by Dennis Crouch
The Federal Circuit recently issued an important decision refining how courts should analyze infringement of system claims when multiple parties are involved in operating different components of the claimed system. This post looks at the court’s new decision in CloudofChange, LLC v. NCR Corp., No. 23-1111 (Fed. Cir. Dec. 18, 2024) and contrasts it Centillion Data Systems, LLC v. Qwest Communications International, Inc., 631 F.3d 1279 (Fed. Cir. 2011).
These cases highlight a recurring challenge in patent law – how to analyze infringement when a claim requires components controlled by different actors. The issue is particularly salient for modern technology systems that often involve both vendor-operated backend servers and customer-operated frontend devices.
The Centillion Framework
Centillion involved a system for processing and delivering telephone billing information. The claimed system required both backend components (operated by the service provider) and frontend components (operated by customers). The court had to determine whether Qwest, which operated only the backend system, could be liable for direct infringement when customers controlled the frontend components.
The Federal Circuit established that to “use” a system under 35 U.S.C. § 271(a), a party must “control the system as a whole and obtain benefit from it.” Importantly, this does not require “direct or physical control over each individual element,” but rather the ability to put the system as a whole into service for its intended purpose. Applying this framework, the Centillion court held that Qwest’s customers, not Qwest itself, “used” the system because they initiated queries that caused the backend processing and benefited from the results. The court explained that “but for the customer’s actions, the entire system would never have been put into service.”
The court then considered whether Qwest could be vicariously liable for its customers’ use. Finding that Qwest “in no way directs its customers to perform nor do its customers act as its agents,” the court held there was no vicarious liability. The fact that Qwest provided software and technical assistance was insufficient – the key point was that customers decided independently whether to install and operate the software.
The CloudofChange Decision
The CloudofChange case presented a similar but distinct scenario involving NCR’s point-of-sale (POS) system called NCR Silver. The claimed system in CloudofChange’s patents required both vendor-operated web servers and subscriber-operated POS terminals. Like in Centillion, NCR operated the backend servers while merchants controlled the frontend POS terminals.
Like the customers in Centillion, NCR’s merchants put the system into service because they initiate at the POS terminal a demand for service (for example, building or editing a POS) and benefit from the back end providing that service.
CloudofChange pursued only a direct infringement theory based on “use,” arguing that NCR controlled and benefited from each component of the claimed system under Centillion. The district court agreed, finding NCR vicariously liable for its merchants’ use because the NCR Silver Merchant Agreement required merchants to obtain and maintain internet access.
On appeal, however, the Federal Circuit reversed, providing important clarification of the Centillion framework. First, the court affirmed that NCR’s merchants, not NCR, “used” the claimed system under Centillion because they initiated actions at the POS terminals and benefited from the system’s POS building functionality. The court emphasized that the relevant “benefit” under Centillion is not general commercial benefit (like subscription fees) but rather the benefit of putting the claimed system to its intended use.
On the question of vicarious liability, the Federal Circuit held that requiring merchants to maintain internet access does not amount to directing or controlling their use of the system as a whole. The court explained that the district court erred by focusing its vicarious liability analysis on control of a single element (internet access) rather than control over the use of the entire system.
Importantly, the court distinguished between vicarious liability analysis for method claims versus system claims. For method claims, courts are instructed to take an step-by-step approach — i.e., the defendant must direct or control each step of the method. See Akamai Technologies, Inc. v. Limelight Networks, Inc., 797 F.3d 1020 (Fed. Cir. 2015). The insight here is that method and system claims fundamentally differ in how they are “used” or “practiced.” A method claim consists of a series of steps that must be performed sequentially or in a particular order – it’s a process that unfolds over time. In contrast, a system claim describes a collection of components working together as a unified whole at a single point in time. This fundamental difference affects how the Federal Circuit thinks about vicarious liability and control. For a method claim, each step represents a discrete action that must be performed by someone. When these steps are divided among multiple parties, we need to ensure that one party is actually controlling or directing the performance of each individual step – otherwise, no single party can be said to be “performing” the complete method.
For system claims, however, the analysis focuses on control over the system as a whole rather than individual components. What matters is whether someone is putting the complete system into service and benefiting from its collective operation.
In its seminal NTP v. RIM decision from 2005, the Federal Circuit established key principles for divided infringement — but from a different perspective. NTP, Inc. v. Research in Motion, Ltd., 418 F.3d 1282 (Fed. Cir. 2005). Rather than having different entities performing aspects of the infringing act, NTP involved infringement actions that occurred across national boundaries. This created a legal question mark because US law only creates an action for uses “within the United States.” The court distinguished between system and method claims, explaining that while a system claim could be infringed when components were split across national boundaries (as long as control and benefits were in the U.S.), method claims required each step to be performed within U.S. territory. This watershed case provided the analytical framework for cross-border divided infringement that continues to guide courts today.
I am a little confused by the opening paragraph (and the title) in the recent article about the recent CloudofChange decision. Both the opening paragraph and the title imply that there is a contrast in the Federal Circuit’s decisions in Centillion Data Systems, LLC. and CloudofChange, LLC. After reading the article and both decisions, I must admit that I do not see the contrast. Can someone please enlighten me?
The law of divided infringement has developed in a similar way to the doctrine of equivalents. In the early days, when patent law was sleepy and there was a fairly low general level of skill amongst patent attorneys, the patent law sometimes needed to save the patentees from the gaps in skill of their patent attorneys. Over time the general level of skill grew, somewhat in parallel with the value of patents. Accordingly, with less need to help the patentees from the skill deficits of their patent lawyers, the patentees got less wiggle room. Who today writes claims expecting that the doctrine of equivalents will save them? Likewise, nobody today should be writing claim sets expecting that divided infringement will work in their favor. Sure, it’s ok to have two-sided claims, but then look closely at them and write another set from one side or the other, or maybe from both sides.
As a product of “the early days, when patent law was sleepy and there was a fairly low general level of skill amongst patent attorneys” as you put it, I have to agree that it was sleepy in the sense that patent application preparation was not so cost-driven by clients and by high patent attorney salaries. More time could be spent on it. Are todays young law firm associates, under far more such pressure, and so little partner supervision, really much better at it? Ask an older PTO examiner if they think application quality has improved or actually gotten worse – I have.
Is not the biggest change in the type of inventions now being patented the main source of the subject divided infringers issues? No longer primarily machines or other products made by one company, but rather sofware and internet information selection and display ideas like the one here? Also, in other areas, technology in general is now far more advanced and often far more complex, with far more prior art, without comensurate increases in application billings. But even in the 18th century [even before my time] patent enforcement was expensive, slow, and difficult, especially to try to win with DOE.
Paul, you claim to be from the early days of US patent law. That would make you easily over 150 years old. Perhaps when considering time it would be better to consider a broader arc, and not merely your own lifetime. From the broader arc, your challenge of my hypothesis loses weight. Your comments also seem to reflect a disregard of patent enforcement, which was the focus of the post and my comment, and a narrow view of patent prosecution. US examiners are probably a poor basis for assessing patent quality, but they are a good basis to assess how patent applications have changed while keeping in mind that examination has also changed a lot during your career and theirs.