact same dilemma in Take-Two. Take-Two
`
`at *9 (“Plaintiff counters that the server is not a participant,” but “Plaintiff’s own expert wrote that
`
`the relay servers ‘are participants.’”). And this Court concluded the only resolution was a finding
`
`of non-infringement of the m-regular limitation. Id. (“[T]here is no genuine dispute that the servers
`
`are participants in the NBA 2K networks, and the networks do not literally meet the m-regular
`
`limitation.”). Accordingly, the issue sought to be precluded in this case is the same exact issue
`
`from Take-Two. After Take-Two, the Court should thus grant EA summary judgment of collateral
`
`estoppel of non-infringement.
`
`B.
`
`Acceleration’s Supposed “Application Layer Overlay” Network Does Not
`Infringe
`
`The motion address Acceleration’s supposed “Application Layer Overlay” Network and
`
`shows how it does not satisfy the claims, even as Acceleration asserts it. EA does not focus on a
`
`different layer than Acceleration. EA is focused on Acceleration’s alleged application layer
`
`overlay network, just as in Take-Two. D.I. 581 at 8 (“For purposes of this motion—just as Take-
`
`Two did on its motion for summary judgment—EA will analyze only Acceleration’s alleged
`
`application layer overlay network.”). EA’s opening brief only cites the reports and testimony of
`
`Acceleration’s experts. Accepting that evidence as true, Acceleration’s supposed application layer
`
`overlay network does not show infringement under its own terms.
`
`4
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 8 of 14 PageID #: 47857
`
`
`
`C.
`
`Any Distinctions Between EA’s Voice-Squelching and Sub-Channels Features
`and the Take-Two Games Are Immaterial
`
`The “voice squelching” feature of all the accused EA games and the voice sub-channel
`
`feature in accused PvZ games, the only two aspects of the accused EA games discussed in
`
`Acceleration’s opposition, are distinctions from the Take-Two games that make no difference and
`
`are immaterial. Neither feature creates infringement, because Acceleration’s contentions depend
`
`on the same arguments rejected in Take-Two.
`
`1.
`
`Voice-Squelching Feature Is an Immaterial Distinction
`
`Voice squelching is a feature that prevents too many players from talking at once so that
`
`individual voices can be heard by the other players. Ordinarily, when a player talks, the player’s
`
`console sends the player’s voice data to the server, which forwards the voice data to the other
`
`players. But if more than four players speak at the same time, the server selects four players and
`
`only forwards the voice data from the selected four to all the other players. D.I. 428, Ex. A-1
`
`(Medvidović opening report) ¶¶ 154, 168, 177, 193 (“When [the] DirtyCast [server] detects that
`
`there are four VoIP streams being directed to a participant, it will prevent additional VoIP data
`
`from being sent to that participant at that time. EA refers to this process as ‘voice squelching.’”)
`
`(internal citations omitted). Acceleration contends that this makes the network “4-regular, because
`
`each player only has 4 voice data connections, irregardless [sic] of the players’ actions.” D.I. 582
`
`at 13.
`
`There are two main problems with this theory. First, this feature is not “irregardless of the
`
`players’ actions.” Voice squelching only activates if more than four players happen to decide to
`
`speak at the same time. Thus, just like Acceleration’s infringement arguments for the “proximity
`
`connection rules” of GTAO in Take-Two, this does not infringe the m-regular limitation because
`
`“the players’ actions determine how connections are formed, and the network is not ‘configured
`
`5
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 9 of 14 PageID #: 47858
`
`
`
`to maintain’ any particular state.” Take-Two at *8. Just as in Take-Two, the system is not
`
`configured to try to make such connections by requiring four or more players to talk at the same
`
`time. Whether and when to speak is entirely driven by the individual players; the system does not
`
`induce more than four players to speak so that the server would squelch any voices.
`
`But, even setting this aside, a second problem with Acceleration’s theory is that even the
`
`voice-squelching feature does not meet the m-regular claim limitation. The Court construed
`
`“connection” as “connection between two participants [computers], with no other participants
`
`[computers] in between, through which data can be sent and received.” D.I. 352 at 7. Here, all
`
`communication flows to and from the server participant (identified as such by Acceleration’s own
`
`experts), which applies the voice-squelching rules. See D.I. 428, Ex. A-1 (Medvidović opening
`
`report) ¶¶ 154, 168, 177, 193. Thus, the same problem arises as in Take-Two: the server connects
`
`to all player participants, while the players connect to some fewer number. This, too, makes the
`
`network not m-regular.
`
`And even ignoring the server connections, there is a third problem with Acceleration’s
`
`theory that negates infringement. Under Acceleration’s theory—contrary to its own experts—that
`
`the server is not a participant, the four chosen players directly connect to every other player
`
`because they send their voice data to every other player. But the unchosen players cannot send
`
`voice data to any other player and cannot receive voice data from any other player. That means the
`
`unchosen players have four connections (to the four chosen players), but the chosen players have
`
`connections to all players (chosen and unchosen), which is more than four. That would also make
`
`the network not m-regular, because some of the participants have a different number of
`
`connections.
`
`6
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 10 of 14 PageID #: 47859
`
`
`
`2.
`
`PvZ Voice Sub-Channel Feature Is an Immaterial Distinction
`
`Acceleration’s PvZ voice sub-channels theory fares no better. Under this theory,
`
`Acceleration contends that players “are directly connected only to the other members of [their]
`
`squad and not to all of the other participants in the game.” D.I. 582 at 6-7. When the squads have
`
`the same number of players, Acceleration alleges an m-regular network. But, as with voice
`
`squelching, Acceleration’s experts have acknowledged that the server implements the sub-channel
`
`functionality. D.I. 428, Ex. A-1 (Medvidović opening report) ¶ 218 (“The DirtySDK and DirtyCast
`
`sever also limit the number of players that can access certain VoIP Sub-Channels.”). This means
`
`that for Acceleration to argue that the sub-channels create an m-regular network, it must
`
`acknowledge that the server is a participant, just as with voice squelching, and just as with the
`
`NBA 2K relay server in Take-Two. See Take-Two at *9. This, in turn, means that—just as with
`
`voice squelching and the NBA 2K relay server in Take-Two—the server connects to all of the other
`
`participants, but those other participants connect to fewer participants, so the network is not m-
`
`regular.
`
`But an even more fundamental flaw exists in Acceleration’s sub-channels theory: EA does
`
`not infringe the “broadcast channel” limitation of all the remaining asserted claims. D.I. 1, Ex. 1
`
`(’344 Patent), 30:5, 30:8; D.I. 1, Ex. 2 (’966 Patent), 30:37, 30:40, D.I. 1, Ex. 3 (’147 patent),
`
`28:53. The sub-channels exist so that voice communications remain confidential to that team.
`
`Indeed, as Acceleration characterizes it, the players are only connected to members of their squad
`
`“and not to all of the other participants in the game.” D.I. 582 at 6-7. That is, the sub-channels
`
`ensure that the voice data from players on one team stays with that team and is not shared with the
`
`players on the other team (i.e., not broadcast to all participants):
`
`Q. And what are these three different channels used for?
`A. The Plants and the Zombies channels are team channels for
`each team for their respective sides.
`
`7
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 11 of 14 PageID #: 47860
`
`
`
`The common channel is used at the end of round and we put all the
`users at the—in the common channel so that they can discuss how
`the game went.
`Q. So, the team channels would allow just members of one team to
`communicate; is that correct?
`A. That is how—the best intended usage of the team channels,
`right.
`Q. Do the, like, for example, the plants team would only hear other
`players on the plants team during gameplay; is that right?
`A. Yes—
`MR. TOMASULO: Objection.ꞏ Form.
`THE WITNESS: A. Yes, they—they generally do, yes.
`
`D.I. 433, Ex. E-5 (Lo Tr.) at 33:3-23 (emphasis added). The benefits of confidential
`
`communications are obviously helpful in a game where one team is coordinating an assault on the
`
`other team. If the defense team could hear the plans of offense team, they could thwart those plans.
`
`Thus, these sub-channels are not “broadcast channels” as required by the claim and Acceleration’s
`
`infringement theory fails for this additional reason. See D.I. 361 at 3 (construing “broadcast
`
`channel(s)” as “communications network consisting of interconnected [participants / computers]
`
`where each [participant / computer] receives all data broadcasted on that communication
`
`network”) (emphasis added). Under Acceleration’s theory, the players on the other squad do not
`
`receive all data broadcasted on that communication network as their experts conceded.
`
`Putting aside
`
`the shortcomings with Acceleration’s
`
`literal
`
`infringement
`
`theory,
`
`Acceleration failed to address prosecution history estoppel, which sinks its equivalents
`
`infringement theory. See D.I. 582 at 13-14. Because the applicants added “m-regular” to the claims
`
`by prosecution amendment for the ’344, ’966, and ’147 patents—the only remaining patents
`
`asserted against EA—prosecution history estoppel forecloses any doctrine of equivalents theory
`
`for the “m-regular” limitation in this case. See Take-Two at *9 (Acceleration’s doctrine of
`
`equivalents theory was “especially weak for the ’344, ’966, and ’147 patents because the patentee
`
`8
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 12 of 14 PageID #: 47861
`
`
`
`added the m-regular limitation during prosecution.”). Acceleration offers no reason why this
`
`holding from Take-Two does not apply to EA as a result of collateral estoppel.
`
`II.
`
`In Addition to Summary Judgment of Collateral Estoppel, the Court May Also Grant
`Summary Judgment of Non-Infringement in the First Instance
`
`For the reasons discussed above, the Court should grant summary judgment of collateral
`
`estoppel of non-infringement based on Take-Two. Alternatively, as discussed in EA’s opening
`
`brief, should the Court decide collateral estoppel does not apply, the Court may also grant summary
`
`judgment of non-infringement in the first instance because Acceleration cannot meet its burden of
`
`establishing infringement of the claims.
`
`Acceleration’s opposition brief offers no good reason not to grant summary judgment of
`
`non-infringement. Acceleration argues that the Court should not grant summary judgment because
`
`EA has already asked for summary judgment before. D.I. 582 at 15. It also argues that summary
`
`judgment of non-infringement is inappropriate because EA’s experts did not review all the source
`
`code that Acceleration’s experts did. Id. Both arguments are immaterial. Acceleration has the
`
`burden of proof on infringement. Medtronic, Inc. v. Mirowski Fam. Ventures, LLC, 571 U.S. 191,
`
`191 (2014) (“A patentee ordinarily bears the burden of proving infringement.”). EA’s motion
`
`establishes that Acceleration and its experts cannot prove infringement because Acceleration’s
`
`own experts agree the EA games use a client-server network through their identification of the
`
`network participants, and because Acceleration’s attorney argument that the server is not a network
`
`participant should be rejected for the same reason the Court rejected it in Take-Two. It does not
`
`matter that the server is not executing slap shots, shooting zombies, or scoring goals—all network
`
`communication goes through the server. The server connects to all participants, and therefore has
`
`more connections than any player—even under Acceleration’s theory. Acceleration cannot
`
`establish that a client-server network infringes these patents because it is mathematically
`
`9
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 13 of 14 PageID #: 47862
`
`
`
`impossible for a network to be both m-regular and incomplete if one participant connects to all
`
`others. Because Acceleration cannot carry its burden, the Court should grant summary judgment.
`
`MORRIS, NICHOLS, ARSHT & TUNNELL LLP
`
`/s/ Jack B. Blumenfeld
`_______________________________________
`Jack B. Blumenfeld (#1014)
`Cameron P. Clark (#6647)
`1201 North Market Street
`P.O. Box 1347
`Wilmington, DE 19899
`(302) 658-9200
`jblumenfeld@morrisnichols.com
`cclark@morrisnichols.com
`
`Attorneys for Defendant
`
`
`
`
`
`OF COUNSEL:
`
`David P. Enzminger
`Michael A. Tomasulo
`Gino Cheng
`Joe S. Netikosol
`WINSTON & STRAWN LLP
`333 South Grand Avenue, 38th Floor
`Los Angeles, CA 90071
`(213) 615-1700
`
`Louis L. Campbell
`WINSTON & STRAWN LLP
`255 Shoreline Drive, Suite 520
`Redwood City, CA 94065
`(650) 858-6500
`
`Joseph C. Masullo
`WINSTON & STRAWN LLP
`1700 K Street, N.W.
`Washington, DC 20006
`(202) 282-5000
`
`January 7, 2022
`
`10
`
`

`

`Case 1:16-cv-00454-RGA Document 586 Filed 01/07/22 Page 14 of 14 PageID #: 47863
`
`
`
`
`
`CERTIFICATE OF SERVICE
`
`I hereby certify that on January 7, 2022, I caused the foregoing to be electronically filed
`
`with the Clerk of the Court using CM/ECF, which will send notification of such filing to all
`
`registered participants.
`
`
`
`I further certify that I caused copies of the foregoing document to be served on January 7,
`
`2022, upon the following in the manner indicated:
`
`Philip A. Rovner, Esquire
`Jonathan A. Choa, Esquire
`POTTER ANDERSON & CORROON LLP
`1313 North Market Street, 6th Floor
`Wilmington, DE 19801
`Attorneys for Plaintiff
`
`Paul J. Andre, Esquire
`Lisa Kobialka, Esquire
`James Hannah, Esquire
`Hannah Lee, Esquire
`Yuridia Caire, Esquire
`Greg Proctor, Esquire
`KRAMER LEVIN NAFTALIS & FRANKEL LLP
`990 Marsh Road
`Menlo Park, CA 94025
`Attorneys for Plaintiff
`
`Aaron M. Frankel, Esquire
`Marcus A. Colucci, Esquire
`Cristina Martinez, Esquire
`KRAMER LEVIN NAFTALIS & FRANKEL LLP
`1177 Avenue of the Americas
`New York, NY 10036
`Attorneys for Plaintiff
`
`
`VIA ELECTRONIC MAIL
`
`VIA ELECTRONIC MAIL
`
`VIA ELECTRONIC MAIL
`
`
`
`/s/ Jack B. Blumenfeld
`______________________________________
`Jack B. Blumenfeld (#1014)
`
`
`
`