`
`IN THE UNITED STATES PATENT & TRADEMARK OFFICE
`______________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`______________________
`
`
`TALARI NETWORKS, INC.,
`
`Petitioner,
`
`v.
`
`FATPIPE NETWORKS PRIVATE LIMITED,
`
`Patent Owner.
`______________________
`
`Case IPR2016-00976
`Patent U.S. 6,775,235
`______________________
`
`
`PATENT OWNER’S REQUEST FOR REHEARING
`
`
`
`
`
`
`
`
`
`
`
`
`
`INTRODUCTION
`Pursuant to 37 C.F.R. § 42.71(d), Patent Owner FatPipe Networks
`
`Private Limited and exclusive licensee FatPipe, Inc. (collectively,
`
`“FatPipe” or “Patent Owner”), respectfully request rehearing of the
`
`Board’s Final Written Decision of November 1, 2017. (Paper 32,
`
`“Decision”) for the following matters relevant to claims 4 and 9:
`
`
`
`First, in determining “that it would have been obvious to modify
`
`Karol by limiting the routing decision to an analysis of the packet’s
`
`source address,” the Decision overlooks the fact that claim 4 of the ’235
`
`patent expressly requires “select[ion] between network interfaces on a
`
`per-packet basis according to at least: a destination of the packet.”
`
`(Decision, pp. 19-20, emphasis added).
`
`
`
`Second, in determining that dependent claim 9 is obvious for the
`
`same reason as claim 4, the Decision similarly overlooks that claim 9
`
`depends from claim 5, which requires analysis of the “destination
`
`location” and “destination address.”
`
`
`
`Third, the Decision overlooked or misapprehended the passages of
`
`the ’235 patent that describe path section based on the origin of the
`
`1
`
`
`
`packet as “coarse” and that distinguish such per-department and per-
`
`router methods from per-packet and per-session path selection. (See Ex.
`
`1001, col. 4:15-23, 7:38-42, 11:33-38).
`
`
`
`Fourth, the Decision overlooked and/or misapprehended that
`
`routing based on the source address will forward all packets from the
`
`same source to the same network, which is similar to the flow-based
`
`routing that the Board correctly distinguished from per-packet selection
`
`in claims 4 and 9. (Ex. 1001, col. 6:62–7:5, 7:38-42, 11:33-38).
`
`ARGUMENT
`
`A. If Karol is modified to analyze only the source address of
`a packet, then the modified system excludes the subject
`matter of claim 4, which requires selection based on “a
`destination address.”
`Patent Owner argued in its response that “Karol does not disclose
`
`or render obvious ‘a packet path selector which selects between network
`
`interfaces on a per-packet basis according to at least: a destination of
`
`the packet ….’” (PO Resp., pp. 23, 27, emphasis added). The Decision
`
`overlooked this limitation when it relied upon Petitioner’s assertion
`
`“that it would have been obvious to modify Karol by limiting the routing
`
`2
`
`
`
`decision to an analysis of the packet’s source address.” (Decision, pp.19-
`
`20, emphasis added).
`
`Claim 4 of the ’235 patent requires “a packet path selector which
`
`selects between network interfaces on a per-packet basis according to at
`
`least: a destination of the packet, an optional presence of alternate paths
`
`to that destination, and at least one specified criterion for selecting
`
`between alternate paths when such alternate paths are present.” The
`
`Board agreed with Patent Owner that “per-” packet means “selecting a
`
`network path/interface for each packet” (Decision, p. 9) and
`
`subsequently concluded that Karol does not disclose the per-packet
`
`selection required in claim 4 (Decision, p. 18-19). But, the Decision
`
`adopted Petitioner’s fallback position that “that it would have been
`
`obvious to modify Karol by limiting the routing decision to an analysis
`
`of the packet’s source address.” (Decision, pp. 19-20).
`
`This fallback position was conditioned on a claim construction
`
`that neither party advocated and that the Board did not adopt, namely,
`
`that per-packet basis means “regardless of the session with which the
`
`packet is associated.” (Petition, p. 45). Nonetheless, Patent Owner was
`
`concerned by that possibility and urged that, if such a claim
`
`3
`
`
`
`construction was adopted (which it wasn’t), a “POSITA would have
`
`found substituting [a] the packet-by-packet path selection process that
`
`considers multiple criteria including associated flows as explicitly
`
`disclosed in Karol with [b] a much simpler and known packet path
`
`selection process that considers only source address regardless of the
`
`session to yield a highly successful and predictable result” because it
`
`was an obvious substitution of one known element for another and
`
`obvious to try. (Petition, p. 46, citing Ex. 1005, ¶¶ 194-196, emphasis
`
`added). In any case, the Board didn’t adopt that construction, and thus,
`
`“a packet path selector which selects between network interfaces on a
`
`per-packet basis according to at least: a destination of the packet …”
`
`remains a limiting element of claim 4.
`
` Because claim 4 requires that the destination of the packet is
`
`used to select between network interfaces on a per-packet basis,
`
`modifying Karol to limit the routing decision to an analysis of the
`
`packet’s source address excludes the first of the three criteria specified
`
`in claim 4 for selecting a network—the destination of the packet.
`
`Neither the Petition nor the Decision addresses how this new deficiency
`
`4
`
`
`
`is cured if Karol’s routing decision is limited to an analysis of the
`
`packet’s source address.
`
`Thus, the Board should reconsider its findings and hold that claim
`
`4 is not unpatentable.
`
`B. If Karol is modified to analyze only the source address of a packet,
`then the modified system excludes the subject matter of claim 9,
`which depends from claim 5 and thus requires network selection
`based on “a destination address.”
`On the issue of network path selection, Patent Owner made the
`
`same arguments for claims 4 and 9: “Independent of the arguments
`
`above for claim 5, claim 9 is also patentable over Karol because Karol
`
`does not disclose or render obvious ‘repeated instances of the selecting
`
`step make network path selections on a packet-by-packet basis.’ ” (PO
`
`Resp., p. 51). In concluding that claim 9 is obvious, the Decision relied
`
`on the same rationale that it used for claim 4. (Decision, p. 31).
`
`As discussed above, this rationale—that it would have been
`
`obvious to modify Karol by limiting the routing decision to an analysis
`
`of the packet’s source address—was based on a claim construction that
`
`neither party advocated and that the Board did not adopt. In any event,
`
`a system that limits the routing decision to an analysis of the packet’s
`
`5
`
`
`
`source address excludes the method of claim 9 because of the following
`
`limitation in independent claim 5 that require analysis of the
`
`destination address/location:
`
` “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”;
`
`The “known location address range” and “associated networks” are
`
`used in conjunction with the destination address or destination location,
`
`which are elements of the first step of claim 5. Thus, the proposed
`
`6
`
`
`
`modification of Karol to limit the routing decision to an analysis of the
`
`packet’s source address implicates every step of claim 5, which
`
`repeatedly relies upon the destination address or the destination
`
`location, neither of which is relevant to Karol’s system as modified.
`
`Neither the Petition nor the Decision addresses how this new
`
`deficiency is cured if Karol’s routing decision is limited to an analysis of
`
`the packet’s source address. Thus, the Board should reconsider its
`
`findings and hold that claim 4 is not unpatentable.
`
`C. The background art of the ’235 patent does not disclose
`selecting between networks on a “per-packet” (claim 4)
`or “packet-by-packet” (claim 9) basis.
`In its response, Patent Owner pointed out that the background art
`
`described in the ’235 patent at col. 4, lines 15-23, used coarse routing of
`
`traffic or flows between networks instead of per-packet or packet-by-
`
`packet routing. (PO Resp., pp. 4, 23). The Board misconstrued this
`
`portion of the ’235 patent to describe selecting between networks on a
`
`per-packet/packet-by-packet basis.
`
`In relying upon the conclusion, “that it would have been obvious to
`
`modify Karol by limiting the routing decision to an analysis of the
`
`packet’s source address,” (Decision, pp. 19-20), the Board specifically
`
`7
`
`
`
`relies upon Petitioner’s citation to column 4, lines 15-23 of the ’235
`
`patent, which describes the following:
`
`But better tools and techniques are needed for use in
`architectures such as that shown in FIG. 5. In particular,
`prior approaches for selecting which network to use for
`which packet(s) are coarse. For instance, all packets from
`department X might be sent over the frame relay
`connection 106 while all packets from department Y are sent
`over the Internet 500. Or the architecture might send all
`traffic over the frame relay network unless that network
`fails, and then be manually reconfigured to send all traffic
`over a VPN 502. (Emphasis added).
`
`The Petition cites this passage in support of its statement that the
`
`“The ’235 Patent also discloses that a POSITA would have known that
`
`the prior art discloses routing decisions that are independent of the
`
`particular flows or sessions of particular packets.” (Petition, p. 45). But
`
`the Petition neglects to mention that, in fact, the ’235 patent criticizes
`
`this per-department or per-router approach as “coarse” to distinguish
`
`the way the invention selects network paths on a per-session or per-
`
`packet basis. (Col. 4:15-18). The ’235 patent describes that the more
`
`8
`
`
`
`“granular” techniques of per-packet and per-session network selection
`
`are distinct from, and preferred over, per-department selection:
`
`Load-balancing is preferably done on a per-packet basis for
`site-to-site data traffic over the Internet or frame relay net,
`or done on a TCP or UDP session basis for Internet traffic,
`as opposed to prior approaches that use a per-department
`and/or per-router basis for dividing traffic. (Col. 7:37-42).
`
`Therefore, the ’235 patent’s description of per-department network
`
`selection does not lend any teaching or rationale to modify Karol to
`
`make a decision on a per-packet basis.
`
`If anything, column 4, lines 15-23, of the ’235 patent shows that,
`
`even when using the source address or the origin of the packet as the
`
`basis for making a routing decision, there is still a problem that path
`
`selection is too coarse because a routing decision is broadly applied to
`
`all packets of a certain origin rather than making a new selection for
`
`each individual packet or session. Thus, the Board misapprehended the
`
`meaning of this passage of the ’235 patent in concluding that per-
`
`department routing is an example of per-packet selection, when in fact,
`
`this passage describes per-department routing to distinguish per-packet
`
`selection, as required by claims 4 and 9.
`
`9
`
`
`
`Therefore, the Board should reconsider and reverse its finding
`
`that at least claims 4 and 9 are obvious over Karol.
`
`D. The modification of Karol to analyze only the source
`address of a packet is inconsequential to achieving the
`feature of per-packet or packet-by-packet selection.
`Patent Owner’s response emphasized the difference between
`
`forwarding packets on pre-computed routes and selecting packets on a
`
`per-packet basis:
`
`Karol’s CL-CO gateway uses the forwarding database’s pre-
`computed routes to make network path decisions and,
`therefore, does not “select” a network when a packet arrives
`at the gateway. (P. 24).
`
`Accordingly, when a packet arrives at Karol’s CL-CO
`gateway there is no “select[ing]” process because the packet
`is simply routed to the network path that was pre-selected
`prior to the packet’s arrival…. (P. 26).
`
`The Decision overlooked or misapprehended this distinction and offers
`
`no explanation how the substitution of source routing into Karol could
`
`amount to per-packet or packet-by-packet selection, as required by
`
`claims 4 and 9.
`
`10
`
`
`
`If Karol’s system were modified to analyze only the source
`
`address, then there could be no per-packet path selection at the CL-CO
`
`gateway because the routing decisions would be predetermined, based
`
`on the source. For example, “all packets from department X might be
`
`sent over the frame relay connection 106 while all packets from
`
`department Y are sent over the Internet 500.” (Ex. 1001, col. 4:18-21).
`
`The Board correctly determined Karol’s routing decision, applied
`
`to a flow of packets, is not per-packet network path selection: “Thus, we
`
`determine that Karol’s routing decisions are made for a flow of packets
`
`and not for an individual packet.” (Decision, p. 18). In other words, if a
`
`routing table or flow database forwards all packets with a particular
`
`destination address to a particular network interface, there is no per-
`
`packet selection taking place. Similarly, if a routing table or flow
`
`database forwards all packets from a particular origin or source address
`
`to a particular network interface, there is no per-packet selection taking
`
`place because the selection is made for an entire group of packets
`
`having the same source address. This is exactly what the ’235 patent
`
`considers to be a coarse packet selection criteria because the selection is
`
`11
`
`
`
`broadly applied to all packets of a particular origin rather than
`
`separately making a network selection for each packet.
`
`For this reason, the Board should reconsider and reverse its
`
`finding that at least claims 4 and 9 is obvious over Karol.
`
`CONCLUSION
`For all the foregoing reasons, the Board should reconsider and
`
`reverse its finding that at least claims 4 and 9 are unpatentable.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Dated: December 1, 2017
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Respectfully submitted,
`
`Oblon, McClelland, Maier &
` Neustadt, LLP
`
`
`
`
`
`
`
`/Robert C. Mattson/
`Robert C. Mattson
`Reg. No. 42,850
`(703) 412-6466 (direct)
`
`
`12
`
`
`
`
`
`
`
`CERTIFICATE OF SERVICE
`
`Pursuant to 37 C.F.R. § 42.6(e), the undersigned certifies service
`
`of PATENT OWNER’S REQUEST FOR REHEARING on the counsel of
`
`record for the Petitioner by filing this document through the PTABE2E
`
`System as well as delivering a copy via electronic mail to the following
`
`address:
`
`
`
`
`Andy H. Chan (chana@pepperlaw.com)
`Charles F. Koch (kochc@pepperlaw.com)
`Thomas F. Fitzpatrick (fitzpatrickt@pepperlaw.com)
`Pepper Hamilton LLP
`
`Dated December 1, 2017
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`/Robert C. Mattson /
`Robert C. Mattson
`Reg. No. 42,850
`
`
`
`
`
`
`
`
`
`
`
`