dissenting.
While I agree with much of the majority opinion, I respectfully dissent from the majority’s decision upholding the verdict of infringement. In my view there was insufficient evidence that the requirement of an “upstream manager” was satisfied. The majority has broadened a poorly drafted patent to cover an invention that was not actually claimed or described in the specification.
I
The ’804 patent claims a “Method and Apparatus for Scalable, High Bandwidth Storage Retrieval and Transportation of Multimedia Data on a Network.” It relates to an “information storage and transport system[ ]” which delivers “data streams over a network,” from a server to client devices. Col. 1 11. 10-11. The system supports various types of client devices, including “set-top boxes” (which permit consumers to download video and view it on them television sets), personal digital assistants, and video phones. Col 2 11. 18-20; Col 3 11. 23-28. Claim 1 of the ’804 patent recites:
1. A high bandwidth, scalable server for storing, retrieving, and transporting multimedia data to a client in a networked system, said server comprising: an upstream manager receiving messages from said client and routing said messages to an appropriate service on said server, said upstream manager being coupled to a first network; a downstream manager sending a stream of said multimedia data from said appropriate service on said server to said client, said downstream manager being coupled to a second network; and
a connection service for maintaining information to connect said client, said upstream manager, said downstream manager, and said appropriate service on said server.
Col. 25 II. 1-16 (emphases added). The server stores and manages various types of data, including audio and video, and sends that data to client devices upon request. Col. 2 11. 31-38. The component of the server that sends the data to the clients is called a “service.”
The client and server communicate via components called the “upstream manager” and “downstream manager.” “The upstream manager 220(USM) accepts messages from [client devices] and routes them to services on the media server 100.” Col. 16 11. 11-13. The media server then supplies the requested service to the client device via the “downstream manager,” which “sends a stream of data [from the services on the media server] to a [client device].” Col. 16 II. 17-18.
The key innovation of the ’804 patent relates to how the upstream and downstream data are “addressed” to their recipients. Col. 13 11. 11-16. In computer networks, data is sent between the client and server in the form of “data packets,” each of which contains its own destination address. Col. 16 11. 24-39. The '804 innovation provided an approach that did not utilize the “physical” address of the recipient of the data packets. The physical *1327address is “[t]he actual, machine address of an item or device.” A. Freedman, The Computer Glossary (9th ed.2001). It “reflects the physical topology of the network .... ” Dictionary of Computing 9 (4th ed.1996) (defining “addressing”). However, keeping track of physical addresses is complicated when data packets must be sent between different types of networks that use different physical addressing schemes. Col. 13 11. 11-15. The present invention resolves this problem by “de-finfing] its own independent address space,” called a “logical” address space. Id.; see generally Col. 12 1. 31 — Col. 15 1. 56. A logical address “reflect[s] the administrative or functional relationships [among the addressed entities].” Oxford, supra at 9.1 Data packets that are sent between the server and client are routed using this logical addressing scheme. The specification states that “[i]t is important to note that all routing is accomplished based on logical addresses, not physical addresses.'” Col. 23 11. 1-2 (emphasis added).
The “upstream manager” is a key component in this system, and is required by every claim in the ’804 patent. The district court construed the term “upstream manager” as “a computer system component that (a) accepts messages from a client bound for services on a server; (b) routes messages from a client to services on a server; and (c) is distinct from the downstream manager.” nCUBE Corp. v. SeaChange Int’l, No. 01-11-JJF, slip op. at 9 (N.D.Del. May 23, 2002). Under this claim interpretation, it sustained the jury’s verdict that Seachange infringed claims 1-4, 6, 7, 9, 10, 12, and 14 of the ’804 patent. The majority upholds the district court’s claim construction. Seachange argues, and I agree, that this claim construction erroneously omitted the further requirement that the upstream manager route messages bound for services on the server using logical addresses. Under the proper claim construction, there was insufficient evidence to support the verdict of infringement because there is no evidence the accused device uses logical addresses for any purpose.
II
Under Phillips, “the words of a claim ‘are generally given their ordinary and customary meaning,’ [which] ... is the meaning that the [words] would have to a person of ordinary skill in the art in question at the time of the invention, i.e., as of the effective filing date of the patent application.’ ” Phillips v. AWH Corp., 415 F.3d 1303, 1312-13 (Fed.Cir.2005) (en banc) (internal citations omitted). The specification plays a key role in determining this meaning: “the specification is ‘always highly relevant to the claim construction analysis. Usually, it is dispositive; it is the single best guide to the meaning of a disputed term.” Id. at 1315 (quoting Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed.Cir.1996)). Ultimately, “[t]he construction that stays true to the claim language and most naturally aligns with the patent’s description of the invention will be, in the end, the correct construction.” Id. at 1316; see also V-Formation, Inc. v. Benetton Group SpA, 401 F.3d 1307, 1311 (Fed.Cir.2005) (restricting the meaning of the claim term “releasably attaching” to require that the attached components be easily removed and replaced). Here, the construction of “upstream manager” that most naturally aligns with *1328the ’804 patent’s description of the invention requires the use of logical addressing.
The use of logical addressing is a critical part of the invention embodied in the ’804 patent. A section of the patent entitled “The Network Protocol of the Present Invention” describes logical addressing, and explains why routing by logical addressing is important: “[b]ecause a packet may travel through several different types of underlying networks, each with their own [physical] addressing schemes, the network protocol of the present invention defines its own independent [logical] address space. This technique hides the many different types of addresses in use for each type of data link.” Col. 13 11. 11-16; see also Col. 12 l 31 — Col. 15 1. 56. Hence, “[t]he [logical addressing scheme] of the present invention provides the communication backbone that allows services scattered across heterogeneous, asymmetric networks to communicate with each other transparently.” Col. 12 1. 32-35. It “enables [services] to communicate transparently across the complex asymmetrical networks ....” Col. 15 11. 50-53.
As noted above, “[t]he upstream manager 220(USM) accepts messages from [client devices] and routes them to services on the media server 100.” Col. 16 11. 11-15. The specification reveals that the upstream manager accomplishes this routing using logical addressing. The patent describes the upstream manager and downstream manager as a “gateways that bridge ... different types of networks.” The upstream manager bridges these different types of networks by routing messages (which contain only logical addresses) from clients to their server destinations. Col. 16 11. 24-25. The downstream manager completes the bridge by directing the downstream data stream to its ultimate destination. As noted above, the use of the independent, logical addressing scheme “hides the many different types of [physical] addresses in use” in the underlying networks. Col. 13 11. 14-15. Because these physical addresses are hidden, the client device is relieved of the burden of identifying the upstream physical destinations of the messages it sends; it relies on the upstream manager to perform this routing function.
Nowhere does the ’804 patent disclose or suggest that the upstream manager routes messages from a client using physical, as opposed to logical addresses. Indeed, if the upstream manager could not route messages from the client using logical addresses, then the purpose of the invention — to substitute logical for physical addresses — would be defeated. See col. 13 11. 14-15. The patentee here offered no expert testimony suggesting that those skilled in the art would not read the patent to include a device using logical addresses.
Ill
The majority relies on three theories in support of the proposition that “the [upstream manager] may route messages using either logical or physical addresses.” Maj. Op., ante, at 1321 -1322.
First, although the majority appears to recognize that there is no reference in the specification to the use of physical addresses, the majority suggests that the specification’s explicit statement that “all routing is accomplished based on logical addresses, not physical addresses” applies only to the preferred embodiment. Col. 23 11. 1-2.
Of course, that a patent describes only a single embodiment does not mean the claims of the patent must be construed as limited to that embodiment. See, e.g., Liebel-Flarsheim Co. v. Medrad, Inc., 358 F.3d 898, 906 (Fed.Cir.2004). However, the claims will be read restrictively if the patentee has demonstrated a clear intention to limit the claim scope. Id.; see also *1329Teleflex, Inc. v. Ficosa N. Am. Corp., 299 F.3d 1313, 1327 (Fed.Cir.2002). Moreover, “the characterization of [a limitation] as part of the ‘present invention’ is strong evidence that the claims should not be read to encompass the opposite structure.” SciMed Life Sys., Inc. v. Advanced Cardiovascular Sys., Inc., 242 F.3d 1337, 1343 (Fed.Cir.2001); see also Wang Labs., Inc. v. America Online, Inc., 197 F.3d 1377, 1383 (Fed.Cir.1999); Modine Mfg. Co. v. U.S. Int’l Trade Comm’n, 75 F.3d 1545, 1551 (Fed.Cir.1996).
Here, the patentee clearly demonstrated that the upstream manager accomplishes routing by logical addressing. The specification characterizes the “present invention” as including the logical addressing limitation:
“[b]eeause a packet may travel through several different types of underlying networks, each with their own [physical] addressing schemes, the network protocol of the present invention defines its own independent [logical] address space. This technique hides the many different types of addresses in use for each type of data link.”
Col. 13 11. 11-16 (emphasis added). Under SciMed, the use of the term “present invention” is strong evidence that the use of logical addressing applies to the invention as a whole, not just the preferred embodiment. Moreover, as noted above, if the client could not send messages using logical addresses, the purpose of the “present invention” — to substitute logical for physical addresses — would be defeated.
Second, the majority urges that reading claim 1 to include a logical addressing-limitation would “impermissibly read the ‘virtual connection’ limitation of claim 2 into claim 1, making these claims redundant.” Maj. Op. at 1321 -1322. Claim 2 recites: “The server in claim 1 wherein said connection service further creates a virtual connection between an upstream address and a downstream address for said client.” Col. 25 11. 17-20. Even assuming that logical addressing is sufficient to create a virtual connection, reading claim 1 to include logical addressing does not make the two claims redundant, because claim 1 does not contain a requirement that the connection service create a virtual connection. Claim 2 adds simply that requirement. There is thus no inconsistency in reading “upstream manager” in claim 1 to require logical addressing.
Third, the majority points out that unas-serted claims 5 and 11 specifically refer to logical addresses. Maj. Op. at 1321 -1322. In contrast, the asserted claims do not. The majority suggests that the presence of the references to logical addresses in the unasserted claims indicates that the omission of logical addresses from the asserted claims was intentional; hence, the asserted claims do not require logical addressing. Id.
In my view the majority misreads claims 5 and 11 to add the requirement that routing be accomplished by logical addressing. Id. Rather, these claims simply recite a specific means of using logical addresses that are already required by the upstream manager. Claim 5 recites:
5. The computer-implemented method in claim 4 wherein further comprising the steps of:
receiving a service request message from said client to said server via said upstream manager, said service request corresponding to said service on said server, said service request message including said client downstream logical address and a service destination logical address;
generating a response message from said server to said client, said response message including said client downstream logical address; and
sending said response message to said client via said downstream manager. *1330Col. 25,11.43-55 (emphases added). Claim 5 describes a method of handling logical addressing when a specific type of message — a service request message — is sent by the client device. It describes a situation in which the client logical address and downstream logical address are provided in the service request message, and a response is generated which includes the downstream logical address. The purpose of the claim is to claim that particular method, not to add a requirement for logical addressing. Therefore the omission of the term “logical address” from the asserted claims does not reflect the claimant’s intent that logical addressing not be part of those claims. The same analysis applies to Claim 11, which rewrites Claim 4 in means-plus-funetion form.2
In my view the upstream manager includes a requirement that the upstream manager route messages using logical addresses. Because nCube presented no evidence that the alleged upstream manager in the accused device uses logical addresses to send messages to any service on the server, I would reverse the verdict of infringement. I thus would not reach the question — addressed by the majority— whether all client communications must be routed through the upstream manager.
. A telephone number is an example of a logical address, while the port to which the telephone is connected is a physical address. Newton’s Telecom Dictionary 645 (21st ed.2005). "A logical address ... may have no fixed physical address.” Id. That is, a person might move from one home to another, keeping the same telephone number.
. Claim 11 recites:
11. The server as claimed in claim 10 further including:
means for receiving a service request message from said client via said upstream manager, said service request corresponding to said service on said server, said service request message including said client downstream logical address and a service destination logical address; means for generating a response message to said client, said response message including said client downstream logical address; and
means for sending said response message to said client via said downstream manager.
Col. 26, 11. 32-46. Claim 11 describes a server which includes several means directed to the handling of logical addressing when a specific type of message is sent by the client device. The purpose of the claim is to claim those particular means, not to add a requirement for logical addressing.