dissenting.
The majority finds that the ciatos of all four asserted patents are directed to eligible subject matter. To make its determination, the majority undertakes “to examine earlier eases in which a parallel descriptive nature can be seen—what prior cases were about and which way they were decided.” Majority Op. at 1294. In application, the majority’s approach involves the mechanical comparison of the asserted claims in this case to the claims at issue in some, but not all, of the cases where we have addressed patent eligibility after the Supreme Court’s decision in Alice Corp. v. CLS Bank Int'l — U.S.-, 134 S.Ct. 2347, 2355, 189 L.Ed.2d 296 (2014).
The majority avoids determining whether the asserted claims are directed to an abstract idea, or even identifying what the underlying abstract idea is. I believe that approach to section 101 is contrary to the Supreme Court’s direction in Alice, 134 S.Ct. at 2355 (“First, we determine whether the claims at issue are directed to one of those patent-ineligible concepts.”). Declining to engage in the step 1 inquiry also ignores and undermines this court’s holdings in Enfish, LLC v. Microsoft Corp., 822 F.3d 1327 (Fed. Cir. 2016), McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299 (Fed. Cir.2016), Affinity Labs of Texas, LLC v. DirecTV, LLC, 838 F.3d 1253 (Fed. Cir.2016), and Affinity Labs of Texas, LLC v. Amazon.com Inc., 838 F.3d 1266 (Fed. Cir.2016).
The majority also relies on the specification to import innovative limitations into the claims at issue. For each of the four patents at issue, the majority’s eligibility determination rests on the use of a “distribution architecture.” As explained below, however, this limitation is insufficient to satisfy Alice step two. Indeed, that limitation- does not exist in all of the claims at issue. This contravenes the fundamental principal that the section 101 inquiry is about whether the claims are directed to a patent-eligible invention, not whether the specification is so directed. See Synopsys, Inc. v. Mentor Graphics Corp., 839 F.3d 1138 (Fed.Cir. 2016) (“The § 101 inquiry must focus on the language of the Asserted Claims themselves.... complex details from the specification cannot save a claim directed to an abstract idea that recites generic computer parts.”) (citing Accenture Global Servs., GmbH v. Guidewire Software, Inc., 728 F.3d 1336, 1345 (Fed. Cir. 2013)).
Because I do not agree that the ’065 and ’797 patents are § 101 eligible, nor with the basis expressed by the majority for finding all four patents subject matter eligible under § 101,1 dissent.
Background
The patents-in-suit disclose a system for monitoring activity on computer networks and for creating accounting records reflecting the activity.1 The system gathers *1308raw activity data from various devices on the network (e.g., “routers, switches, firewalls, authentication servers, LDAP, Web hosts, DNS, and other devices”), and it uses that raw activity data to derive the desired accounting records. ’984 patent at col. 2 1, 65-col. 3 1. 11. In certain embodiments, the system stores the records in a central database, which the network provider can use, for example, for purposes such as billing, operational support, fraud detection, network monitoring, traffic engineering, and the like. Id, at col. 3 11. 20-27, col. 8 1.40-col. 9 1. 41; ’797 patent at col. 3-16-20.
Rather than storing all the raw data in a central database, as in prior art systems, the disclosed system uses a distributed architecture to process the raw data in parallel,' closer to the points of collection. The system associates a distinct Information Source Module (“ISM”) with each network device that records relevant activity data. Id. at col. 5 11. 3-17. The network devices include any devices in the network. Id. at col. 4 11. 49-50. The ISMs are software components that “Represent modular, abstract interfaces that are designed to be platform neutral.” Id: at col. 5 11. 6-8.
Each ISM collects data from the associated network device and passes the data to a respective “gatherer” component. Id. at col. 5 11. 10-11. The gatherer component “can be any hardware and/or software,” for gathering data from the ISMs and cooperating with other components to process the data to form the desired records. Id. at col. 6 11. 25-31. To reduce the additional network traffic created by the monitoring, each gatherer is preferably placed logically or physically near the network devices from which it collects information. Id. at col. 6 11. 32-35.
To derive the values necessaiy to create the desired accounting records, a gatherer may-manipulate the raw data it receives from the ISM by filtering, aggregating, and/or “enhancing” the data. Id. at col. 611. 25-col. 7 11. 50, col. 10 11. 13-col. 11 11. 35. “Enhancing” includes “applying zero or more functions” to a value before storing the resulting value in a field of the record. Id. at col. 1011. 63-65. For instance, simply placing a raw value in the record is referred to as “one-step field enhancement.” Id. at col. 10 11. 66-67. In contrast, using the raw value to query another ISM for the value to place in the record is an example of “two-step field enhancement.” Id. at col. 11 11. 3-7. A gatherer may “enhance” the data through any number of steps.
A Central Event Manager (“CEM”) provides centralized control and management of the system. Id. at col. 7 11. 51-cob 8 11. 39. The CEM provides a graphical user interface for system administrators to query the central database or to configure the system. Id. at col. 9 11. 42-60. For example, administrators can use the user interface to define enhancement procedures for implementation by the gatherers and ISMs. Id. at col. 1111. 36-col. 13 11. 30.
The patents explain that, because the disclosed system distributes the work of collecting and processing the raw activity data among multiple components, it is able to process more information more quickly than do previous designs, in which “all the [raw] network information flows to one location.” Id. at col. 4 11. 9-13. In contrast to these previous designs, the distributed architecture reduces the storage and computational resource requirements of the central repository, which need no longer “keep up with the massive record flows from the network devices” or maintain “huge databases.” Id. at col. 4 11. 7-13. Moreover, the distributed architecture reduces network traffic overhead “by reduc*1309ing the volume of data sent on the network to the OEM.” Id. at col. 6 11. 49-50. The end result is a system that can monitor, process, and create database records reflecting network activity at large scale.
Network operators can use the ultimate records to get an accurate and dependable picture of network usage. The operators can use this information for any number of purposes, such as setting the right price for network services, implementing usage-based charging models, deploying new services based on usage trends, planning network resource provisioning, and usage auditing. Id. at col. 21. 65-col. 31. 27.
Legal FhamewoRK
The Supreme Court has outlined a two-step framework for analyzing whether a claim is eligible. See Alice, 134 S.Ct. at 2355. First, we determine whether the claim at issue is directed to a judicial exception, such as an abstract idea. Id. If so, we next consider all the claim elements in combination to determine whether they recite an inventive concept sufficient to ensure that the patent in practice amounts to significantly more than a patent upon the ineligible concept itself. Id. As this Court recently explained, this two-step formulation contemplates that step one is meaningful, and that a substantial class of claims are not directed to patent ineligible concepts. Enfish, 822 F.3d at 1335; see also McRO, 837 F.3d at 1312-16.
The Alice framework leaves open at least three questions: (1) what makes an idea “abstract”; (2) what it means for a claim to be “directed to” an abstract idea; and (3) what limitations provide an “inventive concept?” To answer these questions we first look to the foundational principles of the abstract idea exception.
For well over a century, the Supreme Court has repeatedly and consistently used the abstract idea exception to prevent patenting a result where “it matters not by what process or machinery the result is accomplished.” O’Reilly v. Morse, 56 U.S. 62, 113, 15 How. 62, 14 L.Ed. 601 (1854). The Court has explained that a patent may issue “for the means or method of producing a certain result, or effect, and not for the result or effect produced.” Diamond v. Diehr, 450 U.S. 175, 182 n.7, 101 S.Ct. 1048, 67 L.Ed.2d 155 (1981). “A patent is not good for an effect, or the result of a certain process” because such patents “would prohibit all other persons from making the same thing by any means whatsoever.” Le Roy v. Tatham, 55 U.S. 156, 175, 14 How. 156, 14 L.Ed. 367 (1853).
Hence, the abstract idea exception must be applied in a way that reserves patent protection for means rather than for ends and thus maintains the incentive of “some future inventor, in the onward march of science” to discover new ways of achieving the same result more cheaply and efficiently than has the patentee. Morse, 56 U.S. at 113; see also Dolbear v. Am. Bell Tel. Co., 126 U.S. 1, 533, 8 S.Ct. 778, 31 L.Ed. 863 (1888) (“Other inventors may compete with him for the ways of giving effect to the discovery.”). This basis of the abstract idea exception runs clear through the Supreme Court’s jurisprudence from the nineteenth century to the present day.
Based on the Supreme Court’s use of the abstract idea exception, it is apparent that a desired goal (i.e., a “result or effect”), absent structural or procedural means for achieving that goal, is an abstract idea. Not every abstract idea is naturally phrased as a goal, and indeed, the Supreme Court has treated somewhat disparate ideas, such a “mathematical formula,” Gottschalk v. Benson, 409 U.S. 63, 71, 93 S.Ct. 253, 34 L.Ed.2d 273 (1972), and a “fundamental economic practice,” Bilski v. Kappos, 561 U.S. 593, 611, 130 S.Ct. 3218, 177 L.Ed.2d 792 (2010), under the abstract idea rubric. Nevertheless, long-standing *1310Supreme Court precedent clearly establishes that a desired goal without means for achieving that goal is an abstract idea. With this in mind, I turn back to the first step of the eligibility inquiry.
Step one of the eligibility inquiry asks whether the claim is “directed to” a judicial exception, such as an abstract idea. The answer is not automatically “yes” simply because a claim involves an abstract idea, and it is not automatically “no” simply because a claim recites limitations beyond the abstract idea. See McRO, 837 F.3d at 1312-13. The Supreme Court has recognized that “[a]t some level, all inventions embody, use, reflect, rest upon, or apply laws of nature, natural phenomena, or abstract ideas.” Alice, 134 S.Ct. at 2354 (internal quotation .marks and ellipses omitted). Unless step one is a nullity, the phrase “directed to” must therefore mean more than merely “embody, use, reflect, rest upon, or apply.” At the same time, the phrase “directed to” must apply even where the claim does not wholly pre-empt the abstract idea. For example, it is well settled that the prohibition against patenting abstract ideas cannot be circumvented by limiting the use of the idea to a particular technological environment or adding insignificant extra-solution activity. Bilski, 561 U.S. at 610-11, 130 S.Ct. 3218. Consequently, the step one inquiry cannot be settled in the affirmative by the observation of an underlying abstract idea nor in the negative by recitation of just any additional limitations.
Rather, the step one inquiry is a legal analysis that must focus on determining “what type of discovery is sought to be patented.” Parker v. Flook, 437 U.S. 584, 593, 98 S.Ct. 2522, 57 L.Ed.2d 451 (1978). For example, a claim is “directed to” an abstract goal if the claim fails to describe how—whether by particular process or structure—the goal is accomplished.2 Even if the claim recites additional limitations, the claim is nevertheless directed to the underlying goal if those limitations fail to restrict how the goal is accomplished. Conversely, where the claim recites specific structure or function for accomplishing the desired goal in a particular way, the claim is more likely directed to a means than to the underlying abstract goal.3 See McRO, 837 F.3d at 1313-15. In those cases, concerns of patent eligibility are resolved at step one, and there is no need to proceed to step two. See Enfish, 822 F.3d at 1339.
Post-Alice, wé have only twice held that a patent was eligible under § 101 based on a determination during step one that the claims were not directed to an abstract idea. In Enfish, we held that the claims at issue were directed to “a specific implementation of a solution to a problem in the software arts” designed to “improve the way a computer stores and retrieves data in memory,” as opposed to an abstract idea implemented with general-purpose computer components. Id. In McRO, we held that the claims at issue were eligible under Alice step one because they were directed to “a specific asserted improvement in *1311computer animation, i.e., the automatic use of rules of a particular type.” McRO, 837 F.3d at 1314. The scarcity of cases resolved under step one should not be interpreted as an indication that step one creates a particularly high bar. .
The inquiry moves to the careful limitation-by-limitation analysis of step two, where there is a credible concern that the additional limitations fail to direct the claim to an eligible invention—e.g., a particular means for accomplishing an underlying goal—or to otherwise obviate concerns of pre-emption. The purpose of the step-two analysis is to ensure that the claim recites an “inventive concept,” which the Supreme Court has defined as “an element or combination of elements that is sufficient to ensure that the patent in practice amounts to significantly more than a patent upon the [ineligible concept] itself.” Alice, 134 S.Ct. at 2355.
To be clear, the concept of inventiveness is distinct from that of novelty. Novelty is the question of whether the claimed invention is new. Inventiveness is the question of whether the claimed matter is invention at all, new or otherwise. The inventiveness inquiry of § Í01 should therefore not be confused with the separate novelty inquiry of § 102 or the obviousness inquiry of § 103. Accordingly, the Supreme Court has cautioned that “[t]he obligation to determine what type of discovery is sought to be patented must precede the determination of whether that discovery is, in fact, new or obvious.” Flook, 437 U.S. at 593, 98 S.Ct, 2522.
Claims that fail to recite how a desired goal is accomplished do not recite an inventive concept. For example, limitations on the context—as opposed to the manner—of accomplishing a desired result is typically not inventive, even if that context is novel. The Pythagorean Theorem cannot be made eligible by confining its use to existing surveying techniques, Flook, 437 U.S. at 590, 98 S.Ct. 2522, nor can the business practice of hedging risk be patented by confining its use to the commodities and energy markets, Bilski, 561 U.S. at 612, 130 S.Ct. 3218, nor the goal of “gathering and combining data” by confining its use to particular types of photographic information, Digitech Image Technologies, LLC v. Electronics for Imaging, Inc., 758 F.3d 1344, 1351. (Fed. Cir. 2014). Even though such field-of-use limitations prevent a claim from wholly pre-empting an abstract idea, they are not inventive because they describe only the context rather than the manner of achieving a result. For similar reasons, limitations that recite only insignificant extra-solution activity also cannot supply an inventive concept because extra-solution activity, by definition, describes activity unrelated to how the solution is achieved. See Flook, 437 U.S. at 590, 98 S.Ct. 2522; see also Mayo, 132 S.Ct. at 1300. It is therefore well established that “limiting an abstract idea to one field of use or adding token postsolution components [does] not make the concept patentable.” Bilski, 561 U.S. at 612, 130 S.Ct. 3218.
■ Illusory limitations, which describe only procedure or structure common to every means. of accomplishing a given result, also eannot provide an inventive concept. Put another way, limitations that simply “comprise the abstract concept” are not inventive. See Ultramercial Inc. v. Hulu, LLC, 772 F.3d 709, 715 (Fed. Cir. 2014). For example, a claim cannot become eligible by reciting that physical automation is accomplished by a “machine” or that logical automation is accomplished by a “computer,” see OIP Technologies, Inc. v. Amazon.com, Inc., 788 F.3d 1359, 1363 (Fed. Cir. 2015), because physical automation requires a machine and logical automation requires a computer. Because such elements cannot restrict a claim to a particular way of automating, recitation of a *1312machine or computer “to lend speed or efficiency to the performance of an otherwise abstract concept does not meaningfully limit claim scope for purposes of patent eligibility.” CLS Bank Int’l v. Alice Corp., 717 F.3d 1269, 1286 (Fed. Cir. 2013).
Post-Aiice, we have only once found that a claim’s additional limitations provide an inventive concept. See DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245 (Fed. Cir. 2014).4 In DDR, we held that “a specific way to automate the creation of a composite web page” was patent eligible even though the underlying abstract idea of “increasing sales by making two web pages look the same” was not. DDR, 773 F.3d at 1259 (emphasis added). In doing so, we distinguished our precedent on the basis that the DDR claims “do not broadly and generically claim ‘use of the Internet’ ” to achieve the desired result, but instead “specify how interactions with the Internet are manipulated to yield a desired result.” Id. at 1258. We cautioned that “not all claims purporting to address [technological] challenges are eligible for patent.” Id. Instead, only claims specifying how to overcome those technological challenges are eligible.
In summary, the eligibility inquiry requires us to first determine whether the claim is “directed to” an abstract idea (such as a result) rather than to an application (such as a particular means of accomplishing that result). If the claim is clearly directed to an application, the inquiry may end. If doubt remains, the inquiry moves to step two, where we carefully consider all the implementation details to determine whether they define an inventive concept. The case law has identified several types of limitations that frequently fail to provide an inventive concept, including illusory limitations (e.g., generic computer implementation) and contextual limitations (e.g., field of use, extra-solution activity). The step-two inquiry is a flexible and fact-specific one focused on whether the claims unduly foreclose future innovation.
DISCUSSION
If I were to examine only the written description of the asserted patents, I would conclude that the network monitoring system disclosed therein is eligible for patenting. The specifications disclose a distributed system architecture comprising special-purpose components configured to cooperate with one another according to defined protocols in a user-configurable manner for the purpose of deriving useful accounting records in a more scalable and efficient manner than previously possible. The disclosed system improves upon prior art systems by creating a specific “distributed filtering and aggregation system ... [that] eliminates capacity bottlenecks” through distributed processing. ’984 patent at col. 6 11. 45-50. The disclosed system is patent eligible.
But the inquiry is not whether the specifications disclose a patent-eligible system, but whether the claims are directed to a patent ineligible concept. See Synopsys, 839 F.3d at 1138 (“The § 101 inquiry must focus on the language of the Asserted Claims themselves.... complex details from the specification cannot save a claim directed to an abstract idea that recites generic computer parts.”) (citing Accenture, 728 F.3d at 1345); Alice, 134 S.Ct. at *13132355 (“First, we determine whether the claims at issue are directed to one of those patent-ineligible concepts.”); Diamond v. Diehr, 450 U.S. 175, 189, 101 S.Ct. 1048, 67 L.Ed.2d 155 (1981) (“In. determining the eligibility of respondents’ claimed process ..., their claims must be considered as a whole.”); McRO, 837 F.3d 1299 (“If the claims are “directed to” an abstract idea, then the inquiry proceeds to the second step.... In step two we consider whether the claims contain an ‘inventive concept’_ To do so we look to both the claim as a whole and the individual claim elements_”); see also McCarty v. Lehigh Valley R.R. Co., 160 U.S. 110, 116, 16 S.Ct. 240, 40 L.Ed. 358 (1895) (“if we once begin to include elements not mentioned in the claim, in order to limit such claim ..., we should never know where to stop”).
Answering this inquiry requires a court to step through each claim to determine whether it is directed to an abstract idea, and if so, to determine whether the claim recites structural or procedural limitations sufficient to ensure that the claim “amounts to significantly more than a patent upon the ineligible concept itself.” Alice, 134 S.Ct. at 2355.
A. ’065 Patent
Amdocs asserted claims 1, 4,, 7, 13, and 17 of the ’065 patent. Claim 1 is representative:
1. A computer program product embodied on a computer readable storage medium for processing network accounting information comprising: computer code for receiving from a first source a first network accounting record;
computer code for correlating the first network accounting record with accounting information available from' a second source; and
computer code for using the accounting information with which the first network accounting record is correlated to enhance the first network accounting record.
The underlying goal of claim 1 is to combine particular information from two different sources. But the step one question is not whether claim 1 involves that abstract idea, but whether claim 1 is directed to it.
Claim 1 recites a software product embodied on a storage medium, but it provides no structural limitations of either the physical medium or the digital software. All software products are stored on a physical storage medium, and claim 1 recites no limitations concerning that physical structure. Likewise, claim 1 discusses only very broad, high-level functionality rather than details about how exactly that functionality is implemented, providing no information about the structure of the software. That the recited information concerns network accounting also provides no particular structure. Claim 1 is therefore not directed to any specific structure, whether physical or digital.
Rather than reciting structure, claim 1 defines the program product using only functional limitations. Looking at those limitations, I find no specific process for accomplishing the abstract goal of combining data from two sources. The recited software performs three- steps: (1) receiving information from a first source, (2) correlating the information with information available from a second source, and (3) using that available information to “enhance” the first information. Under the district court’s construction, to “enhance” includes simply retrieving and recording information, in a field. The three steps therefore only “comprise the abstract concept” of combining data from different sources. Ultramercial, 772 F.3d at 715. Claim 1 is therefore directed to an ab*1314stract idea. Accordingly, the inquiry continues under step two.
Turning to step two, I see no limitations confining the claim to a particular means of combining information from different sources. Limiting the abstract idea to the context in which the information relates to network accounting records is a field-of-use limitation that does not supply an inventive concept. See Flook, 437 U.S. at 590, 98 S.Ct. 2522. The use of “computer code” to automate logic is likewise not an inventive concept because “recitation of generic computer limitations does not make an otherwise ineligible claim patent-eligible.” DDR Holdings, 773 F.3d at 1256. The abstract idea of “gathering and combining data” with a computer is ineligible when only limited by the type of data. See Digitech, 758 F.3d at 1351. The concept of gathering and combining data is all that claim 1 recites.
Amdocs argues that the “enhance” step provides an inventive concept because the district court’s construction of the term “enhance” requires applying zero or more functions “in a distributed fashion.” Br. of Appellant at 59. Amdocs thus renews its argument from the trial proceedings that “the asserted claims are patentable, in part, due to the manner in which the claims facilitate the generation of network accounting records—i.e., ⅛ a distributed fashion.’” J.A. 1567 (emphasis original).
But the “distributed fashion” limitation cannot provide an inventive concept because it has no meaning in the context of claim 1. Claim 1 only requires adding a single piece of information to an accounting record, and it is unclear what doing this “in a distributed fashion” could mean. Moreover, claim 1 recites no components or structure over which the work might be “distributed.”
I agree with the district court that claim 1 is ineligible because it fails to recite any structure or process limiting the claim to a particular means of combining accounting data from different sources. For that reason, I would affirm, the district court’s determination that claims 1, 4, 7, 13, and 17 of the ’065 patent are ineligible.
B. ’510 Patent
Amdocs asserted claims 16,17, and 19 of the ’510 patent. Claim 16 is representative:
16. A computer program product stored in a computer readable medium for reporting on a collection of network usage information from a plurality of network devices, comprising: computer code for collecting network communications usage information in real-time from a plurality of network devices at a plurality of layers; computer code for filtering and aggregating the network communications usage information;
computer code for completing a plurality of data records from the filtered and aggregated network communications usage information, the plurality of data records corresponding to network usage by a plurality of users;
computer code for storing the plurality of data records in a database;
computer code for submitting queries to the database utilizing predetermined reports for retrieving information on the collection of the network usage information from the network devices; and computer code for outputting a report based on the queries;
wherein resource consumption queries are submitted to the database utilizing the reports for retrieving information on resource consumption in a network; and wherein a resource consumption report is outputted based on the resource consumption queries.
In step one, the district court identified the abstract idea underlying claim 16 as *1315“using a database to compile and report on network usage information.” J.A. 22. I agree that this is the goal of the claimed invention. Indeed, the claim’s preamble recites that the invention is for “reporting on a collection of network usage information.” But again, the step 1 question is not whether claim 16 has a goal, but whether claim 16 is directed to that goal rather than to a means of achieving that goal.
As discussed above, one way for a claim to be directed to a means rather than to an abstract end is to recite process limitations defining a specific way of arriving at that end. See Diehr, 450 U.S. at 182-83, 101 S.Ct. 1048 (holding that “a process may be patentable, irrespective of the particular form of the instrumentalities used”). Such limitations may obviate concerns of preemption because they leave room for future inventors to develop new paths to the same end without infringing the patent. See Morse, 56 U.S. at 113. Because § 101 is a “coarse eligibility filter,” Research Corp. Technologies v. Microsoft Corp., 627 F.3d 859, 869 (Fed. Cir. 2010), the recited way of accomplishing the goal need not be extensively detailed or even complete. Rather, it must meaningfully limit the claim to a manner of achieving the desired result without unduly foreclosing future innovation,
Amdocs argues that claim 16 is eligible because it recites procedural limitations, including “filtering and aggregating” “in real time ... at á plurality of layers,” and using the filtered and aggregated information to “complete” data records “in a distributed fashion.” Br. of Appellant at 52-53. It therefore argues that the claims “prescribe a particular inventive manner by which network accounting information is collected, processed, and transformed into meaningful records.” Id. at 53-54 (emphasis original). I agree.
The disclosed invention improves upon the manner in which prior art systems collected and processed network usage information. Unlike those prior art systems, which used centralized processing, the invention improves performance by distributing the processing work among cooperating components. But the invention cannot be merely the idea of distributing the processing—it must describe how. The idea of improving performance through distributed processing is just an abstract goal because the benefits of distributed processing can be attained only through a specific distributed architecture and protocol. The issue here is whether the claims recite enough of that distributed architecture or protocol.
Claim 16 captures enough of the distributed protocol disclosed in the specification to pass through the coarse eligibility filter of § 101. First, claim 16 recites that the network information is collected from a specific source—“a plurality of network devices at a plurality of layers.” Next, claim 16 recites that the distributed system operates on the collected information by applying two specific types of functions—filtering and aggregating. Then, claim 16 recites that the filtered and aggregated information is further processed by enhancing it “in a distributed fashion.” See Amdocs, 761 F.3d at 1338 (upholding the district court’s construction of “completing” as requiring distributed enhancement). Unlike claim 1 of the ’065 patent, claim 16 of the ’510 patent recites “a plurality of network devices” over which the enhancement work may be distributed. Taken together, the limitations of claim 16 capture at least some of the process by which the disclosed system collects, processes, and transforms network accounting information, in a distributed fashion, into usable accounting records.
The district court held that claim 16 “does not add any specific implementation beyond the abstract idea that information *1316is collected and stored, and reports are generated,” because “[collecting, filtering, aggregating, and completing network information amounts to ‘electronic record-keeping.’” J.A. 22. I agree that claim 16 embodies a method of electronic record keeping, but I disagree that the claim is directed to that abstract goal rather than to a particular process for achieving it. Simply because computers are frequently called upon to perform operations such as “Collecting, filtering, aggregating, and completing,” this does not mean that any claim reciting these steps in any order and for any purpose is necessarily directed to that abstract concept. We must consider the claim as a whole and ask “what type of discovery is sought to be patented?” Flook, 437 U.S. at 593, 98 S.Ct. 2522 (emphasis added). Here, the type of invention is a distributed software system that collects and processes network activity in a particularly scalable manner.
Openet argues that the “distributed fashion limitation should be given no weight because a “distributed architecture” is “a generic type of architecture.” Br. of Appellee at 43. However, the claimed invention is not that the work is distributed, but how that distributed architecture is applied. Even if distributed processing generally was a known approach for improving system performance, claim 16 recites a way of applying distributed processing to the problem of activity monitoring, by collecting activity data “in real time from a plurality of network devices at a plurality of layers,” then filtering and aggregating the data, and then using the filtered and aggregated data to assemble accounting records using a distributed “enhancement” protocol. To whatever extent this claimed approach was old, obvious, too broadly claimed, or unsupported, these considerations are apart from the eligibility inquiry and best reserved for other parts of the patentability analysis.
Like the claims at issue in Enfish and McRO, independent claim 16 and its dependent claims 17 and 19 of the ’510 patent are “directed to” a particular process that improves upon the manner in which systems collect and process network usage information, and the claimed process is limited in a specific way. As such, the claims are patent-eligible under step one of the Alice test, and there is no need to consider step two. Id. For that reason, I would reverse the district court’s holding to the contrary.
C. ’984 Patent
Amdocs alleged infringement of claims 1, 2, 7, 8, and 13 of the ’984 patent. Claims 1 and 13 are independent, and claim 1 is representative:
1. A method for reporting on the collection of network usage information from a plurality of network devices, comprising:
(a) collecting network communications usage information in real-time from a plurality of network devices at a plurality of layers utilizing multiple gatherers each including a plurality of information source modules each interfacing with one of the network devices and capable of communicating using a protocol specific to the network device coupled thereto, the network devices selected from the group consisting of routers, switches, firewalls, authentication servers, web hosts, proxy servers, netflow servers, databases, mail servers, RADIUS servers, and domain name servers, the gatherers being positioned on a segment of the network on which the network devices coupled thereto are positioned for minimizing an impact of the gatherers on the network;
(b) filtering and aggregating the network communications usage information;
*1317(c) completing a plurality of data records from the filtered and aggregated network communications usage information, the plurality of data records corresponding to network usage by a plurality of users;
(d) storing the plurality of data records in a database;
(e) allowing the selection of one of a plurality of reports for reporting purposes;
(f) submitting queries to the database utilizing the selected reports for retrieving information on the collection of the network usage information from the network devices; and
(g) outputting a report based on the queries.
Claim 1 of the ’984 patent is analogous to claim 16 of the ’510 patent, except that it adds limitation (a), which recites details of the distributed architecture.
In step one, the district court identified the abstract idea underlying claim 1 as “reporting on the collection of network usage information from a plurality of network devices.” J.A. 27. In step two, the district court found no inventive concept because the additional limitations recite only that “the genetic computer collects information from conventional devices to create records,” using “gatherers, which are software,” and then “filtering, completing, storing, allowing, submitting, and outputting,” all of which are actions that are “conventional for both generic computers and generic databases.” J.A. 27. It applied the same reasoning to claim 13. Id.
I see no error in the district court’s articulation of the underlying abstract idea, which duplicates the preamble of claim 1. But again, after identifying the underlying idea, a court must still ask whether the claim is directed to that idea or to a specific means.
Because claim 1 of the ’984 patent'includes the same process limitations’ as the ’510 claims, it is eligible for at least the same reasons. It was error for the district court to dismiss these process limitations solely on the basis that “filtering, completing, storing, allowing, submitting, and outputting” are “conventional” types of activities for computers. Id. If this analysis were sufficient, no software invention could be eligible because every software invention comprises at most the “conventional” activities of receiving, storing, manipulating, and outputting information. These activities are all that computers can do. But “a new combination of steps in a process may be patentable even though all the constituents of the combination were well known and in common use before the combination was made.” Diamond v. Diehr, 450 U.S. at 188, 101 S.Ct. 1048. Whether a process is performed by software, hardware, machine, or man, the eligibility requirements are identical. The claimed invention must be limited to a specific means (i.e., process or structure) for achieving its underlying purpose. In other words, the claim must be limited “by what process or machinery the result is accomplished.” Morse, 56 U.S. at 113.
It is worth noting that the “process or machinery” by which a result is accomplished need not be tangible to be patent eligible. Though the Supreme Court’s early Information Age jurisprudence incorporated the Industrial Age requirement that eligible inventions must use or manipulate tangible materials,5 the Court’s subse*1318quent case law has questioned that requirement. See Bilski, 561 U.S. at 605,130 S.Ct. 3218 (“But there are reasons to doubt whether the [machine-or-transformation] test should be the sole criterion for determining the patentability of inventions in the Information Age”). A software program is a digital machine. Like a physical machine, a digital machine is made of specific parts that interact with one another to achieve a specific result in a specific way. A claim to either type of machine is eligible only if the claim recites structural limitations detailing those specific parts, process limitations detailing that specific way, or a combination of the two. Such structure or process may be found in the recited components individually as well as in their arrangement and interaction with one another as a system. But the district court considered neither possibility.
Claim 1 recites a distributed architecture, including three types of components (i.e., network devices, gatherers, and ISMs) with given interrelations. The gatherers are coupled to the network devices and positioned on the same segment of the network as those devices. Moreover, each gatherer includes multiple ISMs in a one-to-many relationship, and the ISMs interface with respective network devices using a protocol specific to that device. Because such software structure and process can confer eligibility, the district court erred by dismissing the recited components. on the sole basis that they “are software” without considering whether these architectural aspects are inventive structure or process. J.A. 27.
For the forgoing reasons, I would find that claim 1 of the ’984 patent and its dependent claims 2, 7, and 8 are patent eligible. Independent claim 13 is also eligible because, as the district court acknowledged, it “is directed to essentially the same invention.” J.A. 27.1 would therefore reverse the district court’s holding that claims 1, 2, 7, 8, and 13 of the ’984 patent are not patent eligible.
D. ’797 Patent
Amdocs alleged infringement of claims 1, 2, 7, 8, and 19 of the ’797 patent. Claims 1, 7, and 19 are independent, and claim 1 is representative:
1. A method for generating a single record reflecting multiple services for accounting purposes, comprising:
(a) identifying a plurality of services carried out over a network;
(b) collecting data describing the plurality of services; and
(c) generating a single record including the collected data, wherein the single record represents each of the plurality of services;
wherein the services include at least two services selected from a group consisting of a hypertext transfer protocol (HTTP) session, an electronic mail session, a multimedia streaming session, a voice over Internet Protocol (IP) session, a data communication session, an instant messaging session, a peer-to-peer network application session, a file transfer protocol (FTP) session, and a telnet session;
wherein the data is collected utilizing an enhancement procedure defined utilizing a graphic user interface by:
listing a plurality of available functions to be applied in real-time prior to end-user reporting,
allowing a user to choose at least one of a plurality of fields, and
allowing the user to choose at least one of the listed functions to be applied to the chosen field in real-time prior to the end-user reporting.
In step one, the district court identified the underlying abstract idea as “gener-*1319at[ing] a single record reflecting multiple services.” J.A. 24. In step two, the district court found that the claim adds “only conventional computer functions operating in a conventional manner,” and therefore “amounts to electronic record keeping,” which is “one of the most basic functions of a computer.” Id. The court found nothing inventive about the “enhancement procedure” or about defining that procedure using á graphical user interface (“GUI”), which it reasoned is a conventional way to interact with a computer. Id.
I see no error with the district court’s articulation of the underlying abstract idea, which tracks the preamble of claim 1. I also agree that claim 1 is directed to an abstract idea rather than to a particular process or structure. Steps (a)-(c) utilize nebulous terms to describe a process of “identifying” “services,” collecting data “describing” those services, and generating a “record” that “represents” the services. These three steps merely corriprise the abstract concept of collecting information about network services, but the goal of “gathering and combining data” is not patent-eligible. See Digitech, 758 F.3d at 1351.
The next question is whether the two wherein clauses redirect the claim to a particular method or structure. They do not. The first wherein clause limits the subject of the collected data, but it _ does not define any particular process or structure. The second wherein clause recites that the data is collected utilizing a distributed enhancement procedure and that the procedure is customized by a user’s selection of the fields and functions to apply. Like the ’065 claims, claim 1 of the ’797 recites no distributed architecture over which the enhancement might be performed. Moreover, the user’s pre-solution configuration does not clearly redirect the claim to a particular method of gathering data—at least there is a credible concern that it does not.
Moving to step two, the central question is whether the second wherein clause contains some inventive concept such that claim 1 “amounts to significantly more than a patent upon the” idea of collecting information about network services. Am-docs argues that the “enhancement procedure” provides this inventive concept because it requires combining data from multiple network devices. Br. of Appellant at 63-65. But this argument is not persuasive because the abstract idea of “gathering and combining data” is not patent-eligible, see. Digitech, 758 F.3d at 1351, regardless of the number of sources from which the data is gathered. Lastly, Am-docs argues that the claims “do not recite the general use of a GUI, but also specifically limit how the GUI is used.” Br. of Appellant at 65 (emphasis original). I do ■not agree. The limitations of the second wherein clause do not limit how the GUI is used, but for what purpose. That purpose is to allow the user to choose the enhancement functions. Nothing in these limitations evinces an inventive way of permitting the user to select the functions or otherwise customize the enhancement. At best, the user’s pre-solution customization amounts to insignificant pre-solution activity. See Bilshi, 561 U.S. at 612, 130 S.Ct. 3218. I see no inventive concept in claim 1.
For the foregoing reasons, I would hold that claim 1 of the ’797 patent is ineligible. Claims 2, 7, 8, or 19 are likewise ineligible because Amdocs has not argued that any of these claims add anything more to claim 1. Accordingly, I would affirm the district court’s determination that claims 1, 2, 7, 8, and 19 of the ’797 patent are ineligible.
For these reasons, I dissent.
. All the patents are descendant from U.S. Pat. No. 6,418,467 and they share its common specification, with some variation not relevant here. The '797 patent is a continuation-in-part that contains additional disclosure *1308concerning the content of the accounting records. See '797 patent at col. 2 1. 33-col. 6 1. 9.
. The same concern applies regardless of how narrow the goal. See Mayo, 132 S.Ct. at 1302 (holding that even "narrow laws that may have limited applications” "nonetheless implicate this concern” of pre-emption); buySAFB, Inc. v. Google, Inc., 765 F.3d 1350, 1353 (Fed. Cir. 2014) ("exclusion applies if a claim involves a natural law or phenomenon or abstract idea, even if the particular natural law or phenomenon or abstract idea at issue is narrow”).
. The terms "means” and "function,” as used here, are not to be strictly understood in the context of "means plus function” claiming under 35 U.S.C. § 112(f). When considering whether a claim is directed to an abstract idea or is limited to a means of achieving an underlying abstract goal, we necessarily take into consideration whether the claim includes means-plus-function limitations.
. In one recent case, we found that a patentee made allegations of an inventive step that, when unrebutted, were sufficient to survive a motion to dismiss for ineligibility under Fed. R, Civ. P. 12(b)(6). Bascom Global Internet Services, Inc. v. AT & T Mobility LLC, 827 F.3d 1341, 1352 (Fed. Cir. 2016). Of course, the alleged infringer may yet prevail in invalidating the patent under section 101.
. See, e.g., Diehr, 450 U.S. at 183, 101 S.Ct. 1048 ("A process is a mode of treatment of certain materials") (quoting Cochrane v. Deener, 94 U.S. 780, 787-788, 24 L.Ed. 139 (1877)); Gottschalk v. Benson, 409 U.S. 63, 70, 93 S.Ct. 253, 34 L.Ed.2d 273 (1972) ("Transformation and reduction of an article ‘to a different state or thing’ is the clue to the patentability of a process claim that does not include particular machines.”).