][co-author: Diego F. Freire]
In 2006, Akamai Technologies ("Akamai") sued Limelight Networks, Inc. ("Limelight") in the U.S. District Court for the District of Massachusetts, alleging infringement of U.S. Patent No. 6,108,703. The '703 patent is assigned to the Massachusetts Institute of Technology ("MIT") and is exclusively licensed to Akamai.
Claim 34 of the '703 patent recites:
A content delivery method, comprising:
distributing a set of page objects across a network of content servers managed by a domain other than a content provider domain, wherein the network of content servers are organized into a set of regions;
for a given page normally served from the content provider domain, tagging at least some of the embedded objects of the page so that requests for the objects resolve to the domain instead of the content provider domain;
in response to a client request for an embedded object of the page;
resolving the client request as a function of a location of the client machine making the request and current Internet traffic conditions to identify a given region; and
returning to the client an IP address of a given one of the content servers within the given region that is likely to host the embedded object and that is not overloaded. [Emphasis added.]
The claimed invention is directed to delivering electronic data using a content delivery network (CDN). It purports to facilitate faster delivery of the data by separating the content of a website onto multiple servers. The content that requires greater network capacity (such as photos and videos) can be assigned to servers ("tagged") that provide this content at faster speeds. The remaining content can be provided by non-specialized servers.
Limelight operates a CDN, and content providers are its customers. Limelight carries out three of the four claimed steps (the distributing, resolving, and returning), but did not tag components of its customers' websites -- instead, Limelight contractually required its customers to do their own tagging, if those customers wanted to exploit the faster servers. The relevant language of the contract stated "Customer [i.e., content provider] shall be responsible for identifying via the then current [Limelight] process all [URLs] of the Customer Content to enable such Customer Content to be delivered by [Limelight]," and "Customer shall provide [Limelight] with all cooperation and information necessary for [Limelight] to implement the [Content Delivery Service]."
At trial, the jury found that Limelight and its customers jointly and directly infringed '703 patent under 35 U.S.C. § 271(a), and awarded $40 million in damages. Following this verdict, the Federal Circuit decided Muniauction, Inc. v. Thomson Corp. In Muniauction, the Court held that direct infringement of a claimed method requires that a single entity perform every step of the claim (the "single entity rule"). But, this requirement is satisfied if steps are performed by multiple parties provided that a single defendant exercises "control or direction" over entire process. Thus, neither party is liable for infringement if they perform all of the steps, but merely engage in an arms-length relationship to do so.
Limelight moved for judgment of non-infringement as a matter of law (JMOL) in view of Muniauction, and the District Court granted the motion, holding that because (i) no single entity performed all of the claimed steps, and (ii) the contractual relationship between Limelight and its customers did not rise to the level of "control or direction" there was no liability. On appeal, a Federal Circuit panel affirmed, but the en banc Court heard the case, reversed, and remanded the case for further proceedings. Particularly, the en banc majority reasoned that Limelight and its customers did not directly infringe, but "the evidence could support a judgment in its favor on a theory of induced infringement [under 35 U.S.C. § 271(b)]" because "inducement does not require that the induced party be an agent of the inducer or be acting under the inducer's direction or control." The Court, however, also stated that "there can be no indirect infringement without direct infringement."
In a June 2014 appeal, the Supreme Court took issue with this apparent contradiction, and held that a defendant is not liable for inducing infringement under 35 U.S.C. § 271(b) when no one party has directly infringed the patent under § 271(a) (see "Limelight Networks, Inc. v. Akamai Technologies, Inc. (2014)"). Thus, the high court reversed the Federal Circuit finding that Limelight had infringed the '703 patent. On remand, the Federal Circuit considered whether Limelight has infringed under § 271(a). Judge Linn authored the opinion of the court, and was joined by Chief Judge Prost. Judge Moore dissented.
2. The Outcome
Ultimately, the majority ruled that "because Limelight . . . did not perform all of the steps of the asserted method claims . . . and because the record contains no basis on which to impose liability on Limelight for the actions of its customers who carried out the other steps, Limelight has not directly infringed the '703 patent under § 271(a)." The majority confirmed that "direct infringement liability of a method claim under 35 U.S.C. § 271(a) exists when all of the steps of the claim are performed by or attributed to a single entity -- as would be the case, for example, in a principal-agent relationship, in a contractual arrangement, or in a joint enterprise." Here, there was no liability "[b]ecause this case involves neither agency nor contract nor joint enterprise."
In dissent, Judge Moore disagreed. She believes that "§ 271(a) includes joint tortfeasor liability." Characterizing the majority's rule as creating "a gaping hole in what for centuries has been recognized as an actionable form of infringement," she opined that the single entity rule "is a recent judicial creation inconsistent with statute, common law, and common sense."
Instead, the majority found that "[e]ncouraging or instructing others to perform an act is not the same as performing the act oneself and does not result in direct infringement [as] § 271 has separate subsections addressing induced and contributory infringement." In coming to this conclusion, the majority relied heavily on the structure of § 271, noting that "[w]hen a party participates in or encourages infringement but does not directly infringe a patent, the normal recourse under the law is for the court to apply the standards for liability under indirect infringement." But, "indirect infringement requires, as a predicate, a finding that some party is directly liable for the entire act of direct infringement."
Akamai attempted to argue that "an accused infringer 'directs or controls' a third party if the accused infringer goes beyond loosely providing instructions and specifically tells a third party the step or steps to perform." The majority disagreed based on the organization and legislative history of the statute. Notably, "§ 271(a) includes only the principles of vicarious liability, as embodied in the single entity rule." Despite being presented with "numerous conflicting theories of joint liability that existed in the common law prior to 1952, Congress enacted specific rules for inducement and contributory liability in § 271(b) and (c), respectively." The majority also explained its disagreement with the dissent's argument -- "[w]hile the dissent believes this leaves a 'gaping hole,' it is not our position to legislate or contravene Congress' choice -- right or wrong -- by importing other theories of joint liability into § 271(a)." Particularly, "the dissent advocates holding a customer jointly and severally liable for patent infringement based on its performance of a single step of a claimed method, even when it has no knowledge of the patent."
3. The Majority's Reasoning
The majority carried out its substantive analysis of the case in three steps. First, in reviewing the statutory scheme of § 271, the majority pointed out that patent infringement is not a creation of common law but a tort defined by statute. While § 271(a) is "a declaration of what constitutes infringement," subsections (b) and (c) codify the doctrines of inducement and contributory infringement respectively. The Court states that "in this way, Congress carefully crafted subsection (b) and (c) to expressly define the only ways in which individuals not completing an infringing act under § 271(a) could nevertheless be liable." And "[t]he fact that Congress chose to impose some forms of secondary liability, but not others, indicates a deliberate congressional choice with which the courts should not interfere."
The majority then addressed the dissent's argument (which was also promulgated by Akamai) that the word "whoever" in § 271(a) is plural and therefore undermines the single entity rule. In the majority's view, despite "whoever" being plural, more than one entity can be independently liable for direct patent infringement if each entity practices every step of the claim. Furthermore, interpreting § 271(a) to include actors who do not independently infringe would make § 271(b) and (c) redundant. In the majority's view, such a construction contravenes the canon against surplusage in statutes.
Second, the majority analyzed divided infringement case law, which is rooted in traditional principles of vicarious liability. This case law sets forth the single entity rule and that "multiple actors could together infringe a patent only if one controlled the other(s)." Particularly, "when a contract mandates the performance of all steps of a claimed method, each party to the contract is responsible for the method steps for which it bargained . . . [h]owever, this type of contractual arrangement will typically not be present in an arms-length seller-customer relationship." Still, "the concerns over a party avoiding infringement by arms-length cooperation can usually be offset by proper claim drafting [because] a patentee can usually structure a claim to capture infringement by a single entity."
Third, the majority found errors in the notion of importing joint tortfeasor liability into § 271(a). It acknowledged that "liability exists under traditional principles of vicarious liability, such as where a mastermind directs or controls another to perform all steps of a claimed method." But, the majority indicated that the dissent departed from the common law limits on joint tortfeasor liability in order to import this law into § 271(a). Particularly, "actors whose innocent actions coordinate to cause harm generally are not subject to liability at common law," but would be under the dissent's interpretation. Further, the common law "requires both parties to know the others' actions to act in concert" such that liability is found. Thus, "when parties act independently for their own benefit, such as in arms-length seller-customer relationships," the necessary concerted action between the parties is missing. Additionally, for liability under § 271(a), the dissent would require the parties to know that their joint conduct would cause damage. This is inconsistent with the statute and the Supreme Court's precedent that "a direct infringer's knowledge or intent is irrelevant."
The majority further indicated that the dissent's reasoning would lead to unexpected results. In one scenario, "a customer who performs a single step of a patented method by merely using a product as intended would . . . liable for direct infringement." Doing so would fill the "gaping hole" left open by divided infringement, but open a much larger door where innocent customers would be liable for patent infringement for their simple actions.
Lastly, the majority pointed out that the dissent's position would conflict with the long established rule that "a dependent claim cannot be infringed unless the independent claim from which it depends is also infringed." Applying this logic a hypothetical set of claims in which an independent claim recites a replicating step and a dependent claim adds a tagging step, "the dissent's rule . . . would result in a party that performs the tagging step but not the replicating step being liable for infringing [the dependent claim] while not being liable for infringing the broader claim from which it depends."
The majority concluded by pointing out that "Limelight's customers decide what content, if any, they choose to have delivered by Limelight's CDN and only then perform the [tagging step]." This "form contract does not obligate Limelight's customers to perform any of the method steps." Based on this interpretation, it was not established that "Limelight's customers were acting as agents of or otherwise contractually obligated to Limelight or that they were acting in a joint enterprise."
As the law regarding divided infringement currently stands, it is possible to use a contract between a company and its customer to avoid infringement of a patented method when those parties split carrying out the patented steps. The fact pattern of this case provides a recipe for doing so. As a result, it is more critical than ever that claim drafters focus method claims so that their steps are likely to be carried out by a single entity. For example, Claim 34 of the '703 patent could be written as "for a given page normally served from the content provider domain, obtaining a tagged representation of at least some of the embedded objects of the page . . . ."
Still, this case may not be at an end. The dissent's final argument was that an en banc action is required to do what the majority proposes, as it is at odds with binding precedent from a previous panel. On June 12th, Akamai requested such an en banc review.
Akamai Technologies, Inc. v. Limelight Networks, Inc. (Fed. Cir. 2015)
Panel: Chief Judge Prost and Circuit Judges Linn and Moore
Opinion for the court by Circuit Judge Linn; dissenting opinion by Circuit Judge Moore