ORDER GRANTING-IN-PART AND DENYING-IN-PART DEFENDANTS' MOTIONS FOR SUMMARY JUDGMENT

RONALD M. WHYTE, District Judge.

Defendants A10 Networks, Inc. ("A10") and F5 Networks, Inc. ("F5") (collectively "defendants") bring motions for Summary Judgment of Invalidity, Dkt. No. 1391, and Summary Judgment of Noninfringement, Dkt. No. 137 (A10), Case No. 13-2024, Dkt. No. 91 (F5), against plaintiffs Radware, Inc. and Radware Ltd. (collectively "Radware"). For the reasons explained below, the court DENIES the Motion for Summary Judgment of Invalidity, GRANTS-IN-PART and DENIES-IN-PART A10's Motion for Summary Judgment of Noninfringement, and GRANTS-IN-PART and DENIES-IN-PART F5's Motion for Summary Judgment of Noninfringement.

I. BACKGROUND

Radware brings this patent infringement action against its competitors A10 and F5, alleging infringement of claims 1, 2, 6-9, 13 and 14 of U.S. Patent No. 6,665,702 ('702 Patent); claims 1-7, 9-19, and 21-32 of U.S. Patent No. 8,266,319 ('319 Patent); and claims 1-4, 6-12, 14, and 15 of U.S. Patent No. 8,484,374 ('374 Patent) (collectively asserted patents). All three patents are entitled "Load Balancing" and relate to the "management of networks that have multiple connections to the Internet through multiple Internet Service Providers (ISPs)." '702 col.15 ll.53-56. The '319 Patent is a division of the '702 Patent and the '374 Patent is a continuation of the '319 Patent. The '702 and '319 Patents have the same specification (other than some formatting variances) and the '374 Patent shares that same specification other than the "Summary" section.

The technology at issue relates to link load balancing in a multi-homed environment. A "multi-homed" network is a network with multiple connections to the Internet. '702 col.15 ll.53-56. "Link load balancing" is a process for allocating network communications across these connections.

The asserted patents relate to techniques and systems for selecting a specific route from the multi-homed network to the Internet and from the Internet into the multi-homed network. The claimed inventions describe both "outbound" and "inbound" link load balancing. The claims of the '702 Patent and claims 24-28 of the '319 Patent are directed to outbound link load balancing. Claims 1-23 and 29-32 of the '319 Patent and all claims of the '374 Patent are generally directed to inbound link load balancing. The court explains outbound link load balancing in detail; inbound link load balancing is essentially the reverse.

The patents claim link load balancing as both a method and system. Representative Claim 1 of the '702 patent describes a method for outbound link load balancing:

In outbound link load balancing, an example of which is depicted in Figure 3B of the asserted patents, client 105 is situated within a multi-homed environment and is connected to the Internet 110 through three ISPs 115, 120, and 125. '702 col.15 ll.61-64. In this example, each ISP provides a single route 1, 2, or 3, to the Internet through routers 130, 135, and 140, respectively. Id. col.15 l.64-col.16 l.1. Each router has its own IP address range, 20.x.x.x, 30.x.x.x, and 40.x.x.x, respectively. Id. col.16 ll.4-6.

Client 105 has an IP address of 10.1.1.1 and seeks to connect to remote server 150, with an IP address of 192.115.90.1. When the client 105 connects to remote server 150 over the Internet, content router 145 sends three "polling requests" to server 150 through each of the three routers and ISPs. Id. col.16 ll.10-14. "When sending the polling requests, content router 145 assigns respective network addresses 20.1.1.1, 30.1.1.1 and 40.1.1.1 to client 105. Thus three polling requests are sent: one from each of the sources 20.1.1.1, 30.1.1.1 and 40.1.1.1 to destination 192.115.90.1." Id. col.16 ll.10-14.

The server 150 replies to each of the three polling requests, which are returned through the ISPs. The polling results are then translated by content router 145 into a rating for each route. Id. col.16 ll.26-28. In this example, the polling replies are "measured for latency and number of hops," but the patents also disclose other measureable parameters. Id. col.16 ll.18-20.

The number of hops refers to the number of networking elements between the source and the destination along a particular connection. Dkt. No. 156-2 (Peles Depo.) at 95:23-96:4. Latency is a measure of the time it takes for a communication over the network to travel from one point to another. Id. at 65:9-12. Another measurement used is "time to live" or "TTL," which is the number of hops a packet is allowed to travel before expiring.

Based on the polling results, the content router selects one of the three routes for connecting the client 105 with the server 150. '702 col.16 ll.18-20.The polling results are stored in a "proximity table" 155, shown in Figure 3D. The polling results are saved so that "when a new client 160 with IP address 10.2.2.2 on the private network attempts to connect to a server 165 with IP address 192.115.90.2, through a content router 145, content router 145 determines from proximity table 155 that the best router to use is router 135." Id. col.16 ll.28-34.

Another aspect of the invention ensures that when the content router sends the client request out to the remote server, it also sets the client IP address to correspond to the specific route chosen. For example, if the best route, as determined by the polling requests and selected by the content router is "2", the content router will send the request from the client through router 135 and ISP 120, and sets the client IP address to 30.1.1.1, so that when the remote server replies to the client the information returns through the same route. Id. col.16 ll.40-46.

The court held a Markman hearing on April 8, 2014 and issued its Claim Construction Order on April 18, 2014. Dkt. No. 185. Having held a hearing on May 2, 2014, the court addresses the defendants' various Motions for Summary Judgment.

II. LEGAL STANDARD

Summary judgment is proper where the pleadings, discovery, and affidavits demonstrate that there is "no genuine issue as to any material fact and that the moving party is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(c); see also Celotex Corp. v. Catrett, 477 U.S. 317, 322 (1986). At the summary judgment stage, the Court "does not assess credibility or weigh the evidence, but simply determines whether there is a genuine factual issue for trial." House v. Bell, 547 U.S. 518, 559-60 (2006). Material facts are those which may affect the outcome of the case. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). A dispute as to a material fact is genuine if there is sufficient evidence for a reasonable jury to return a verdict for the nonmoving party. Id.

III. INVALIDITY

A10 and F5 make several invalidity arguments based on 35 U.S.C. § 112. See Dkt. No. 139. Following the Claim Construction Order, Dkt. No. 185, A10 and F5 withdrew invalidity arguments related to at (1) the terms "ISP links," "one or more criteria," "one load balancing criterion," and "weighed function of at least one of," and (2) the term "the IP address for the client within the private network." Dkt. No. 187. The court addresses the remaining arguments.

A. Improper Mixing of Statutory Classes

1. Legal Framework

A single claim that covers both an apparatus and a method of use is invalid under 35 U.S.C. § 112; IPXL Holdings, LLC v. Amazon.com, Inc., 430 F.3d 1377 (Fed. Cir. 2005). "[S]uch a claim is not sufficiently precise to provide competitors with an accurate determination of the `metes and bounds' of protection involved and is ambiguous and properly rejected under section 112, paragraph 2." Id. at 1384 (citation and quotation omitted).

The claim invalidated in IPXL read:

Id. (emphasis and brackets in opinion). The Federal Circuit found the claim invalid because "it is unclear whether infringement of claim 25 occurs when one creates a system that allows the user to change the predicted transaction information or accept the displayed transaction, or whether infringement occurs when the user actually uses the input means to change transaction information or uses the input means to accept a displayed transaction." Id.

2. Claims 8-14 of the '702 Patent; claims 1-12, 24-25, and 29-32 of the '319 Patent; claim 8 of the '374 Patent

Defendants argue that claims 8-14 of the '702 patent; claims 1-12, 24-25, and 29-32 of the '319 patent; and claim 8 of the '374 patent are invalid because they claim both an apparatus and the method steps of using the apparatus. Each of the claims is identified as a system claim in the preamble, but the body of each claim includes language like "receiving", "selecting", "looking up", "assign", and "sets". For example, claim 8 of the '702 patent reads:

'702, claim 8 (emphasis added).

The use of functional language—generally the gerund form of a verb—does not automatically convert the claims into method claims. See, e.g., Apple, Inc. v. Samsung Elecs. Co., Ltd., 876 F.Supp.2d 1141, 1150-1151 (N.D. Cal. 2012). The claims at issue here do not call out affirmative steps that must be taken to infringe, as in IPXL. IPXL, 430 F.3d at 1384 (system claim included limitation that "the user uses the input means"). Infringement of claim 8 (and the other system claims at issue) occurs when "one creates a system that allows the user [to perform the claimed functions]." Id.

Defendants point to Rembrandt Data Technologies, LP v. AOL, LLC, 641 F.3d 1331 (Fed. Cir. 2011) in arguing that a gerund can be interpreted as an affirmative method step. In Rembrandt, the claim recited:

Id. at 1339 (emphasis added). The Federal Circuit found that the final limitation, "transmitting the trellis encoded frames," was a method step and invalidated the claim under IPXL's reasoning. Id.

Rembrandt is distinguishable because the "transmitting" limitation was claimed as an affirmative step, rather than as a function that the system or a component thereof could perform. For instance, claim 8 of the '702 Patent recites "a network controller receiving a client request from within a client computer network directed to a remote server computer, within a second computer network and selecting one of a plurality of routes through which to route the client request." This is in contrast to the '702 patent's method claims, which do not link the "receiving", "looking up", and "selecting" steps to any component. See, e.g., '702 Patent claim 1. Because the functions of "receiving" and "selecting" are linked to the "network controller" component in the system claim, it is clear that the claim is describing the functionality of the "network controller" rather than affirmatively calling out steps for the "network controller" to carry out. See also Eolas Techs., Inc. v. Adobe Sys., Inc., 810 F.Supp.2d 795, 812 (E.D. Tex. 2011) (citing Microprocessor Enhancement Corp. v. Tex. Instruments Inc., 520 F.3d 1367, 1374-75 (Fed. Cir. 2008)) ("It is well-established that for a limitation to introduce a method step, the limitation must require action, or `actual use' of something instead of merely requiring or setting forth a particular capability.").

Although each party argues that HTC Corp. v. IPCom GmbH & Co., KG, 667 F.3d 1270 (Fed. Cir. 2012), supports their interpretation of the claims, HTC does not add to the analysis of the claims under Rembrandt and IPXL. HTC dealt with a somewhat unique "mobile station" device claim where the preamble included a detailed description of the functionality of a separate "network" within which the mobile station was used. Id. at 1274. There the Federal Circuit returned to the inquiry first identified in Rembrandt: whether one of ordinary skill would understand when infringement occurs. Id. at 1277 (citing IPXL, 430 F.3d at 1384). Again, the claims at issue in this case recite a device or system with various components that perform specified functions. There is no confusion over when infringement occurs: "when one makes, uses, offers to sell, or sells the claimed apparatus: the [network controller.]" Id. at 1277.

Defendants next argue that because the claims do not include modifiers like "for receiving" or "capable of receiving" or "configured to receive," they must recite method steps. The Federal Circuit has not adopted a strict test for separating functional language from improper method-step language. The Federal Circuit has found that claims including a gerund without any modifiers are functional limitations. See Microprocessor Enhancement Corp. v. Texas Instruments, Inc., 520 F.3d 1367 (Fed. Cir. 2008); see also CSB-Sys. Int'l Inc. v. SAP Am., Inc., 864 F.Supp.2d 335, 351 (E.D. Pa. 2012) ("The mere fact that the claims failed to use the terminology `capable of' or `intended to' prior to the active terms does not amount to a fatal flaw comparable to that in Rembrandt."). For example, in Microprocessor Enhancement, the court found that the following claim was not invalid:

Id. at 1371-72 (emphasis added). The bolded portion of the apparatus claim includes functional limitations and uses the term "determining" in contrast to "for determining," "capable of determining," or "configured to determine," etc. Nonetheless, the court again applied the IPXL and Rembrandt methodology, emphasizing that the claims place the public on notice of when infringement occurs:

Id. at 1375 (citations ellipsed). Again, the court finds that there is no confusion over when infringement of the system or device claims occurs: when one makes, uses, offers to sell, or sells the claimed apparatus capable of performing the claimed functions. Id.; HTC, 667 F.3d at 1277. Therefore, the claims are not invalid under § 112.

B. Lack of Written Description and Indefiniteness2

1. Legal Framework

A patent's written description must "clearly allow persons of ordinary skill in the art to recognize that [the inventor] invented what is claimed." Vas-Cath Inc. v. Mahurkar, 935 F.2d 1555, 1562-63 (Fed. Cir. 1991). "[T]he test for sufficiency is whether the disclosure of the application relied upon reasonably conveys to those skilled in the art that the inventor had possession of the claimed subject matter as of the filing date." Ariad Pharm., Inc. v. Eli Lilly & Co., 598 F.3d 1336, 1351 (Fed. Cir. 2010) (en banc). The adequacy of the written description is a question of fact. Id.

A patent claim is indefinite if "its claims, read in light of the specification delineating the patent, and the prosecution history, fail to inform, with reasonable certainty, those skilled in the art about the scope of the invention." Nautilus, Inc. v. Biosig Instruments, Inc., No. 13-369, 572 U.S. ___ (2014) (Slip Op. at 1).

2. "the plurality of routes assign respective IP addresses to the computer network"

The phrase "the plurality of routes assign respective IP address to the computer network" appears in claims 1 and 8 of the '702 Patent. Defendants argue that the specification does not teach one of ordinary skill in the art how a "route" can assign an IP address. Instead, an ISP or a router assigns an IP address. Although the defendants did not request construction of this term, the court construed the similar term "sets/setting the source IP address of the client request corresponding to the selected route on the client side." Dkt. No. 185 at 24. The court explained that one of ordinary skill in the art would understand the patent to teach that a portion of the route (the ISP or router) assigns an IP address to a specific route. See Dkt. No. 156-11 (BPAI trans.) at 4:22-5:23; Dkt. No. 155-3 (Zisapel Depo.) at 191:24-192:12; Dkt. No. 155-2 (Peles Depo.) at 329:25-330:7.

Defendants' indefiniteness argument is that the claims are invalid because persons of ordinary skill in the art have more than one understanding of the phrase "the plurality of routes assign respective IP addresses." Defendants point to the deposition testimony of the inventors and a declaration from a technical expert that all "disagree" as to what the claim language means. See Dkt. No. 166 at 13-14. However, the artisans explain that the patents teach one of ordinary skill how selecting a route is necessarily selecting an IP address and vice versa. See Dkt. No. 155-4 (Rubin Decl.) ¶ 20 (explaining that one of ordinary skill would look to the specification and see that an ISP, a part of a route, assigns an IP address; further explaining that the specification identifies routes based on their IP addresses); Dkt. No. 137-8 (Peles Depo.) at 330:14-25 (explaining that the network administrator, via an ISP, can assign an IP address); Dkt. No. 166-2 (Zisapel Depo.) at 190:23-191:8 (explaining how an IP address can designate a specific route). The phrases "the plurality of routes assign respective IP addresses" and "sets/setting the source IP address of the client request corresponding to the selected route" embody these concepts in less than ideal, but still readily understandable, language.

Defendants fail to present "clear and convincing evidence" of invalidity on the basis of either indefiniteness or lack of written description. See 35 U.S.C. § 282; Microsoft Corp. v. i4i Ltd. P'ship, 131 S.Ct. 2238, 2242 (2011).

3. "remote computers"

Claims 3, 15, 29, and 32 of the '319 Patent (and dependent claims 4-5, 16-17, and 30) all recite a step of determining proximities of "remote computers" (plural). For example, claim 3 recites:

Defendants argue that the specification only discusses determining the proximity of a single remote computer. This is not persuasive. The patents clearly contemplate determining the proximity of more than one computer, or determining the proximity of a group of remote computers, such as computers located within the same subnet. For example, the Patent describes determining the proximity of remover server computers using a "subnet IP address" of the remote computer. '319 col.6 l.61 and Fig. 3D. A "subnet IP address" identifies a group of computers within a network that have a common subnet. The Patent also describes connecting to networks with one public IP address and using network address translation ("NAT") to connect to specific computers within the network. '319 cl. 1, col.7 ll.30-37.

At the summary judgment hearing, defendants made two additional arguments. First, defendants argued that NAT and subnet proximities were only used with inbound link load balancing, while the claims at issue here cover outbound link load balancing. However, Figures 3A-3F show outbound link load balancing, with proximity measurements indexed by subnet address. See Figure 3D. Defendants' second argument was that there is no disclosure of using the proximity of an "unrelated remote computer" to select a route. It is true that the patents do not teach how the proximity of an unrelated remote computer could be used to select a route between a client and remote server. However, the claims cannot be read to cover such an embodiment. Defendants cannot unilaterally construe the term "remote computers" to include "unrelated remote computers." This straw man does not change the conclusion that the patents describe determining the proximity of multiple remote computers, such as a computer within the same subnet, and therefore claims of the '319 Patent are not invalid for lack of written description.

C. Lack of Antecedent Basis

1. Legal Framework

A claim that contains a term which lacks antecedent basis may be indefinite "where such [antecedent] basis is not otherwise present by implication or the meaning is not reasonably ascertainable. . . .The common thread in all of these cases is that claims were held indefinite only where a person of ordinary skill in the art could not determine the bounds of the claims . . . ." Halliburton Energy Servs., Inc. v. M-I LLC, 514 F.3d 1244, 1249 (Fed. Cir. 2008) (citation omitted). "When the meaning of the claim would reasonably be understood by persons of ordinary skill when read in light of the specification, the claim is not subject to invalidity upon departure from the protocol of `antecedent basis.'" Energizer Holdings, Inc. v. Int'l Trade Comm'n, 435 F.3d 1366, 1370 (Fed. Cir. 2006)

2. "a remote computer"

Dependent claims 4, 9, 16, 21, and 30 of the '319 Patent all recite the term "a remote computer." The claims depend from claims that also recite the term "a remote computer." For example, claim 9 depends from claim 1 as follows:

Defendants argue that the lack of antecedent basis for "a remote computer" in claim 9 renders the claims indefinite. First, this is not a typical lack of antecedent basis situation. Lack of antecedent basis generally occurs when a claim first refers to an element using the words "the" or "said," with no earlier recitation of the claimed element. See Manual of Patent Examination Procedure § 2173.05(e) (9th ed. Mar. 2014). Here, the claims recite "a remote computer" twice. Defendants are essentially making the same argument addressed above: the claim as written recites multiple remote computers, which is not supported by the specification.

As discussed above, the phrase "remote computer" can encompass multiple computers in a remote server farm or remote subnetwork. In that situation, the network controller would receive a DNS resolution query from a remote computer in the remote network. When determining proximity, the network controller would send a query to the remote network, but may actually receive a response from a different remote computer within the same remote network. Thus, the proximity is not necessarily limited to "a proximity of the [identical] remote computer." One of ordinary skill in the art would understand how to determine the proximity of a remote computer within a remote network, as discussed above with regard to "remote computers" (plural). Defendants failed to demonstrate that the claims are indefinite to one of ordinary skill in the art.

IV. NONINFRINGEMENT

A. Legal Framework

As the Federal Circuit has noted, summary judgment of noninfringement is a two-step analysis. "First, the claims of the patent must be construed to determine their scope. Second, a determination must be made as to whether the properly construed claims read on the accused device." Pitney Bowes, Inc. v. Hewlett-Packard Co., 182 F.3d 1298, 1304 (Fed. Cir. 1999) (internal citation omitted). "[S]ummary judgment of non-infringement can only be granted if, after viewing the alleged facts in the light most favorable to the non-movant, there is no genuine issue whether the accused device is encompassed by the claims." Id. at 1304. "Whether a claim is infringed under the doctrine of equivalents may be decided on summary judgment if no reasonable jury could determine that the limitation and the element at issue are equivalent." Zelinski v. Brunswick Corp., 185 F.3d 1311, 1317 (Fed. Cir. 1999) (citing Warner-Jenkinson Co. v. Hilton Davis Chem. Co., 520 U.S. 17, 39 n. 8 (1997)).

B. Claims Radware No Longer Asserts

A10 and F5 argue that they are entitled to summary judgment of noninfringement on all claims that Radware no longer asserts. The court agrees. Radware's original infringement contentions asserted every claim of the three asserted patents, and A10 and F5's declaratory judgment counterclaims were directed toward every claim as well. Under MedImmune, "a declaratory action is available when the facts as alleged `under all the circumstances, show that there is a substantial controversy, between the parties having adverse legal interests, of sufficient immediacy and reality to warrant the issuance of a declaratory judgment.'" Revolution Eyewear, Inc. v. Aspex Eyewear, Inc., 556 F.3d 1294, 1297 (Fed. Cir. 2009) (quoting MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118, 127 (2007)).

Radware argues that because it removed certain claims from its infringement contentions, there is no longer a controversy between the parties. Radware also notes that it would be estopped from asserting the silently withdrawn claims in future patent litigation. Nonetheless, in light of the Federal Circuit's case law that even some covenants not to sue do not eliminate declaratory judgment jurisdiction, the justiciable controversy did not end when Radware silently abandoned some claims in its infringement contentions. See, e.g., Revolution Eyewear, 556 F.3d at 1297-98.

As the Supreme Court recently affirmed, Radware bears the burden of proving infringement, even with respect to A10 and F5's declaratory judgment of noninfringement counterclaims. Medtronic, Inc. v. Mirowski Family Ventures, LLC, 134 S.Ct. 843, 849-50 (2014). Because Radware presents no evidence that A10 or F5's products infringe any claims Radware no longer asserts, Radware has not met its burden and A10 and F5 are thus entitled to summary judgment of noninfringement as to claims Radware no longer asserts.3

The court grants F5 summary judgment of noninfringement as to the following claims: claims 3, 4, 5, 10, 11, and 12 of the '702 Patent; and claims 5 and 13 of the '374 Patent. The court grants A10 summary judgment of noninfringement as to the AX Series products on the following claims: claims 2, 3, 4, 5, 9, 10, 11, and 12 of the '702 Patent; claims 8, 20, 24, 25, 26, 27, and 28 of the '319 Patent; and claims 5 and 13 of the '374 Patent. The court grants A10 summary judgment of noninfringement as to the EX Series products on the following claims: claims 3, 4, 5, 10, and 12 of the '702 Patent; claims 8, 9, 20, and 21 of the '319 Patent; and claims 5 and 13 of the '374 Patent.

C. Direct Infringement of Method Claims

The defendants' first broad noninfringement argument is that they do not directly infringe any asserted method claims because they only supply finished products to customers—the defendants do not use the products themselves. "[F]or a party to be liable for direct patent infringement under 35 U.S.C. § 271(a), that party must commit all the acts necessary to infringe the patent, either personally or vicariously." Aristocrat Technologies Australia Pty Ltd. v. Int'l Game Tech., 709 F.3d 1348, 1362 (Fed. Cir. 2013) (quoting Akamai Technologies, Inc. v. Limelight Networks, Inc., 692 F.3d 1301, 1307 (Fed. Cir. 2012) rev'd on other grounds, Limelight Networks, Inc. v. Akamai Technologies, Inc., 134 S.Ct. 2111 (2014). For method claims, to prove direct infringement, "a patent holder must establish that an accused infringer performs `all the steps of the claimed method, either personally or through another acting under his direction or control.'" Aristocrat, 709 F.3d at 1362 (quoting Akamai, 692 F.3d at 1307).

A10 and F5 argue that they are entitled to summary judgment of no direct infringement of the asserted method claims because end users must perform some of the claimed method steps. Radware responds that A10 and F5 can be liable for direct infringement of the method claims if they perform all of the claimed steps during product testing. See Aristocrat, 709 F.3d at 1363 (recognizing that product testing can constitute direct infringement of method claims, despite affirming trial court's grant of summary judgment on claim language-specific grounds). As discovery has yet to close in this case, Radware asserts that it continues to pursue discovery regarding the defendants' testing to determine whether either defendant directly infringes the asserted method claims. In their replies, A10 and F5 acknowledge that the court should defer ruling on whether A10 and F5 "use" the method steps until a later date, but still request summary judgment on the make, sell, offer to sell, or import prongs of 35 U.S.C. § 271(a). Dkt. No. 165, at 3-4; 13-cv-2024, Dkt. No. 110, at 8-9. However, because discovery has yet to close and Radware may still present evidence of one or both defendants' alleged direct infringement of the asserted method claims, the court denies without prejudice A10 and F5's motions for summary judgment of no direct infringement of the asserted method claims. It does not appear, however, that Radware contends infringement other than by use.

D. Direct Infringement of Apparatus Claims

Defendants' second broad argument is that they do not directly infringe the asserted apparatus claims because those claims require configurations that are not supplied by A10 or F5. For example, claims 8-14 of the '702 Patent require "[a] network management system for managing a computer network connected to the Internet through a plurality of routes." '702 cl. 8. The defendants argue that they are entitled to summary judgment of no direct infringement of these claims because their devices do not include a computer network, nor are they connected to the Internet.

Although the parties agreed that the preamble of the claims is limiting, none of the claims when properly interpreted affirmatively recite "the Internet" or a connection to the Internet as a component of the device or system. A preamble does not recite a component of the claim "where a patentee defines a structurally complete invention in the claim body and uses the preamble only to state a purpose or intended use for the invention." Rowe v. Dror, 112 F.3d 473, 478 (Fed. Cir. 1997); see also MPEP § 2111.02 (9th ed. Mar. 2014). The preambles at issue here state the intended use of the device and do not call out additional structural requirements.

For example, claim 8 of the '702 Patent recites a network management system "for managing a computer network connected to the Internet." This preamble clearly discloses the intended use of the system, and does not affirmatively recite the Internet as part of the claimed system. Similarly, claim 9 of the '374 Patent recites a device "for load balancing" across multiple ISP links. Again, this use of functional language in the preamble of the claims "only state[s] a purpose or intended use for the invention." Rowe, 112 F.3d at 478. Multiple ISP links are not required for a structurally complete device. The same is true of claim 29 of the '319 Patent.

Claim 1 of the '319 Patent is slightly different. Claim 1's preamble requires "a device for managing a computer network, said device connected to the Internet through a plurality of routes . . ." (emphasis added). Here, the phrase "said device connected to the Internet" does not use the typical functional language just analyzed. However, Claim 1 is limited to "a device." The preamble clearly does not contemplate that the Internet and multiples Internet connections are structural components of the device. If a person of ordinary skill in the art was asked to construct the claimed "device for managing a computer network," he would build the discrete "content router" 145 shown in Figures 3A-3F of the Patent. See '319 col.15 ll.38-51. The ISP connections, and Internet cloud, though necessary to implement the functions described in the Patents, are illustrated and described as separate from the device. Id. col.15 ll.53-55. Any person of ordinary skill in the art reading the claims would understand that infringement occurs upon production of a device designed to be used with multiple connections to the Internet. The preambles at issue here are limiting in the sense that they recite the intended use of the system or device, but do not recite additional components (i.e. Internet connections) that are a part of the system or device. See Rowe, 112 F.3d at 478

Defendants also argue that they do not infringe the product claims because the products as sold are not configured to perform the claimed functions.4 However, "in an apparatus claim involving software, the infringement occurs because the software includes the patented feature: `although a user must activate the functions programmed into a piece of software by selecting those options, the user is only activating means that are already present in the underlying software.'" Brocade Comm'ns Sys., Inc. v. A10 Networks, Inc., C 10-3428 PSG, 2013 WL 831528, at *11 (N.D. Cal. Jan. 10, 2013) (quoting Fantasy Sports Properties, Inc. v. Sportsline.com, Inc., 287 F.3d 1108, 1118 (Fed. Cir. 2002) and Finjan, Inc. v. Secure Computing Corp., 626 F.3d 1197, 1205 (Fed. Cir. 2010)). Here, the user merely activates functionality present in the underlying software when he connects the accused devices to the Internet.

The defendants also rely on Ball Aerosol & Specialty Container, Inc. v. Ltd. Brands, Inc., 555 F.3d 984, 994-95 (Fed. Cir. 2009) for their argument that there is no direct infringement until the device is connected to the Internet. Ball Aerosol is distinguishable. The claims in Ball Aerosol were directed to a candle-holding tin and required very specific placement of various components. Id. at 987-88. It was undisputed that the patentee had "no proof that the [accused product] was ever placed in the infringing configuration, and it [was] clear that the [accused product] [did] not necessarily have to be placed in the infringing configuration." Id. at 995.

In the instant case, neither party disputes that the accused devices must be connected to the Internet to perform the functions disclosed in the specification. Moreover, the Ball Aerosol case dealt with tangible structures, whereas this case, along with the Brocade, Fantasy Sports, and Finjan cases Radware cites, concern software. As such, summary judgment of no direct infringement of the apparatus claims is improper.

E. '702 Patent

1. Setting the source IP address

Turning now to defendants' patent-specific arguments, A10 argues that its AX Series products do not infringe the '702 Patent because they do not "set[] the source IP address of the client request corresponding to the selected route on the client side." '702 cl. 1 and 8. In its claim construction order, the court adopted A10's proposed construction, construing this "setting the source IP address" language as "setting/sets the source IP address of the client request to one associated only with the selected route that connects the client computer network to the Internet." CCO at 22-25. As the court observed, "part of the invention is ensuring that responses will return to the server using the same route . . . ." Id. at 25.

The thrust of A10's contention is that its AX Series products can set a source IP address, but that the AX Series products do not set a source IP address specific only to the selected route. See Dkt. No. 165, A10 Noninfringement Reply, at 14-15. Radware responds by presenting evidence raising an issue of material fact over whether the AX Series products set a source IP address specific to the selected route. In particular, Radware points to A10 documentation indicating that "to configure the AX Series to perform outbound LLB [link load balancing] with optional QoS [quality of service]," the product "create[s] a Source NAT [network address translation] (SNAT) pool for each link." Dkt. No. 192-7, at 3. The same document later notes that each testing "session has a different return source IP address that comes from the source NAT pool corresponding to each WAN [wide area network] link." Id. at 8. According to Dr. Izhak Rubin, Radware's expert, this document supports his view that A10's AX Series products set a source IP address corresponding only to the selected route. Dkt. No. 192-9, Supplemental Rubin Decl. ¶ 22. A10 disputes Dr. Rubin's characterization of the evidence, but the dispute merely reveals the existence of a genuine issue of material fact for trial. Therefore, summary judgment is denied as to A10's "setting the source IP address" argument.

2. Proximity arguments

The defendants raise several arguments related to the proximity limitations in the '702 Patent. All claims of the '702 Patent require a "proximity table," which the court construed in part as "logically organized electronically stored information expressing a proximity." CCO at 15-16. The disputes generally focus on whether the alleged proximity table expresses a proximity. The court construed "proximity or proximities" as "a measurement or measurements based on hops, latency, TTL, or a combination thereof." CCO at 6-11.

a. F5's accused products

F5 contends that its accused products do not infringe the '702 Patent because its products do not measure any "proximities," as required by all claims of the '702 Patent. Radware does not contest that F5's products do not measure hops or TTL. Rather, Radware argues that F5's "fastest link" variable is a latency measurement, or, in the alternative, the "fastest link" variable infringes under the doctrine of equivalents. The court is not persuaded, and therefore the court finds that summary judgment of noninfringement of F5's products as to the '702 Patent is proper.

In support of its motion for summary judgment, F5 presents the uncontroverted declaration of Pete Thornewell, a senior architect at F5, who testifies that the "fastest link" metric "is based on the least number of outstanding requests through that pool member." No. 13-2024, Dkt. No. 110-2, Supplemental Thornewell Decl. ¶ 4. In other words, the "fastest link" variable measures the size of the queue of outstanding requests through that pool member—it makes sense that a pool member with a shorter queue of requests will be a faster link, and therefore F5 includes that variable in its software. Given that latency is a measurement of travel time, see Dkt. No. 156-2, Peles Depo. at 65:9-12, the "fastest link" metric, which is not a measurement of travel time, is not a latency measurement as contemplated by the '702 Patent.

Radware in Dr. Rubin's supplemental declaration appears to drop its literal infringement position and argue instead for infringement under the doctrine of equivalents. Dr. Rubin testifies that the "fastest link" variable "directly relates" to latency because a faster link speed—for which the "fastest link" metric only acts as a proxy measurement—is one way to achieve lower latency. Supplemental Rubin Decl. ¶ 14.

The "essential inquiry" in any determination under the doctrine of equivalents is whether "the accused product or process contain[s] elements identical or equivalent to each claimed element of the patented invention." Warner-Jenkinson Co., Inc. v. Hilton Davis Chem. Co., 520 U.S. 17, 40 (1997). For this essential inquiry, the Federal Circuit uses the function-way-result test as set forth in Union Paper-Bag Machine Co. v. Murphy, 97 U.S. 120, 125 (1877), which asks whether an element of an accused product "performs substantially the same function in substantially the same way to obtain the same result" as an element of the patented invention. See, e.g., Am. Calcar, Inc. v. Am. Honda Motor Co., Inc., 651 F.3d 1318, 1338 (Fed. Cir. 2011).

The court holds that no reasonable jury could find that the "fastest link" measurement is equivalent to latency. Even if the fastest link metric performs substantially the same function as a latency metric—a proposition contested by F5—no reasonable jury could determine that the fastest link metric performs that function in substantially the same way. Dr. Rubin acknowledges that the latency measurement as described in the '702 Patent measures "the time it takes for a communication over the network to travel from one point to another." Supplemental Rubin Decl. ¶ 14. More specifically, the '702 Patent refers to "the latency of the received replies in travelling a round trip from the computer network to the remote server and back from the remote server to the computer network." '702 col.5 ll.34-37. By contrast, the fastest link metric measures only the number of outstanding requests, and thus is not a time measurement, nor does it have any relation to the destination server. Consequently, construing the evidence most favorably to Radware, no reasonable jury could find that the fastest link metric performs the function of the latency metric in substantially the same way as the latency metric. Therefore, F5 is entitled to summary judgment of noninfringement of all its accused products as to the '702 Patent.

b. A10's accused products

As the court crafted its own constructions of "proximity" and "ratings [for a plurality of routes]," A10 raises two new noninfringement arguments in its supplemental brief regarding its summary judgment motion, to which Radware replies in its response to A10's brief.

First, A10 contends that its AX Series products do not infringe because the AX Series products only measure the properties of immediately adjacent routers. Resolution of this dispute turns on whether the court's construction of "ratings [for a plurality of routes]" requires the accused device to take a proximity measurement through the entire route from the client computer network and the remote server computer. A10 argues that the proximity measurement must be taken through the entire route, and Radware asserts that the proximity measurement need only be taken through the route to any point along the route, such as an adjacent router.

The '702 Patent contemplates taking proximity measurements through at least the entire route to the remote server.5 The figures of the '702 Patent demonstrate this with clarity. For example, Figures 3A-3F depict an exemplary polling request traveling from the client through the entire route to the remote server. In particular, Figure 3B illustrates a polling request beginning at the content router, passing through the ISP and Internet on its way to the server. The "polling request" box in Figure 3B confirms that the polling request travels along the entire route, as it shows an IP address associated with the destination server (192.115.90.1) as the destination for the polling request, and three source IP addresses associated with the client (20.1.1.1, 30.1.1.1, and 40.1.1.1). The '702 Patent describes Figure 3B as follows:

'702 col.16 ll.6-14 (emphasis added). Figure 3C then depicts three lines, one for each polling response, which travel from the remote server back to the content router. The "polling response" box shows proximity measurements for each route, indicating that the proximity measurement is taken through the entire route to the remote server.

The abstract of the '702 Patent confirms this interpretation. It states that the invention "include[es] the steps of; sending polling requests through a plurality of routes from a computer network to a remote server computer, receiving replies from the remote server computer corresponding to the polling requests, and measuring proximities of the remote server computer to the computer network based on the received replies." '702 Abstract (emphasis added). The body of the '702 Patent's specification also parrots this language. '702 col.6 ll.34-39 ("sending polling requests through a plurality of ISPs from a computer network to a remote server computer, receiving replies from the remote server computer corresponding to the polling requests, and measuring proximities of the remote server computer to the computer network based on the received replies"). Moreover, Radware cannot point to any part of the '702 Patent that discloses polling requests that are directed only at an adjacent router as is found in the AX Series products. Therefore, the court finds that A10 is entitled to summary judgment that its AX Series products do not infringe the '702 Patent because they do not include "ratings based on proximity measurements taken through each of the plurality of routes."

A10's second argument asserts that both its AX and EX Series products do not infringe because those products do not have a proximity table containing route ratings apart from the proximity measurements themselves. In other words, the AX and EX Series products' proximity table only stores the proximity measurements, and not a separate "rating." Here, A10 misapprehends the court's claim construction order. The court did not hold, as A10 asserts, that "ratings are not the proximity measurements themselves." Dkt. No. 187, A10 Supplemental Brief Regarding Summary Judgment Motions, at 3.The court found that "the ratings are not required to be in a preferred order of selection," the ratings must be "based on" the proximity measurements, and that "the ratings are used in the `selecting' step to determine which route to use." CCO at 13. As such, the court never restricted the scope of "ratings" to exclude unaltered proximity values as A10 suggests, nor do any passages of the specification cited at the hearing limit "ratings" as A10 contends. Therefore, the "ratings" contained in the proximity table may be the proximity measurements themselves, and A10's motion for summary judgment is denied as to this "ratings" argument.

In sum, the court grants summary judgment as to A10's AX series and the '702 Patent; denies summary judgment as to A10's EX series and the '702 Patent; and grants summary judgment as to F5's products and the '702 Patent.

F. '374 Patent — "Configured"

A10 and F5 argue that their accused products do not infringe claims 9-12 and 14-15 of the '374 Patent because their products are not "configured" in the way required by the claims. The court construed "configured to" as "programmed to [perform certain functions]. This does not require user intervention if the feature claimed is included in the product as supplied." CCO at 21-22. This claim construction resolves the defendants' arguments in Radware's favor. A10 and F5 contend that their products "are not `plug-and-play' devices," and that the products cannot perform the claimed functionality "out of the box." Dkt. 187, Supplemental Brief Regarding Summary Judgment Motions, at 4-5. F5 elaborates further, stating that "the user must take various action" Before infringement occurs. Id. at 5. However, the court explicitly addressed these summary judgment arguments with the second part of its claim construction, holding that user intervention is not required "if the feature claimed is included in the product as supplied." CCO at 21-22. Because the necessity of user intervention does not mean that the accused products are not "configured" to perform the claimed function under the court's construction, the court denies the defendants summary judgment as to claims 9-12 and 14-15 of the '374 Patent.

G. '374 and '319 Patents — "Cost"

In its claim construction order, the court rejected A10's construction of "cost" or "costing information," giving the terms their "plain and ordinary meaning not limited to monetary price." COO at 25-26. Radware alleges that A10's products select a route based on "costing information" because A10's products use a metric called "BW-Cost" (bandwidth cost). Although A10's summary judgment briefs appear to concede that the "cost" noninfringement argument was dependent on the court accepting A10's construction, A10 re-raised the argument at the hearing, claiming that bandwidth utilization is distinct from "costing information." Even were the court to interpret A10's summary judgment position as independent from its now-rejected claim construction, A10's noninfringement argument raises genuine issues of material fact. A fact-finder must determine whether "BW-Cost" is within the plain and ordinary meaning of "cost," "not limited to monetary price." Therefore, the court denies A10 summary judgment of noninfringement on the "cost" limitation.

V. ORDER

For the foregoing reasons, the court DENIES summary judgment of invalidity. The court GRANTS summary judgment of noninfringement as to the following claims. Summary judgment of noninfringement is DENIED as to all other claims.

Summary Judgment of Noninfringement GRANTED: