`
`
`EXHIBIT 1020
`
`
`
`
`
`Trials@uspto.gov
`571.272.7822
`
`
`Paper No. 32
`Filed: November 1, 2017
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`_______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_______________
`
`TALARI NETWORKS, INC.,
`Petitioner,
`
`v.
`
`FATPIPE NETWORKS INDIA LIMITED,
`Patent Owner.
`____________
`
`Case IPR2016-00976
`Patent 6,775,235 B2
`____________
`
`
`
`Before STACEY G. WHITE, MICHELLE N. WORMMEESTER, and
`CHRISTA P. ZADO, Administrative Patent Judges.
`
`WHITE, Administrative Patent Judge.
`
`
`
`
`
`
`
`FINAL WRITTEN DECISION
`35 U.S.C. § 318(a) and 37 C.F.R. § 42.73
`
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 1
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`I. INTRODUCTION
`
`A. Background
`Talari Networks, Inc. (“Petitioner”) filed a Petition (Paper 1, “Pet.”)
`seeking to institute an inter partes review of claims 4, 5, 7–15, and 19 of
`U.S. Patent No. 6,775,235 B2 (Ex. 1001, “the ’235 patent”) pursuant to
`35 U.S.C. §§ 311–319. FatPipe Networks India Limited. (“Patent Owner”)
`filed a Preliminary Response. Paper 6 (“Prelim. Resp.”). Based on our
`review of these submissions, we instituted inter partes review of claims 4, 5,
`7–15, and 19 on the following specific grounds:
`Reference(s)
`Basis Claims Challenged
`Karol1
`§ 102 4, 5, 7–11, 14, and 19
`Karol
`§ 103 4, 5, 7–15, and 19
`Karol and Stallings2
`§ 103 5, 11–15, and 19
`
`
`Paper 7 (“Dec.”), 22. Patent Owner filed a Patent Owner’s Response (Paper
`22, “PO Resp.”), and Petitioner filed a Reply (Paper 26, “Reply”). An oral
`hearing was held on August 14, 2017. Paper 31 (“Tr.”).
`We have jurisdiction under 35 U.S.C. § 6. This Final Written
`Decision is issued pursuant to 35 U.S.C. § 318(a) and 37 C.F.R. § 42.73.
`For the reasons discussed below, Petitioner has demonstrated by a
`preponderance of the evidence that claims 4, 5, and 7–15 of the ’235 patent
`are unpatentable. Petitioner has not meet its burden to establish the
`unpatentability of claim 19.
`
`
`1 U.S. Patent No. 6,628,617 B1 (“Karol,” Ex. 1006).
`2 William Stallings, Data and Computer Communications, Prentice-Hall, 5th
`Ed, 1997, ISBN-81-203-1240-6 (“Stallings,” Ex. 1011).
`
`2
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 2
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`B. Related Proceedings
`The parties inform us that FatPipe, Inc. v. Talari Networks, Inc., No.
`5:16-CV-54-BO (E.D.N.C.) and FatPipe, Inc. v. Viptela, Inc., No. DED-1-
`16-cv-00182 (D. Del.), may be impacted by this proceeding. Pet. 1; Paper
`30, 2–3. In addition, Petitioner has a pending petition for inter partes review
`of a related patent, U.S. Patent No. 7,406,048 B2 (“the ’048 patent”)
`(IPR2016-00977). Pet. 2. Viptela, Inc. and Cisco Systems, Inc. also have
`filed petitions seeking inter partes review of various claims of the ’235 and
`’048 patents. Paper 30, 3.
`
`C. The ʼ235 Patent
`The ’235 patent describes a system and method for communicating
`using two or more disparate networks in parallel. Ex. 1001, Abstract. For
`example, an embodiment of this system could be composed of a virtual
`private network (“VPN”) in parallel with a frame relay network. Id. at 1:19–
`24. These parallel networks back each other up in case of failure and when
`both networks are operational their loads are balanced between the parallel
`networks. Id. at Abstract. An embodiment of this system is depicted in
`Figure 10, which is shown below.
`
`3
`
`
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 3
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`Figure 10 depicts an example of the network topology described in the ’235
`patent. Id. at 8:29–30. Two sites 102 transmit and/or receive data from one
`another. Id. at 2:38–40. These sites are connected by two disparate
`networks, Internet 500 and frame relay network 106. Id. at 8:30–32. Each
`location has frame relay router 105 and Internet router 104. Id. at 8:32–33.
`“Access to the disparate networks at site A and site B is through an inventive
`controller 602 at each site.” Id. at 6:34–36. Controller 602 “allows load-
`balancing, redundancy, or other criteria to be used dynamically, on a
`granularity as fine as packet-by-packet, to direct packets to an Internet router
`and/or frame relay/point-to-point router according to the criteria.” Id. at
`9:12–17.
`
`Figure 7 of the ’235 patent is reproduced below.
`
`
`Figure 7 depicts controller 602. Id. at 10:59–60. Controller 602 is
`connected to site 102 via site interface 702. Id. at 10:60–63. Packet path
`selector 704 is hardware or software that determines which path a given
`packet is to travel. Id. at 11:2–6. The criteria used to determine which path
`a packet travels may be based on concerns such as redundancy,
`
`4
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 4
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`load-balancing, or security. Id. at 11:9–63. Controller 602 also has two or
`more network interfaces 706 (at least one per each network for which
`controller 602 controls access). Id. at 11:64–67.
`
`D. Illustrative Claim
`As noted above, we instituted review of claims 4, 5, 7–15, and 19 of
`the ʼ235 patent, of which claims 4, 5, and 19 are independent. Claim 5 is
`illustrative of the challenged claims and is reproduced below:
`5. A method for combining connections for access to multiple
`parallel disparate networks, the method comprising the
`steps of:
`obtaining at least two known location address ranges which
`have associated networks;
`obtaining topology information which specifies associated
`networks that provide, when working, connectivity
`between a current location and at least one destination
`location;
`receiving at the current location a packet which identifies a
`particular destination location by specifying a destination
`address for the destination location;
`determining whether the destination address lies within a
`known location address range;
`selecting a network path from among paths to disparate
`associated networks, said networks being in parallel at
`the current location, each of said networks specified in
`the topology information as capable of providing
`connectivity between the current location and the
`destination location;
`forwarding the packet on the selected network path.
`
`
`II. CLAIM CONSTRUCTION
`In an inter partes review, “[a] claim in an unexpired patent shall be
`given its broadest reasonable construction in light of the specification of the
`patent in which it appears.” 37 C.F.R. § 42.100(b). Under this standard, we
`
`5
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 5
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`construe claim terms using “the broadest reasonable meaning of the words in
`their ordinary usage as they would be understood by one of ordinary skill in
`the art, taking into account whatever enlightenment by way of definitions or
`otherwise that may be afforded by the written description contained in the
`applicant’s specification.” In re Morris, 127 F.3d 1048, 1054 (Fed. Cir.
`1997).
`
`A. “selects between network interfaces on a per-packet basis” (claim 4)
`/“make network path selections on a packet-by-packet basis” (claim 9)
`Patent Owner contends that one of ordinary skill in the art would have
`understood these phrases to mean “for each packet, make[] a discrete choice
`between network paths/interfaces.” PO Resp. 9 (citing Ex. 2003 ¶¶ 34–41).
`Petitioner asserts that if we determine that these phrases need construction
`the proper construction is “for each packet a network interface/path is
`chosen.” Reply 6. The parties dispute whether path selections must occur
`once for each individual packet or whether such selection may be made for a
`group of packets. PO Resp. 10; Reply 2–3.
`First, Patent Owner asks that we construe the select/selections terms
`to mean making a discrete choice between two or more possibilities. PO
`Resp. 9. Petitioner asserts that there is no basis for Patent Owner’s proposed
`“discrete choice” language. Reply 4. We are not persuaded that, for the
`purposes of this Decision, there is meaningful information to be gleaned by
`construing the words select/selections to mean “make a choice.” We also
`are not persuaded that there is a basis for inserting the word “discrete” into
`this construction. Therefore, based on the disputes before us, we see no
`reason to provide an express construction for the common terms
`select/selections.
`
`6
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 6
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`Patent Owner contends that the terms “per” and “packet-by-packet”
`would have been understood to require a selection for each packet. PO
`Resp. 10. Thus, Patent Owner asserts that there must be a path selection
`process performed for each individual packet. Id. Petitioner argues that this
`is too narrow of a view of the claim terms and asserts that the specification
`describes making a single selection that applies to each packet in a group.
`Reply. 2–3.
`In support of their arguments both parties direct us to Figure 9 and its
`supporting text. Figure 9 is reproduced below.
`
`
`Figure 9 “is a flowchart illustrating methods of the present invention for
`combining connections to send traffic over multiple parallel independent
`disparate networks.” Ex. 1001, 5:48–50. The parties direct us to the
`discussion of step 908, which states that
`[d]uring a path selecting step 908, the path selector 704 selects
`the path over which the packet will be sent; selection is made
`between at least two paths, each of which goes over a different
`
`7
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 7
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`network 106 than the other. The disparate networks are
`independent parallel networks. This path selecting step 908
`may be performed once per packet, or a given selection may
`pertain to multiple packets.
`Id. at 14:40–46.
`As described in the specification, Figure 9 depicts “methods” for
`selecting paths. See id. 5:48–50. These methods allow for the selection to
`“be performed once per packet, or a given selection may pertain to multiple
`packets.” Id. at 14:40–46 (emphasis added). Thus, we find that the
`specification discloses both selection for each individual packet and
`selection for a group of packets.
`Next, we examine the claims to see whether claims 4 and 9 cover both
`of the embodiments or whether they are directed only to the embodiment
`wherein the selection occurs for each individual packet. Independent claim
`4 recites, in relevant part, “a packet path selector which selects between
`network interfaces on a per-packet basis.” This stands in contrast to claim 5,
`which recites, in relevant part, “selecting a network path from among paths
`to disparate associated networks.” While independent claim 5 refers to
`“receiving at the current location a packet,” its selection step does not make
`reference to this “packet,” but rather it recites the selection of a network path
`without requiring this selection to be made “per-packet.” We are persuaded
`that this difference is significant in that it evidences that the Patentee
`intended claim 4 to have a more narrow scope than claim 5.
`This distinction is supported by the language of claim 5’s dependent
`claims. Claim 9 depends from claim 5 and further recites “wherein repeated
`instances of the selecting step make network path selections on a packet-by-
`packet basis.” Also, claim 10 depends from claim 5 and further recites
`
`8
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 8
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`“wherein repeated instances of the selecting step make network path
`selections on a per session basis.” Thus, claim 9 narrows the breadth of
`claim 5’s selection step by requiring it to occur “on a packet-by-packet
`basis” and claim 10 narrows the breadth of claim 5 by requiring selection to
`occur on a “per session basis.” Therefore, claim 5 is broad enough to
`encompass selection for more than one packet at a time (a session) and
`selection for an individual packet.
`Petitioner argued that claim 4 must be broad enough to encompass the
`embodiment in which selections were performed for multiple packets
`because it would be improper to construe claim 4 in a manner that would
`exclude that embodiment. Reply 2. We do not agree. The Patentee
`described multiple embodiments in the specification and as such, the
`Patentee was free to determine which embodiments would be encompassed
`by which claims. Here, we are presented with evidence that the Patentee
`drafted claim 5 to cover both embodiments and drafted dependent claims to
`focus on the individual embodiments. Such a drafting choice is within the
`purview of the Patentee and we see no reason why we must construe claim 4
`in a manner that would encompass all embodiments. The Patentee’s choice
`to describe the selection as occurring on a “per packet basis” when viewed in
`light of the specification and the other claims indicates a decision to direct
`claim 4 to the embodiment in which routes are selected for each packet. As
`such, we find that the language of claims 4 and 9 indicates that these claims
`are directed to the embodiment wherein path selection is performed for each
`individual packet. Therefore, we construe the disputed phrases to mean
`“selecting a network path/interface for each packet.”
`
`9
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 9
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`B. “dynamic load-balancing” (claims 11–13)
`Patent Owner asserts that the term “dynamic load-balancing” would
`have been understood by a person of ordinary skill in the art to mean
`“distributing packets based on actual traffic assessed after the packet
`arrives.” PO Resp. 15 (citing Ex. 2003 ¶¶ 42–48). Petitioner responds by
`arguing that Patent Owner’s proposed construction is overly narrow and that
`if we determine that a construction is necessary a proper construction would
`be “sending packets in distributions that balance the load of a given network,
`router, or connection relative to other networks, routers, or connections
`available to the controller without manual intervention.” Reply 7 (citing Ex.
`1001, 9:30–33, 11:21–24, 11:33–38).
`Patent Owner directs us to a passage in the specification that it
`contends supports its position that the load is balanced in response to actual
`traffic. PO Resp. 15
`For instance, a local area network (LAN) at site 1 may be set up
`to send all traffic from the accounting and sales departments to
`router A1 and send all traffic from the engineering department
`to router B1. This may provide a very rough balance of the
`traffic load between the routers, but it does not attempt to
`balance router loads dynamically in response to actual traffic
`and thus is not “load-balancing” as that term is used herein.
`Id. (quoting Ex. 1001, 2:61–65). Patent Owner’s declarant, Joel Williams,
`explains that
`[t]he phrase “as that term is used herein” in this passage
`informs a [person of ordinary skill in the art] that the ’235
`specification imposes constraints on the meaning of the term
`“load balancing,” relative to the way that term was used
`conventionally to describe balancing traffic loads between
`routers. In particular, dynamic load-balancing in the context of
`the patented invention requires that load-balancing is performed
`
`10
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 10
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`on the basis of the actual traffic observed at the time of
`balancing on the available lines.
`Ex. 2003 ¶ 43.
`Patent Owner asserts that actual traffic is determined at some point in
`time after the packet arrives. PO Resp. 16. In support of this assertion,
`Patent Owner directs us to the specification’s disclosure that “in some cases
`the path for the next packet may be determined by the packet path selector
`before the packet arrives, e.g., in a round-robin manner, while in other cases
`the path is determined after the packet arrives, e.g., using per-packet
`dynamic load balancing.” Id. (quoting Ex. 1001, col. 14:53–58) (emphases
`added).
`Petitioner disputes this construction and argues that Patent Owner’s
`view of the claim language is “a blatant, improper attempt to read additional
`limitations into the claims.” Reply 7. According to Petitioner, the cited
`“passage only highlights the difference between the prior art (manual
`switchover) and the alleged invention (no manual intervention). It is
`unrelated to balancing traffic based on traffic loads that existed at the time of
`a packet’s arrival or at some undefined period thereafter.” Id. at 8. Thus,
`Petitioner argues that “[t]he specification uses ‘dynamic load-balancing’ to
`mean that no manual switchover is required.” Id.
`We are not convinced by Petitioner’s arguments. The cited portion of
`the specification discusses a configuration in which a LAN “may be set up to
`send all traffic from the accounting and sales departments to router A1 and
`send all traffic from the engineering department to router B1.” Ex. 1001,
`2:55–61. Petitioner argues that the key point here is that this prior art
`configuration required a manual switchover. Reply 8. There, however, is no
`reason for us to conclude that dynamic load balancing should be equated
`
`11
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 11
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`with either a manual or an automated environment. The cited passage
`distinguished routing by departments from dynamic load balancing by
`stating that the department-based routing is not dynamic because in dynamic
`load balancing the “router loads dynamically in response to actual traffic.”
`Id. at 2:63–64. We find that the cited passages show that the contemplated
`load balancing is performed based on actual traffic, but are insufficient to
`establish that the Patentee intended load balancing to be equated with a
`manual selection process. See Id. at 2:61–65, 7:31–32.
`Further, we are not convinced by Patent Owner’s arguments regarding
`the alleged requirement that traffic be assessed at some point after the
`packet’s arrival. The portion of the specification cited in support of this
`assertion refers to multiple scenarios or “cases.” See PO Resp. 16 (citing
`Ex. 1001, 14:53–58). The specification states that in one of those cases the
`path may be determined after the packet arrives and states that, for example,
`the path determination may be done using dynamic load balancing. See
`Ex. 1001, 14:53–58. This passage does not provide any specific temporal
`limitation to the term “dynamic load balancing,” but rather it illustrates an
`example of when load balancing could be used. It is improper to limit the
`claims to this one exemplary embodiment without more express language in
`the claims or the specification that would narrow the scope of this term.
`Thus, we find that the proper construction of dynamic load balancing is
`“distributing packets based on actual traffic.”
`
`III. ANALYSIS
`
`Petitioner bears the burden of proving unpatentability of the
`challenged claims, and the burden of persuasion never shifts to Patent
`Owner. Dynamic Drinkware, LLC v. Nat’l Graphics, Inc., 800 F.3d 1375,
`
`12
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 12
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`1378 (Fed. Cir. 2015). To prevail, Petitioner must establish the facts
`supporting its challenge by a preponderance of the evidence. 35 U.S.C.
`§ 316(e); 37 C.F.R. § 42.1(d). Thus, we examine the full record in this
`matter to determine whether Petitioner has met its burden under 35 U.S.C.
`§ 316(e).
`
`A. Asserted Ground of Unpatentability over Karol
`Petitioner asserts that claims 4, 5, 7–11, 14, and 19 are anticipated by
`the disclosures of Karol. Pet. 10–30. Petitioner also contends that claims 4,
`5, 7–15, and 19 would have been obvious over the teachings of Karol. Id. at
`42–60. Petitioner supports its arguments with a declaration from Dr. Kevin
`Negus. Ex. 1005.
`1. Overview of Karol
`Karol is directed to “the internetworking of connectionless (e.g.,
`Internet Protocol or ‘IP’) and connection oriented (e.g., ATM, MPLS,
`RSVP) networks.” Ex. 1006, 1:7–10. Connectionless (“CL”) networks
`require no explicit connection setup prior to transmitting datagrams. Id. at
`1:19–24. In contrast, connection oriented (“CO”) networks determine a
`route for the connection and allocate bandwidth resources along the route.
`Id. at 1:31–39. Figure 1 of Karol is reproduced below.
`
`13
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 13
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`
`Figure 1 depicts CO and CL networks in a parallel configuration. Id. at
`4:12–14. Datagrams ultimately destined for endpoint 151 may be sent from
`source 101 to node 111 in CL network 110. Id. at 4:39–40. The source or
`destination may be connected directly to CL-CO gateway 140 or they may
`be connected through a node in the network. Id. at 5:5–8. The datagrams
`may be routed over either the CO or CL network in order to arrive at
`endpoint 151. Id. at 4:40–43. CL-CO gateways 140 and 150 interconnect
`the CL and CO networks and “allow[] datagrams (sometimes hereinafter
`called messages) originated on the CL network to be transported . . . on the
`CO network.” Id. at 3:30–37. “When a datagram arrives at CL-CO gateway
`140 of FIG. 1, a determination is made if that packet should be carried by
`CO network 160.” Id. at 5:23–25. CL-CO gateway 140 is described in more
`detail in Figure 4, which is reproduced below.
`
`14
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 14
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`
`
`Figure 4 illustrates the internal arrangements of CL-CO gateway 140. Id. at
`6:31–32.
`Generally speaking, each CL-CO gateway arranged in
`accordance with the present invention includes hardware and
`software modules that typically comprise (a) a switch fabric for
`CO networking, shown in FIG. 4 as CO switch 410, (b) a CL
`packet forwarding engine, shown in FIG. 4 as CL router/switch
`420, (c) a protocol converter 450, (d) a moderately sized
`packet buffer 440 for temporarily storing packets waiting for
`CO network setup or turnaround; and (e) a processor 430 and
`associated database 431 for controlling the gateway packet
`handling operations and for storing forwarding, flow control,
`header translation and other information. Input line cards 401
`and output line cards 402 connect the gateway of FIG. 4 to
`external networks, such that datagrams received in input line
`cards 401 can be directed either to CO switch 410 or CL
`router/switch 420, and such that output line cards 402 can
`receive datagrams from either of the last mentioned elements
`and direct them to external networks.
`
`Id. at 6:32–50. The elements depicted in Figure 4 are controlled by
`processor 430 and such control is implemented via programs stored in the
`processor. Id. at 6:55–59. The routing procedures used by gateway 140
`may adjust routing dynamically “to divert connections away from
`
`15
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 15
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`overloaded call processors.” Id. at 17:64–67. In other words, routing “can
`be adjusted to reflect bandwidth availability.” Id. at 18:1–2.
`2. Independent Claim 4
`Claim 4 recites a controller, which controls access to multiple
`networks. Petitioner’s arguments as to independent claim 4 may be
`summarized as follows: Petitioner argues in the alternative that the claimed
`controller that provides access to multiple networks may be either Karol’s
`CL-CO gateway alone or the gateway in combination with one or more
`routers or switches. Pet. 10–12. If the controller is the gateway alone, then
`Petitioner asserts that the site interface is disclosed by one or more of
`Karol’s input line cards 401 or the network connection depicted in Figure 1
`between source 101 and node 111. Id. at 12. If the controller is the gateway
`in combination with routers and/or switches, then Petitioner asserts the site
`interface is a network connection. Id. According to Petitioner, Karol
`discloses at least two output line cards 402 that receive datagrams from the
`CO switch or CL router/switch and directs the datagrams to external
`networks. Id. at 12–13. In addition, we note that Karol discloses that “the
`source or destination may be directly connected to a CL-CO gateway (e.g.,
`gateway 140) as opposed to being connected through a CL node.” Ex. 1006,
`5:5–8. As to the packet path selector, Petitioner points to Karol’s gateway
`processor, CL router/switch, CO switch, packet buffer, protocol converter
`and input line cards to disclose this element of the claim. Pet. 14. Petitioner
`asserts that these items work together in Karol to determine if a packet
`(“datagram”) from a source should be forwarded to either the CL or CO
`network. Id. Petitioner relies on Karol’s disclosure of routing datagrams
`based on “‘bandwidth availability’ that can be ‘dynamically allocated to
`
`16
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 16
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`flows on an as-needed basis’ and can ‘divert[] connections away from
`congested links.” Id. at 15 (citing Ex. 1006, 17:18–26, 17:63–18:2;
`Ex. 1005 ¶ 182).
`Patent Owner argues that Karol does not disclose “a packet path
`selector which selects between network interfaces on a per-packet basis.”
`PO Resp. 23. Patent Owner supports its contentions with declarations from
`Joel Williams. Exs. 2001, 2003.
`Specifically, Patent Owner contends that Karol does not disclose the
`selection of paths on a per packet basis as required by claim 4. PO Resp.
`23–28. This argument is based on Patent Owner’s contentions that (1) Karol
`does not select a network when a packet arrives, but rather it routes packets
`based on precomputed routes; and (2) Karol’s path selection occurs
`infrequently and not on a per-packet basis. Id.
`Petitioner asserts that Karol’s system “selects between network
`interfaces on a per-packet basis” (e.g., packet path selector compares
`information in each packet received at the CL-CO gateway to determine if
`the packet will be routed to the CL or CO network interface output line
`card). Pet. 15–16. Petitioner’s declarant, Dr. Kevin Negus, testifies that this
`path selection occurs by examining the packet’s destination, the optional
`presence of alternative paths to that destination, and at least one specified
`criteria for selecting between alternative paths. Ex. 1005 ¶ 187.
`Karol states that “[w]hen a datagram arrives at a CL-CO gateway 140
`of FIG. 1, a determination is made if that packet should be carried by the CO
`network 160.” Ex. 1006, 5:23–25. Patent Owner, however, asserts that this
`determination is not an individualized selection of a route for a specific
`packet, but rather it is a determination as to whether a packet is part of a
`
`17
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 17
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`group of packets (a flow) for which a routing decision previously has been
`made. PO Resp. 24. Figure 5 of Karol is reproduced below.
`
`
`Figure 5 of Karol depicts the packet forwarding process. Ex. 1006, 3:6–8.
`In step 501, a packet arrives at Karol’s CL router/switch 420. Id. at 8:56–58.
`Step 503, then inquires as to whether the received packet is “a packet from a
`flow that needs CO Service.” Id. at Fig. 5, element 503 (emphasis added).
`Thus, we determine that Karol’s routing decisions are made for a flow of
`packets and not for an individual packet.
`As discussed above, we construe claim 4’s “per packet basis” to
`require “selecting a network path/interface for each packet.” See supra
`§ II.A. Thus, we find that Karol does not disclose the per packet selection
`
`18
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 18
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`required in claim 4. Therefore, we determine that Petitioner has not met its
`burden to establish that claim 4 is anticipated by Karol.
`Petitioner, however, also argues that claim 4 would have been obvious
`over Karol. Petitioner expands upon its anticipation arguments with
`arguments in which it explains why Karol combined with the knowledge of
`one of ordinary skill in the art would have rendered claim 4 obvious. Pet.
`42–60.
`Petitioner argues that if we construe “per-packet basis” to require
`selection for each packet then Karol would have rendered this limitation
`obvious. Pet. 45. Petitioner asserts that this limitation would have been
`obvious because (1) modifying Karol to select networks on a per-packet
`basis would have amounted to nothing more than the simple substitution of
`one known element for another to obtain predictable results (id. at 45); and
`(2) the combination of Karol and the knowledge of one of ordinary skill in
`the art would have been obvious to try (id. at 46).
`Petitioner contends that the ’235 patent describes the “prior art [as]
`disclos[ing] routing decisions that are independent of the particular flows or
`sessions of particular packets.” Id. at 45 (citing Ex. 1001, 4:15–23; Ex. 1005
`¶ 192). Petitioner relies upon a passage from the ’235 patent, which
`describes a “prior approach[] for selecting which network to use for which
`packet(s)” in which decisions are made based on the origin of the packet.”
`Ex. 1001, 4:15–23. In contrast, Karol describes a selection process in which
`the gateway makes a determination as to which network should receive the
`packet based on its examination of the fields of Karol’s flow database, which
`includes “source address” as one of its fields. Id. at 46. Petitioner asserts
`that it would have been obvious to modify Karol by limiting the routing
`
`19
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 19
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`decision to an analysis of the packet’s source address. Id. In support of this
`position, Dr. Negus testifies that such a modification would entail “a much
`simpler and known packet path selection process.” Ex. 1005 ¶ 193.
`Patent Owner asserts that Karol’s gateway uses OSPF (Open Shortest
`Path First) to determine routing prior to a packet being received at the
`gateway. PO Resp. 26. Patent Owner further contends that “[t]he OSPF
`routing protocol expressly excludes the possibility of making packet routing
`decisions on a per-packet basis.” Id. at 27. We find this to be an overly
`narrow view of Karol’s disclosures because Karol is not limited to OSPF.
`See Ex. 1006, 14:20–22 (“the description below assumes the OSPF routing
`protocol, the concept is readily applicable to other IP routing protocols”).
`We determine that Petitioner has shown that it would have been
`obvious to modify Karol to select networks on a per packet basis. Petitioner
`has proposed a modification to Karol that is “much simpler” and therefore,
`we find that one of ordinary skill in the art would have been motivated to
`make this modification in order to reduce system complexity. Petitioner has
`provided arguments and evidence sufficient to show that this modification
`would be within the abilities of one of ordinary skill in the art and that there
`would have been a reasonable expectation of success. Thus, we find that
`Petitioner has set forth sufficient evidence to demonstrate that it would have
`been obvious to modify Karol in a manner that would have taught this
`limitation.
`Patent Owner also argues that Petitioner has not shown that claim 4’s
`“site interface” would have been taught by Karol. PO Resp. 28. Petitioner
`asserts that a “‘site’ in Karol could be either the routers/switches connected
`to the CL-CO gateway and/or the source 101 and/or destination 151
`
`20
`
`IPR2017-00684
`Viptela, Inc. - v. Fatpipe - Exhibit 1020
`Page 20
`
`
`
`IPR2016-00976
`Patent 6,775,235 B2
`
`endpoints, if the CL-CO gateway alone is the ‘controller,’ and the ‘site
`interface’ would be one or more of the input line cards 401 or a network
`connection.” Pet. 44 (citing Ex. 1005 ¶ 175). Patent Owner contends that
`Karol does not teach the recited “site interface” because Karol’s gateway
`(controller) only has “network interfaces” and does not have a “site
`interface.” PO Resp. 28. According to Patent Owner, “[t]he ‘site interface’
`an