When "One or More Processors" Still Requires One: Functional Allocation After Salazar and Finjan

Executive Summary: Patent claims frequently recite “one or more processors configured to” perform multiple functions, but that language does not necessarily permit several processors to divide those functions among themselves. Federal Circuit decisions including Convolve, Salazar, and Finjan distinguish numerical scope from functional identity: although “a processor” ordinarily encompasses one or more processors, at least one qualifying processor may still need to perform—or be configured to perform—every function assigned to the claimed processor. In re Varma’s memorable dog analogy illustrates the point: multiple elements generally cannot split the characteristics attributed to a single claimed element. Collective performance may nevertheless be permitted when the claim assigns the functions to a composite unit, such as a system, server node, or user equipment, and the intrinsic record supports a distributed implementation. This distinction can determine infringement, anticipation, obviousness, and the effect of prosecution arguments. Practitioners should therefore state expressly whether one processor must perform all claimed functions, multiple processors may act collectively, the system as a whole is the functional subject, or particular operations must be performed by distinct processors. A few carefully chosen words during drafting can determine whether a claim reaches modern distributed architectures.

I. Introduction: Distributed Computing Meets Singular Claim Grammar

 

Patent claims commonly recite "one or more processors configured to" perform a sequence of operations. The formulation appears broad and technologically neutral. It avoids limiting the invention to a particular processor count and accommodates devices ranging from a single-chip controller to a cloud platform containing thousands of processors. But the phrase may provide less flexibility than its plural wording suggests.

Consider a claim requiring one or more processors configured to receive data, analyze the data, and generate an output. The accused system uses an edge processor to receive the data, a cloud processor to analyze it, and a client processor to generate the output. The system contains multiple processors, and the processors collectively perform every claimed function. Yet no individual processor performs—or is configured to perform—all three functions. Does the system satisfy the limitation?

Federal Circuit precedent increasingly separates that question into two distinct inquiries. The first concerns numerosity: does "a processor" encompass one or more processors? Ordinarily, it does. The second concerns functional identity: may several processors divide the functions attributed to the claimed processor? The answer is frequently no. Even where "a processor" or "one or more processors" permits multiple processors to be present, the claim may require at least one processor to possess every capability assigned to that processor.

That distinction has significant consequences. Modern computing systems routinely allocate tasks among central processing units, graphics processing units, dedicated accelerators, edge devices, remote servers, and software services. A claim drafted as though a single processor performs every operation may fail to cover a commercial embodiment even when the system as a whole practices the invention. The same drafting choice can affect validity: a prior-art system may collectively disclose all claimed functions without disclosing any one processor capable of performing them all.

The governing cases do not establish an absolute rule against distributed performance. Instead, they make claim grammar, antecedent basis, component hierarchy, and the intrinsic record decisive. The emerging lesson is that "one or more" answers how many processors may exist; it does not necessarily answer how the claimed functions may be allocated among them. This article addresses that allocation question, rather than the separate body of law concerning whether "configured to" requires present programming, structural adaptation, or mere capability.

II. "A" Ordinarily Means "One or More"—But That Is Only the Beginning

 

Patent law starts with a familiar presumption. In an open-ended claim using "comprising," the indefinite articles "a" and "an" ordinarily mean "one or more." KCJ Corp. v. Kinetic Concepts, Inc., 223 F.3d 1351, 1356 (Fed. Cir. 2000). Departures from that rule are narrowly confined and ordinarily require a clear indication in the claims, specification, or prosecution history that the patentee intended a singular-only meaning. Baldwin Graphic Systems, Inc. v. Siebert, Inc., 512 F.3d 1338, 1342-43 (Fed. Cir. 2008).

Later use of a definite article generally does not, by itself, convert the antecedent into a singular element. Thus, if "a processor" means one or more processors, a subsequent reference to "the processor" or "said processor" ordinarily refers back to that potentially plural antecedent. Id. at 1342. The Federal Circuit reaffirmed that principle in holding that "a sample stream," and therefore the later-referenced "the sample stream," could encompass more than one stream. ABS Global, Inc. v. Cytonome/ST, LLC, 84 F.4th 1034, 1040-42 (Fed. Cir. 2023).

This plural rule, however, addresses the permissible number of claimed elements. It does not erase the substantive requirements attached to an element. Nor does the open-ended transition "comprising" make every limitation within a claim open-ended. "Comprising" permits an accused system to contain additional, unrecited features, but the system must still satisfy each recited limitation. Dippin' Dots, Inc. v. Mosey, 476 F.3d 1337, 1343 (Fed. Cir. 2007).

The Federal Circuit explained this distinction memorably in In re Varma. The claims required a statistical-analysis request corresponding to two or more selected investments. The Patent Trial and Appeal Board reasoned that the requirement could be satisfied by separate requests, each corresponding to a different investment, because the claim used "comprising" and introduced "a" request. The Federal Circuit rejected that reading. Although the system could receive multiple requests, at least one request still had to possess the claimed characteristic of corresponding to two or more investments. In re Varma, 816 F.3d 1352, 1362-63 (Fed. Cir. 2016).

"For a dog owner to have 'a dog that rolls over and fetches sticks,' it does not suffice that he have two dogs . . . ." In re Varma, 816 F.3d 1352, 1363 (Fed. Cir. 2016).

The omitted ending explains the point: one dog cannot supply only the rolling-over characteristic while another supplies only the fetching characteristic. The indefinite article permits the owner to have multiple dogs, but at least one qualifying dog must possess both characteristics.

Applied to processor claims, Varma supplies the basic doctrinal distinction. A system may contain several processors without departing from a limitation reciting "a processor." But if the claim describes a processor configured to perform A, B, and C, the plural rule does not necessarily permit one processor to perform A, another to perform B, and a third to perform C. At least one instantiation of the claimed processor may need to satisfy the complete functional description.

That conclusion is not automatic in every claim. The relevant question remains what the claim assigns to the processor, rather than merely whether the noun "processor" appears in singular or plural form. The Federal Circuit's decision in Convolve demonstrates how small differences in claim structure can change the result.

III. Convolve: Two Processor Constructions in One Patent

 

Convolve, Inc. v. Compaq Computer Corp. involved patents directed to reducing hard-drive acoustic noise. The accused computers employed different components in generating a user interface and issuing commands to the hard drive. The district court treated the asserted claims as though they all required one processor associated with the user interface to generate the claimed commands. The Federal Circuit held that this approach overlooked material differences among the claims. 812 F.3d 1313, 1320-21 (Fed. Cir. 2016).

Claims 9 and 15 recited an apparatus comprising "a processor" executing steps to generate a user interface, alter settings, and output commands to the data-storage device. Applying the ordinary plural rule, the court construed "a processor" to mean one or more processors. Nothing in those claims, the specification, or the prosecution history clearly limited the apparatus to a single processor. The specification reinforced the plural construction by disclosing an embodiment in which a separate controller dedicated to the disk drive controlled the relevant seeks. The Federal Circuit therefore held that the district court erred by imposing a single-processor requirement on those claims. Id. at 1320-21.

Claims 1, 3, and 5 produced the opposite result. Those claims introduced "a processor" in the preamble and later used "the processor" in the body. That antecedent structure tied the processor associated with the user interface to the later command-outputting function. The claims accordingly required the same processor to participate in performing the recited steps. Id. at 1321.

The importance of Convolve lies in this internal comparison. The court did not announce that "a processor" always permits functional division, nor did it hold that every processor limitation requires one processor to perform everything. It examined how each claim grammatically assigned the functions. The claims that expressly referred back to "the processor" required a common processor. The claims lacking that same linguistic tie, when read with a specification disclosing a separate controller, permitted multiple processors.

For practitioners, Convolve identifies two recurring sources of claim scope. First, antecedent basis can do substantive work. "The processor" and "said processor" may connect functions that otherwise could have been assigned to different components. Second, the specification can confirm that distributed processing is an intended embodiment. A generic statement that a system may include multiple processors is less useful than an express disclosure that separate processors may perform particular claimed operations.

Convolve nevertheless left room for a further question: what happens when a claim does not repeatedly say "the processor," but its overall grammar still assigns a series of functions to the processor? The Federal Circuit addressed that question more directly in Salazar and Finjan.

IV. Salazar: Numerosity Is Not Functional Identity

 

The claim language in Salazar v. AT&T Mobility LLC introduced "a microprocessor" for generating control signals and then recited functions performed by "said microprocessor," including creating communication protocols, retrieving parameter sets, and generating command-code sets. 64 F.4th 1311, 1313-14 (Fed. Cir. 2023). The parties agreed that the claimed system could contain multiple microprocessors. Their dispute concerned whether the functions could be divided among those microprocessors.

The patent owner argued that the plural meaning of "a microprocessor" permitted one microprocessor to perform one function and another microprocessor to perform a different function, even if no microprocessor could perform every recited function. The district court disagreed and construed the limitation to cover one or more microprocessors, at least one of which was configured to perform all the generating, creating, and retrieving functions. Following a jury verdict of noninfringement, the patent owner appealed. Id. at 1314-15.

The Federal Circuit affirmed. It preserved the general rule that "a microprocessor" does not limit the system to one microprocessor. It also acknowledged that "said microprocessor" refers back to the same potentially plural antecedent. But those principles did not permit the patentee to divide the functions among unrelated members of the group. At least one microprocessor had to be capable of performing every function attributed to "said microprocessor." Id. at 1317-18.

The court relied on both branches of Convolve. The claim resembled claims 1, 3, and 5 of Convolve, where the later reference to "the processor" tied the recited functions to the previously introduced processor. It did not resemble claims 9 and 15, whose language and intrinsic record permitted multiple processors. Salazar, 64 F.4th at 1316-17. The court also invoked Varma: the word "said" could not negate the functional requirements that followed it, just as multiple dogs could not divide the characteristics of rolling over and fetching. Id. at 1318.

Properly understood, Salazar did not adopt a singular construction. An accused system may contain many processors, and more than one processor may independently satisfy the complete limitation. The requirement is existential rather than numerical: among the processors encompassed by the claim, at least one must be capable of performing the entire set of processor functions.

Nor did Salazar squarely construe every possible formulation using the express phrase "one or more processors configured to." Its holding arose from claim language repeatedly referring to "said microprocessor." Nevertheless, the decision sharply limits the argument that plural wording alone authorizes collective performance. After Salazar, a patentee must identify something in the claim structure or intrinsic record beyond the ordinary plural meaning of "a" to support dividing recited functions among processors.

V. Finjan and the Post-Salazar Same-Component Rule

 

Six months after Salazar, the Federal Circuit extended the same reasoning in Finjan LLC v. SonicWall, Inc., 84 F.4th 963 (Fed. Cir. 2023). The asserted claims concerned computer-based malware analysis. Some claims recited steps performed "by a computer" and later "by the computer." Another required program code for causing "a computer to perform" a listed series of steps. The accused products divided the relevant operations among separate remote computers. Finjan argued that the plural meaning of "a computer" permitted those computers to perform the steps collectively. Id. at 973-74.

The Federal Circuit rejected that position. References to "the computer" tied the later functions to the antecedent computer, consistent with Salazar, Convolve, and Traxcell Technologies, LLC v. Nokia Solutions & Networks Oy, 15 F.4th 1136, 1143-44 (Fed. Cir. 2021). Even if the accused system could use one or more computers, at least one computer had to perform all functions listed in the claims. Finjan, 84 F.4th at 974-75.

Claim 22 makes Finjan particularly important. It did not depend on a series of later references to "the computer." Instead, it required program code for causing "a computer to perform" the enumerated steps. That introductory language itself assigned the steps to the computer. The court therefore required one computer to perform all of them. Id. at 975. Finjan accordingly prevents a patentee from avoiding the same-component rule merely by omitting repeated uses of "the" or "said" when the claim's grammar still makes one computer responsible for the listed operations.

The court also distinguished 01 Communique Laboratory, Inc. v. LogMeIn, Inc., 687 F.3d 1292 (Fed. Cir. 2012), which had permitted a software-based location facility to be distributed among multiple locator-server computers. 01 Communique addressed whether the claimed facility could be implemented as a composite, distributed component, supported by express specification language describing multiple computers and subdivided facilities. It did not decide whether separate computers could divide several later limitations attributed to the same computer. Finjan, 84 F.4th at 974-75.

Later decisions demonstrate the practical reach of the rule. In Carrum Technologies, LLC v. Ford Motor Co., the Federal Circuit held that although "a controller" permitted one or more controllers, at least one controller had to perform every function attributed to "said controller." Nos. 2024-1183, 2024-1480, slip op. at 8-12 (Fed. Cir. Oct. 15, 2025) (nonprecedential). The court characterized the distinction directly: the plural rule leaves numerical scope open but does not relax the substantive requirement that at least one instantiation embody all assigned functions when the claim context requires it. Id. at 8-10.

District courts have applied Finjan even where the processor or computer appears only in the preamble. In Equil IP Holdings LLC v. Akamai Technologies, Inc., the claim recited a method "in a host computer" without later repeating "the host computer." The court nevertheless held that at least one host computer had to perform every claimed step because the preamble made that computer responsible for implementing the method. Multiple computers could participate, but one of them still had to perform all steps. No. 22-677-RGA, slip op. at 5-8 (D. Del. July 2, 2024).

The rule likewise creates an evidentiary burden at infringement. In Cellspin Soft, Inc. v. Fitbit LLC, the claims required a "first processor" configured to acquire data, store it, and transmit an event notification and the data. The accused Apple Watch contained multiple processors, but the patentee's evidence did not show that any identified processor was configured to perform all three functions. The Federal Circuit affirmed summary judgment of noninfringement. Nos. 2022-2025, -2028, -2029, -2030, -2032, -2037, slip op. at 17-18 (Fed. Cir. Nov. 1, 2024) (nonprecedential).

The PTAB has also applied Salazar to processor language. In Apple Inc. v. Smart Mobile Technologies LLC, claims recited "a processor" followed by "the processor is configured to" process multiple data streams or channels. The Board held on remand that the same processor had to perform all operations executed by "the processor." IPR2022-01002, Paper 52, at 17-21 (P.T.A.B. Apr. 22, 2026) (final written decision on remand). Director Review was later denied. Apple Inc. v. Smart Mobile Technologies LLC, IPR2022-01002, Order at 2 (P.T.A.B. July 24, 2026).

Together, Salazar and Finjan establish the controlling starting point: plural scope does not itself authorize functional fragmentation. The analysis must identify the claimed functional subject and determine whether the claim assigns the complete set of operations to that subject. When it does, an accused system ordinarily must contain at least one qualifying processor or computer capable of performing them all.

VI. The Collective-Performance Line: When Multiple Machines May Cooperate

 

The same-component rule is not necessarily a single-chip rule. Federal Circuit cases leave room for distributed performance when the claim assigns the functions to a composite element—such as a server node, location facility, user equipment, or system—that may itself comprise several processors or computers. The critical question remains the identity of the claimed element that must possess all recited capabilities.

01 Communique Laboratory, Inc. v. LogMeIn, Inc. remains the principal precedential example. The claims recited "a locator server computer" containing a "location facility" that performed several functions. The accused architecture distributed those functions among multiple computers, and no single computer performed all of them. The Federal Circuit nevertheless rejected a construction requiring one physical computer because the specification expressly stated that the server computer could comprise one or more computers and that its facilities could be subdivided. Those disclosures supported distributing the location facility among multiple physical computers. 687 F.3d 1292, 1296-98 (Fed. Cir. 2012).

Symantec Corp. v. Computer Associates International, Inc. similarly recognized that the ordinary meaning of "computer" and "computer system" was not limited to a single stand-alone machine. 522 F.3d 1279, 1290-91 (Fed. Cir. 2008). But Finjan later made clear that treating a computer or system as composite does not automatically eliminate the requirement that the same claimed component perform the functions assigned to it. Finjan, 84 F.4th at 975.

The nonprecedential decision in Unwired Planet, LLC v. Google Inc. illustrates the distinction between a claimed composite element and the computers implementing it. The district court had construed "server node" to mean one or more computers, each of which performed every receiving, accessing, processing, and transmitting function. The Federal Circuit rejected that requirement. The claim required one server node to perform every function, but it did not require each computer within the node to perform every function. Multiple computers or programs could work in concert as the claimed server node. 660 F. App'x 974, 980-81 (Fed. Cir. 2016).

These decisions do not conflict with Salazar or Finjan. In Salazar, the functions were attributed to "said microprocessor," and therefore at least one microprocessor had to possess all recited capabilities. Salazar, 64 F.4th at 1317-18. In Unwired Planet, the functions were attributed to a server node, and therefore the node—not each constituent computer—had to perform all of them. Unwired Planet, 660 F. App'x at 980-81. The cases can be reconciled by asking whether the claim's functional subject is an individual processor or a higher-level composite element.

The Federal Circuit's nonprecedential decision in Sonos, Inc. v. International Trade Commission underscores the importance of specification support. Sonos argued that separate messages could collectively carry an identifier and a security key. The court disagreed, requiring each qualifying message to contain both. In distinguishing 01 Communique, the court emphasized that distributed performance there rested on an express specification disclosure permitting facilities to be subdivided. No. 2022-1421, slip op. at 8-10 (Fed. Cir. Apr. 8, 2024) (nonprecedential). Generic boilerplate stating that a system may contain multiple processors may therefore be insufficient. The specification should connect the distributed architecture to the particular claimed functions.

Some lower-tribunal decisions have gone further. In Google LLC v. EcoFactor, Inc., the PTAB construed express "one or more processors" language as allowing processors collectively to execute recited instructions, reasoning in part that requiring one processor to execute every instruction would give insufficient effect to "or more." IPR2021-00054, Paper 35, 2022 WL 1157606, at 7-10 (P.T.A.B. Apr. 18, 2022). That decision deserves caution. It preceded Salazar, was not designated precedential, and the appeal was dismissed without a merits decision. EcoFactor, Inc. v. Google LLC, No. 2022-1971, 2023 WL 8438149, at 1 (Fed. Cir. Nov. 30, 2023).

More recently, an Eastern District of Texas court distinguished Salazar and Finjan based on component hierarchy. The claims concerned user equipment containing processors, receivers, and transmitters. Although the court agreed that one claimed element had to possess all functionality, it concluded that the relevant singular element was the user equipment—not one individual processor. Because the specification contemplated multiple processors and antennas, the user equipment could distribute the functions among its components. Intellectual Ventures I LLC v. American Airlines, Inc., No. 4:24-cv-00980, Dkt. 177, at 45-47 (E.D. Tex. July 28, 2026). Although the decision is not controlling, it offers a useful model for distinguishing Salazar: the claim must make the larger device, rather than the processor, the functional subject.

An administrative ruling involving a redesigned smartwatch reached a similar result for language expressly reciting "one or more processors." Customs permitted a processor in a user-worn sensor and another in a connected phone or monitor to constitute the claimed processor group. The ruling relied heavily on the absence of language tying all processors to the worn device and on specification disclosures describing separate sensor and monitor components. U.S. Customs & Border Protection, HQ H338254 (Jan. 7, 2025). The ruling has no precedential judicial force, but it confirms the practical importance of component hierarchy and express distributed embodiments.

VII. A Practical Synthesis: Identify the Qualifying Unit

 

The cases support a framework more precise than either "a means one or more" or "one processor must do everything." The proper inquiry is: What claimed unit must possess the complete collection of recited characteristics?

1.       Identify the grammatical subject. Language stating that "the processor is configured to perform A, B, and C" points toward a processor-level requirement. Language stating that "a system comprising one or more processors is configured to perform A, B, and C" more naturally assigns the functions to the system.

2.        

3.       Trace antecedent basis. Repeated references to "the processor," "said processor," or a processor "further configured to" strengthen the conclusion that the same qualifying processor must possess later capabilities. But Finjan shows that a definite article is not always necessary: a claim may assign an entire series of operations to "a computer" through its overall grammar. Finjan, 84 F.4th at 974-75.

4.        

5.       Determine whether the functional subject is composite. "Processor" ordinarily identifies a lower-level component; "server node," "user equipment," "computing system," and "location facility" more readily describe composite structures. Labels are not dispositive, however. The specification may define a processor to include a distributed processing arrangement, or it may use "server" to identify one physical machine.

6.        

7.       Examine the specification for more than generic multiplicity. A statement that a system may include one or more processors addresses how many processors may exist. It does not necessarily say whether those processors may divide A, B, and C. Stronger disclosure explains that different processors may perform different recited operations, gives representative allocations, and states that no single processor need perform every operation.

8.        

9.       Review related claims and prosecution history. An applicant who distinguishes prior art because it divides functions between processors may surrender a later collective-performance construction. See Omega Engineering, Inc. v. Raytek Corp., 334 F.3d 1314, 1323-26 (Fed. Cir. 2003). Conversely, dependent claims expressly requiring one common processor may support an argument that an independent claim permits collective performance, although claim differentiation cannot overcome contrary claim language or intrinsic evidence.

10.    

VIII. Drafting Claims That Resolve the Question

 

The familiar formulation below is unnecessarily uncertain. It expressly permits multiple processors but does not clearly answer whether they may divide the operations:

one or more processors configured to perform A, B, and C.

A drafter should decide which architecture matters and say so.

A. If One Common Processor Is Required

 

one or more processors, at least one processor of the one or more processors being configured to perform each of A, B, and C.

This formulation permits additional processors while expressly requiring a common processor possessing every capability.

B. If Collective Performance Is Intended

 

one or more processors collectively configured to perform A, B, and C, wherein different processors of the one or more processors are permitted to perform different ones of A, B, and C.

The second clause may be omitted when brevity is important, but "collectively configured" is the clearest textual signal that functional division is permitted.

C. If the System Should Be the Functional Subject

 

a computing system comprising memory and one or more processors, the computing system configured to perform A, B, and C through execution of instructions by the one or more processors.

This formulation tracks the reasoning of Unwired Planet and Intellectual Ventures. It requires the system as a whole to possess all capabilities without necessarily assigning every capability to one processor.

D. If Separation Is Part of the Invention

 

a first processor configured to perform A; and a second processor, distinct from the first processor, configured to perform B and, based on a result of A, perform C.

Explicit allocation avoids turning claim construction into a proxy dispute over architecture.

A useful claim set may include several levels. An independent claim may permit collective processing; one dependent claim may require at least one common processor to perform all operations; another may require different processors to perform identified operations; and a separate independent claim may assign the functionality to a system or device. The specification should support each permutation with concrete examples involving, for example, CPU and GPU execution, edge and cloud processors, separate virtual machines, processor cores, or cooperating devices.

IX. Prosecution Consequences

 

During prosecution, antecedent language should be treated as substantive. Replacing "one or more processors collectively configured" with "a processor" followed by "said processor" may create a narrower claim than intended. Likewise, arguing that a reference fails because no single processor performs all functions may become valuable for allowance but costly during enforcement. Prosecution counsel should make that tradeoff deliberately.

When an examiner maps different functions to different prior-art processors, the response should first state the applicant's construction. If the claim requires a common processor, the applicant can argue that the reference fails to disclose the claimed arrangement. If the claim permits collective performance, disputing the rejection may require a different distinction. The Office's interpretation must remain reasonable in light of the claim and specification; it is not simply the broadest linguistically possible reading. In re Smith International, Inc., 871 F.3d 1375, 1382-83 (Fed. Cir. 2017).

Continuation practice provides an opportunity to preserve alternatives. A continuation can expressly claim collective operation even if an allowed parent claim has acquired same-processor language during prosecution. Conversely, a dependent claim requiring one processor to perform all functions may provide a useful validity position and a narrower enforcement fallback. The disclosure must support both architectures before those alternatives can be claimed safely.

X. Litigation and PTAB Consequences

 

For patent owners, the issue affects proof as much as construction. An infringement chart should identify the processor—or qualifying composite system—responsible for every function. Showing that a product contains several processors and that the product as a whole performs all operations may be insufficient under Salazar and Finjan. Relevant discovery may include source code, firmware, task-scheduling records, architecture diagrams, processor-affinity settings, hardware-acceleration paths, and testimony explaining which component executes each operation.

Accused infringers should examine whether the asserted mapping silently changes its level of abstraction. A patentee may identify a device for one limitation, a CPU core for another, and a cloud service for a third, then label the combination "the processor." Whether that collection constitutes the claimed element must be grounded in the intrinsic record, not merely in the fact that all components contribute to the product's operation.

Distributed architectures may also implicate direct-infringement principles separate from processor identity. A system claim spanning client and server components requires analysis of who puts the system into service and obtains its benefit. Centillion Data Systems, LLC v. Qwest Communications International, Inc., 631 F.3d 1279, 1284-85 (Fed. Cir. 2011). A method claim whose steps are divided among actors requires proof that the acts are attributable to one entity under the governing divided-infringement standard. Akamai Technologies, Inc. v. Limelight Networks, Inc., 797 F.3d 1020, 1022-23 (Fed. Cir. 2015) (en banc). Clear processor language therefore does not, by itself, solve a multi-actor infringement problem.

In an IPR, the construction may determine whether a petition establishes anticipation or must rely on obviousness. If the claim requires one processor to possess A, B, and C, a reference distributing those capabilities among unrelated processors ordinarily does not disclose the claimed arrangement. See Net MoneyIN, Inc. v. VeriSign, Inc., 545 F.3d 1359, 1369-71 (Fed. Cir. 2008). An obviousness theory may propose consolidating operations, but the petitioner should explain why a skilled artisan would have done so and why the resulting processor would possess every capability. A conclusory assertion that consolidation would have been routine may not supply the required articulated reasoning. See KSR International Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007); In re Kahn, 441 F.3d 977, 988 (Fed. Cir. 2006).

Conversely, if the claim permits collective performance, a petitioner may map different operations to cooperating processors—but must still show that the processors form the claimed system or composite element and interact in the required manner. Collective construction does not eliminate the all-elements rule; it changes the unit across which those elements may be found.

XI. Conclusion

 

"One or more processors" answers a counting question, but it does not invariably answer an allocation question. Salazar and Finjan generally require one qualifying processor to possess every function assigned to that processor. 01 Communique, Unwired Planet, and later applications show that multiple machines may nevertheless cooperate when the claimed functional subject is a composite system and the intrinsic record supports that architecture.

The drafting lesson is straightforward: if one processor must do everything, say so. If processors may act collectively, say that instead. A few additional words during drafting can avoid a claim-construction dispute years later—after the accused architecture, the most relevant prior art, and the commercial stakes are already fixed.