Navigation Links: Notes Definitions > Internet > Internet Definition > History > Public Internet > Internet2 See also Internet :: Feedback :: :: Disclaimer :: :: Sitemap ::

Introduction

We start with a problem.  The problem is that so frequently while discussing legal or regulatory issues related to The Internet, we have no idea what it is that we are talking about.

Countless courts, legislative bodies, and packs of pundits have sought to produce definitions of The Internet.  Their attempts have at times been poor and misconstrued.  Many definitions do not agree.  Some talk about applications.  Some talk about packets.  And others are simply too vague to be of value.  And yet, not comprehending the object of the policy, the lawyers go merrily forward constructing on faulty foundation new, imaginative, and perhaps misdirected policy and law.

Is it possible to formulate a good definition of The Internet?  Is The Internet the technical specs that currently make up The Internet?  Is The Internet based on the experience of The Internet by the user?  Can a definition of The Internet reflect the robust and flexible nature, in all of its diversity, as it exists today and what it might become tomorrow?  Is there something wrong with this simple question about the definition of The Internet that makes it so hard to answer?

This paper sets forth on a modest quest: make fun of previous definitions of The Internet; attempt to provide a foundational exploration of what a good understanding of The Internet might look like; and finally, look at why none of this (usually) should or does matter. 

A response to this modest quest might be that what The Internet is depends upon why you are asking.  There may be a degree of truth to this.  The cubist will argue that there are multiple perspectives of a horse, all of which are valid.  The modernist will argue that how we perceive of the horse depends on where we ourselves stand.  Nevertheless, there is a horse there.  While there may be multiple perspectives of a horse, looking at the bird sitting on the horse is not looking at the horse.  Part of the mission here is to say to the courts or to any other legal body, stop looking at the bird.

What it Aint

We have perhaps all heard the popular and tedious definition of The Internet as "a network of networks." [2]   Boring.  This tells us little about The Internet.  The telephone network is a network of networks.  The cable television system is a network of networks.  The lawyers in Washington, D.C. circulate in networks of networks.  This definition provides us little useful information about The Internet (but perhaps this concept will have to be revisited later in the paper).

Nor is The Internet a network of "interoperable packet switched data networks." [3]   This definition was the first codified as a part of the Communications Decency Act. [4]   The first two statutory attempts to define the Internet come from the first two statutory attempts to censor the Internet - both of which have been declared unconstitutional. [5] The definitions remain in the wreckage of those laws. With regards to this first statutory definition, there are lots of packets switched networks using other protocols known as Asynchronous Transfer Mode (ATM), Frame Relay, or the ITU's X.25. [6] There can also be IP networks that are not themselves apart of The Internet. The Internet is just one packet switched network in a family of packet switched networks.  A good definition of The Internet is something more than "a network of networks" and it is something more than "a network of packet switched networks."

The second attempt to censor The Internet, the Children's Online Protection Act, brought the second definition. [7]   And although the CDA and COPA are codified in sections directly next to each other, the two definitions are different.

47 U.S.C. § 230(f)(1) (the Communications Decency Act)	47 U.S.C. § 231(e)(3) (the Children's Online Protection Act)
Internet: The term ''Internet'' means the international computer network of both Federal and non-Federal interoperable packet switched data networks.	The term "Internet" means the combination of computer facilities and electromagnetic transmission media, and related equipment and software, comprising the interconnected worldwide network of computer networks that employ the Transmission Control Protocol/Internet Protocol or any successor protocol to transmit the information.

 The first item to note is that these definitions are not the same.  Sec. 230 is far broader than Sec. 231.  The set of networks covered by Sec. 230 is the over inclusive all "interoperable packet switched data networks." Sec. 231 is better - but not by much.  The problem with Sec. 231 is that all it has done is narrow the set of qualifying networks to all IP networks.  There are IP networks other than The Internet. [8]   Merely looking at a networking and saying, "I spy with mine own eye ... IP" is not enough to conclude that you have the genuine product in front of you. [9]  

 The Internet is defined in federal statute six times.  These definitions follow in the footsteps of Sec. 230 [10] and Sec. 231. [11]

While legislators struggled with a definition of The Internet in terms of the network, some judges have blundered in a different direction.  In a domain name trademark dispute, one district court judge concocted the following definition:

The Internet, or the World Wide Web, is a network of computers that allows people to access information stored on other computers within the network.  … "Information on the Internet is lodged on files called web pages, which can include printed matter, sound, pictures, and links to other web pages.  An Internet user can move from one page to another with just the click of a mouse." [12]  

What is wrong with this definition?  This mistaken definition occurs with unsettling regularity. [13] This is a definition of The Internet network in terms of an application.  There are many applications that can be utilized over the Internet: email, USENET, IRC, chat, games, streaming media, IP telephony, file transfer protocol, peer-to-peer file transfer (aka Napster), and many others.  The Internet network was intentionally designed so that the network transmission capacity would be separate from, disinterested in, and have no involvement in the higher layer applications.  The network just moves the bits creating an open platform for applications.  What you do with that platform is your choice.  You make this choice at your end of the network where you utilize the application of your choice. [14]   And as this stupid network [15] is not optimized for any application, it means that the end user is empowered to utilize any application of desire.

Furthermore, there is a significant limitation in understanding the Internet in terms of a single application.  Each application behaves differently.  To conceive of how the network behaves by referencing one application is to fundamentally miscomprehend how the network will behave with another application.  The interactivity of USENET is different than the interactivity of the World Wide Web which is different than the interactivity of telephony. 

Finally, these applications are not particular to The Internet.  Email, chat, telephony, even the World Wide Web can be provisioned over networks that are not The Internet.

In some ways it is an understandable mistake.  People comprehend a thing by what they can observe about a thing.  People comprehend the Internet by what they observe, which is the applications.  An end user interacting with a computer is conscious not so much of the Internet as the end user is conscious of the email application being used or the web page from which the user is gathering information.  The Internet is made manifest through the applications that are visible on a computer screen.  However, like the man behind the curtain, the Internet is not, in fact, those applications

                  Public

With a given degree of regularity, The Internet is described as The Public Internet. [16]   In looking at what The Internet aint, it is useful to note that The Internet aint public.

There are perhaps four relevant definitions of "public" in this context.

(1)     Communications Networks - public communications networks are common carriers regulated under Title II of the Communications Act. [17]  

(2)     Government Owned - certain property is described as public in that it is government owned and used for the benefit of the public.  This includes, for example, public parks.

(3)     Trespass - this is a concept of physical access to a space.  Where access to a space is highly restricted, it is said to be private.  Where access to a space is free and open, it is said to be public.

(4)     Service - Institutions which serve the public such as stores or restaurants might be considered public and fall under certain laws with regard to the public such as forbidding discriminatory service. [18]

 When the Internet is described as public, it is generally the third definition that people are using.  The Internet is public in that everyone has the opportunity to have access to that network.  There is no trespass barrier to entry; people come and go freely.  While access may be denied specifically via one means of access (for example, a college network that serves only the students but not the town), in a competitive market another means of access is likely to always be available giving individuals unrestricted Network access.

 It certainly does not mean definition (2); the government does not own the Internet.  The government owns individual pieces of the infrastructure and the government funded the development of The Internet, however, most of The Internet's infrastructures are privately owned. 

 It does not mean definition (1); the Internet is not a communications common carrier. [19]   In communications policy, The Internet falls within the category of enhanced service provider (aka Information Service provider).  Information service providers are not regulated by the FCC (although they are primary beneficiaries of FCC policy). [20]

 Finally, with regard to definition (4), while this may be true of some ISPs, it is not true of all ISPs.  Some ISPs are commercial Internet Access Providers who provide Internet access service for anyone who desires it.  But many Internet networks do not.  These include educational campus networks, corporate networks, and other closed networks that provide access only to a select class of subscribers and are otherwise closed.  These are not public networks; would this mean that they are not part of The public Internet?

 Lesson learned? This term public is poorly referenced and easily morphs into something else.  When people mean "a network open to all comers (the trespass concept)," this quickly morphs into definition one: communications carriers.  And some regulator somewhere will say, well if the Internet is public, then it ought to be regulated just like every other public communications network. We were talking trespass and, magic, now we are talking title II FCC regulation.  Neat trick. 

What it is

The Federal Networking Council Speaks

We are not without a good definition with which to start.  In 1995, the Federal Networking Council came up with the following definition:

On October 24, 1995, the FNC unanimously passed a resolution defining the term Internet. This definition was developed in consultation with members of the internet and intellectual property rights communities.

RESOLUTION:

The Federal Networking Council (FNC) agrees that the following language reflects our definition of the term "Internet".

               "Internet" refers to the global information system that --

(i) is logically linked together by a globally unique address space based on the Internet Protocol (IP) or its subsequent extensions/follow-ons;

(ii) is able to support communications using the Transmission Control Protocol/Internet Protocol (TCP/IP) suite or its subsequent extensions/follow-ons, and/or other IP-compatible protocols; and

(iii) provides, uses or makes accessible, either publicly or privately, high level services layered on the communications and related infrastructure described herein. [21]

 This is a good definition.  It contains many of the general characteristics one might think of when one ponders The Internet.  It includes (1) interconnection, (2) IP, and (3) the end-to-end policy.  It can appropriately, therefore, be set up as a straw man for the sake of discussion.

IP

Perhaps the preeminent defining characteristic of The Internet is the Internet Protocol.  This is where The Internet gets its name.  Where you have IP, you can have The Internet; where you do not have IP, you cannot have The Internet (right?).

 But The Network did not always use IP.  Starting in the early 1970s, The Network, at the time known as The ARPANet, used the Network Control Protocol (NCP). [22]   In 1974, Robert Kahn and Vint Cerf introduced a new network protocol, the Transmission Control Protocol (TCP). [23]   However, TCP had technically limitations and was eventually unpacked into two separate protocols, TCP and IP. [24]   In time, this won favor, and the US Government declared that The Network would migrate from NCP to TCP/IP (aka IP version 4) on January 1, 1983 [25] (an early Y2K drill).   But the role of the Internet protocol in transforming communications was not yet assured. Further incidents helped to place this protocol in its current preeminent role. TCP/IP was incorporated into the UNIX operating system, resulting in its propagation throughout the academic research community. [26]   In 1985 the government decided to build NSFNet in response to a need for a national backbone network to connect academic and research networks. The government made the crucial decision that TCP/IP would be the protocol for NSFNet. [27] Finally, Microsoft elected to include the TCP/IP in the Windows operating system.  In a sinfully reductionist history, this is how IP grew to take over the world. Currently The Network uses IP version 4 and is in the process of migrating to IP version 6. [28]

 The Internet network is a virtual network at the logical network layer, [29] overlayed on top of physical networks (ie, Ethernets, telephone networks, cable networks, wireless networks).  Designed to be provisioned over a diversity of physical networks without the need for altering those physical networks, the overlayed logical network provides a means for these networks to exchange information. The Internet supports in the higher layers a diversity of applications and services.  TCP/IP is indifferent to the physical network and the applications utilized.  It is intentionally a thin protocol which leaves the intelligence for the higher layers, at the edge, where the end user is empowered to communicate at will.  The Internet network is simply in the business of passing the bits without acting on the content. [30]

Figure 1: In a simplified example: (1) User at a computer requests the retrieval of certain content.  Computer turns request into data packets and injects them into The Internet over a telephone line.  These packets hit a router and are forwarded to the next router on an Internet path over wireless.  Here the packets are routed over the last Internet path over fiber.  (2) The server responds by delivering the content.  The content is broken down into packets and is transmitted back through the network (not necessarily over the same paths) back to the computer making the request.  The logical network exists from one end to another.  It is overlayed over physical networks.  The physical networks vary at different legs of the journey.

 Internet packets are not products of the network.  The Internet protocol stack is within the computer device of end users.  End user interacts with an application.  The computer then takes that content and breaks it down into Internet packets, with each packet individually addressed with an IP number.  These packets are then injected into the network and are transported until they reach the next point, the next router, in the network.  The router reads the IP number address on the packet (it does not interact with the content), consults an IP number routing table, and routes the packet based on what that router believes to be the best known route to the IP number destination.  The packet will in this way hop through the network from router to router until it (hopefully) reaches the device with that IP number.  Different packets can travel different routes through the network until they are reassembled at the destination.  All that the network is doing is processing the packet based on the IP number and, where appropriate, engaging in error correction.

 The use of IP is a crucial characteristic of The Internet.  Years ago some of the early proprietary commercial computer services (ie, Prodigy, Compuserve, and AOL) did not use IP in their networks.  It was common practice to refer to these services as online service providers but not Internet service providers. [31]  

 The FNC indicated that TCP/IP was a key defining characteristic.  However, as noted, TCP is used for error correction and control of transmission rates. For some applications, the behavior of TCP is not only useless but also destructive.  Traditional Internet applications are not real time.  If the packets arrive now or a few moments from now, it does not matter.  However, newer applications, such as telephony, are time sensitive.  If a packet of data does not arrive now, with the voice that is a part of the conversation, it is no longer useful.  Retransmitting it merely gets in the way. Therefore, many real time applications have migrated from TCP to UDP (User Datagram Protocol). [32]   Thus, while the FNC defined the network protocol in terms of TCP/IP, it might be more appropriate to defined it solely in terms of IP (unless we want to go in the opposite direction and conclude that any application failing to use TCP is not on The Internet, even though it is fully interconnected, reachable, and visible using the IP number addressing system).

 Finally, the FNC nailed the definition of The Internet to a particular protocol.  There is some degree of difficulty being so technically specific. The network protocol is in flux, changing at different moments in The Networks history (ie. NCP, TCP, TCP/IP, Ipv4, Ipv6).  What happens when the technical specification changes? Does each alteration of the technical specification result in an identity crisis where the question must be asked, "is this thing still The Internet"? The Internet protocol itself is almost 30 years old.  Might it be replaced with something novel tomorrow, meeting the demands of the new network?  In order to avoid answering rather bothersome questions, we skip forward to the next consideration.

Extensions, Follow-ons, Successors

In response to the above discussion, an imaginary proponent of the FNC definition might argue, "well yes, but the FNC definition was not limited to TCP/IP.  It anticipated evolution of the protocol with the phrase 'subsequent extensions/follow-ons.'  The FNC was not so daft as to think that the network protocol was static; but recognized that it evolved with time and that the heirs of the Internet protocol ought properly be considered likewise The Internet."

 While this is certainly true, it is also saying that the Internet is the Internet regardless of how it mutates or transforms.  As long as the mutant has some historical basis in IP, well then it is The Internet.

 It creates a bit of a problem if the object of our definition is permitted to evolve into areas unknown.  Assume that we have created regulation X for the Internet.  And assume that in the best of all possible worlds that regulation X is the best of possible regulations.  Now the defenders of regulation X will tell you that this regulation of the Internet was not created in abstraction with no regard for the actual nature of the Internet; rather this regulation, as was required of it, took into account economics, societal values, the public interest, freedoms, and other interests.  This regulation was created specifically with a set of assumptions concerning the context of the Internet.

 Now change those assumptions.  If The Internet morphs into areas unknown, shedding and distorting those things which led to those assumptions, then the best of all possible regulations, regulation X, has just become a problem.

 The law has a need.  In order to effectively address something, it has to know what that something is.  If the definition of that something is vague, then the laws effective intercession with that thing experiences problems.  This is a crucial point.  Many purists would prefer that The Internet remain undefined so that the law lacks a target and The Internet can remain a libertarian paradise.  But this is divorced from reality.  Reality is that legislators and policy makers are making decisions on a day to day basis concerning The Internet and activity over The Internet.  Sen. Exon, who reportedly had never been online, introduced the infamous and ill fated Communications Decency Act. [33]   Congress has every year sought to criminalize Internet gambling.  And there are countless programs where funding is being provided for Internet projects.  Lacking solid definitions of the object of these policies, these policies potentially may go astray and in the end may cause more damage than good.  One who professes to desire to protect the Internet by ensuring that policy makers have before them only vague and misconstrued definitions defeat their own desires.  Good law is based on a good foundation.  As the computer science phrase goes, "garbage in; garbage out."

 Defining a thing for what it might become is problematic.  One might think of a definition of beer.  According to Mr. Webster, beer is defined as "an alcoholic beverage usually made from malted cereal grain (as barley), flavored with hops, and brewed by slow fermentation." [34]   Now consider the suggestion that beer is beer and anything it might become.  Well, it might become Welsh Rarebit. [35]   But there is hardly a drunk Welshman who might confuse a pint and a plate of gooey cheese poured impossibly over some toast.  A thing is properly defined as what it is and not what it might possibly become. 

 After 7 years of trying, let us assume that Congress finally passes a law regulating Internet gambling.  Now this law assumes that the Internet is this vibrant global information system where borders mean nothing and 15 year old boys can gamble away their college savings with a guy named Lou operating a somewhat tilted server on some island.  Everyone assumes that this Internet gambling law is the best of all possible laws and prevents 15 year olds from getting cyberscammed.  Now assume that in 5 years, the mega corporation AOLTWCBSMCDONALDS builds a super new network which is "an extension of or follow-on" of the network.  This new network is everyone's dream come true, a network of highly connected super artificially intelligent toasters. [36]   These toasters have one desire, to toast.  They do not gamble.  And by the time they are 15 years old they have become obsolete.  According to our definition, this mutant evolutionary off shoot is The Internet.  But now our best of all possible regulations designed to protect 15 year old boys (or perhaps to protect Lou) now protects a pack of hyper artificially intelligent toasters.

 Jurisprudence disdains vagueness.  Vagueness leads law into unanticipated and unintended applications. [37]   To define something is the act of "describing, explaining, or making definite and clear" the thing as it exists. [38]   This does not include the rather large set of things that it might potentially become.

Interconnected

The FNC lists as a definition criteria that the network is logically linked. The Internet is a "network of networks."  Ack!  Having previously denigrated use of that phrase, how can I now use it?  The phrase does get to an important concept (although in and of itself the phrase is entirely insufficient and vague).   The concept is interconnection. IP creates a virtual or logical network over the different physical network, logically linking together edge devices and creating the ability for these computers to interact with one another.

 There certainly was a time when a perception of The Internet as a “networks of networks” was valid.  The problem, in the early 1970s, that Robert Kahn and Vint Cerf set out to solve was to develop a network protocol to permit interconnection between otherwise incompatible networks.  They set out with the purpose of connecting ARPANet to SATNET, PRNet, and ALOHANet. [39] Thus, in the 1970s, the perception was of multiple networks that themselves needed somehow to be networked.  Now, I would suggest, the exact opposite perception is true.  No longer are the underlying networks the primary concern.  Now the end to end Internet itself is the key concern, and what underlying network will be used to access it has become the secondary concern.  The Internet has gone from being the glue of other networks to being The Network itself.

 The physical networks are interconnected at the logical network layer.  An ethernet communication goes to an ethernet address at the other side of the network and is terminated at a black box.  There is not necessarily an ethernet at the other side of the black box.  It could be any network (telephone, wireless, cable...).  Any sense of this continuing to be an ethernet communications ceases.  What goes from one edge to the other, traveling over multiple interconnected networks, is The Internet. To be "On Net" or to have Internet access is to be linked with this unique network of interconnected devices and users. [40]

 This criteria is useful.  Confronted with the question, "is a long distance telephony network that utilizes IP in the backbone a part of the Internet," the answer would be no.  The long distance IP telephony network is not interconnected. 

 There is, however, a degree of vagueness here. What is mean by "logically linked" or interconnected?  If someone interconnects three IP networks, would that be The Internet?  This network of networks has IP and is interconnected.  The answer, as suggested by FNC definition, is that not just any interconnection will do.  There is a specific historical networked community known as The Internet that is interconnected.

 Now a hypothetical test: Imagine, if you will, a huge intranet.  Imagine that a class of users had become dissatisfied with the performance of The Internet and therefore, without changing the Internet protocol and using the same IP address system, constructed an private intranet for their own purpose.  This intranet, in order to ensure top performance, was constructed with top of the line electronics, premium design, and sufficient bandwidth.  Having constructed this high performance intranet and in order to maintain its high performance nature (i.e., uncongested), access was restricted.  Only those on the intranet could communicate with others on the intranet.  You could not communicate from the intranet to The Internet and you could not communicate from The Internet to the intranet (though the end user may hardly notice as intranet traffic is simply properly routed over the intranet and outside traffic is simply routed over The Internet).  Would it be fair to conclude that this intranet should not properly be called the Internet - it is not logically linked (interconnected) to the rest of The Internet (you cannot get there from here)? Would it be fair to conclude this when the intranet is called Internet2?

Address Space

The FNC definition states that The Internet "is logically linked together by a globally [41] unique address space based on the Internet Protocol (IP)."  It is compelling to think of a network in terms of its address space. One of the tell-tale signs of what network one is using is how the network is addressing the traffic.  The Public Telephone System uses the North American Numbering Plan, in other words, telephone numbers. [42]   To utilize the public telephone system, one needs a telephone number.  This number is entered into the network in order to transmit communications from point A to point B.  If you are communicating with people using telephone numbers, this is a good indicator that you are on the Pubic Telephone Network. [43]   If an end point lacks a telephone number, it is not on the Public Telephone Network.

 A point that is reachable by a network's addressing system is on that network.  If it is not reachable, if it cannot be "seen," then it is not on that network. 

IANA

What is the "unique address space" of The Internet?  It is the unique set of IP number addresses.  Each device on The Network is assigned an IP number.  Internet packets traveling from one device to another are addressed with the IP number of the destination and the IP number of the originating device.  To be visible a device must be associated with an address; if a device is not associated with an address, it cannot be reached and cannot be said to be on the network.

 In the early days of The Network, Jon Postel [44] at UCLA and then later at ISI nobly volunteered to administer the total set of IP numbers.  This administration of IP numbers came to be called the Internet Assigned Number Authority (IANA), [45] which set the policy and determined who would be allocated what blocks of IP numbers. With the formation of the Internet Corporation of Assigned Names and Numbers, the IANA function, by contract with the Department of Commerce (ICANN), [46] is now under ICANN. [47]  

 The "unique address space" of The Internet is the specific historical set of network addresses that have historically been managed by IANA.

Behind the Curtain: NATs & PBXs

The criteria of addresses begs the following question: there are devices on the telephone network that are behind PBXs.  There are devices on the Internet that are behind NAT boxes.  The PBX permits access to devices behind it through adjuncts to regular telephone number addresses.  The NAT box permits access to devices behind it through adjuncts to the IP number addresses.  The devices behind the curtain otherwise have interconnection to the networks. 

 Are these devices "On Net?"  These devices are interconnected.  They are reachable using the addressing system of the network.  While it is true that address may only function for part of the network, to either the NAT Box or to the PBX, nevertheless, the address (plus an adjunct) is used to reach the end point.

 This is a matter of perception.  Would it be coherent to suggest that someone behind a NAT box who has full access to Email, Web, streaming media, and other applications over the Internet is not "On Net"?  Would it be coherent to say that someone whom I can reach by dialing a telephone number over the telephone network is not "On Net"? 

 I would suggest that these devices are "On Net" and are apart of The Internet.  They are logically linked.  They are visible (at least to some degree).  The network address may provide only partial delivery of the communication but that network address is at the core of the address of that device.

Domain Names

The Internet has an interesting network addressing system.  The devices on the network are reachable using IP numbers.  Those are the network addresses.  But for all practical purposes, the humans are unaware of those network addresses.  Instead, recognizing years ago that the IP numbers are difficult for the humans, the domain name system (DNS) was developed.

 The DNS is a database that maps letter strings to IP numbers. [48]   The domain name cybertelecom.org is mapped to a particular IP number, for example 192.16.0.192.  An individual who wishes to view the Cybertelecom website can remember the domain name and enters it into a web browser.  The first thing the browser does is a DNS look up, asking a DNS server what IP number that domain name is mapped to.  Having the IP number, only now can the browser make a query to the cybertelecom.org web server for the content.  The web server responds by sending packets of content to the IP number of the device that originated the query.  In other words, while the human may know only the domain name, the devices on the network communicate with each other with only the IP number.

 There are two wrinkles with this.  First, while the DNS is the compelling way with which humans communicate with each other over the network, it is not necessary.  Communication can be achieved with only IP numbers or a different alias system can be mapped to IP numbers, such as instant messaging names.  While the DNS is fantastically convenient, it is still an adjunct to and not a necessary component of the network address system.

 Second, domain names have the added advantage of remaining constant regardless of what IP number they are mapped to.  The website cybertelecom.org has changed web hosts several times.  The humans are not aware of this.  They simply enter the domain name, do a DNS look up, get the new IP number of the new host, and they are happy.  In this way, destinations on The Net can have smooth portability.  A user can move from service to service with the domain name remaining constant while the IP address changes.

 It is also true that humans on intranets do not always have the same IP number even though they might retain the same domain name.  Joe may log on to the network with his email address joe@cybertelecom.org.  When Joe logs on, the intranet, using dynamic host control protocol, assigns Joe a new IP number from its pool of IP numbers.  Thus, Joe's network address is changing on a regular basis within the confines of the pool of IP numbers allocated to that network, while again, his domain name email address remains constant.

 This demonstrates how powerfully the DNS compliments the IP number system.  But Joe is still reached by IP numbers.  Fred, trying to reach Joe, enters Joe's email address in the application.  A DNS look up is conducted and, magically, the same IP number is produced every time.  But doesn't Joe's number change on a daily basis?  Yes, but the IP number produced for Fred is the IP number not of Joe but of the Cybertelecom email server.  Fred's email application uses that IP number to send the email, not to Joe's computer, but to Cybertelecom email server where it is stored until the next time the Joe logs on.

 The point is not to get too technical with the functionality of the DNS.  The point is that ultimately the DNS is an adjunct to the IP number system that makes the IP number system useful to humans.  The network itself communications with IP number addresses.  So close is DNS to the addresses system of The Internet that it is difficult to talk about Internet addresses without talking about DNS.  But DNS is an adjunct, it is not necessary, and packets are routed through the network using IP numbers, not domain names.

E2E

Another defining characteristic as noted by the FNC is the end-to-end design of the network. The Internet religious among us will suggest that end-to-end is an essential characteristic of The Internet. [49] The Internet protocol is a very thin and stupid protocol.  It does very little beyond pass packets back and forth.  It does not care what physical network the packets originate on.  It does not care what application they are from.  The Internet Protocol virtual network simply transmits the packets back and forth to destinations on the network without change in form or content of the data transmitted by the network itself.  There is no interactivity or interference with content at this layer.

 As described originally by J.H. Saltzer, David Clark, David Reed, [50] application functionality is reserved for higher protocols layer in the Internet protocol stack. [51]   Application functionality is removed from the logical network layer so that all that the logical network does is pass packets without interactivity.

 This means that instead of having a network with intelligence in the core where the network is optimized for a single application, like the telephone network, the intelligence resides at the end. End users are confronted with a virtual network over which they can utilize any application. The Internet is not optimized for any application.  All packets from applications are equal and are transmitted with best effort.  The end user can select any application and indeed must select some application before communication can be initiated.  The processing and interactivity involved in that application occurs at an edge and not within the network itself.

 The Internet is The Internet all the way to the end.  At the end generally is a computer.  It is in that computer that the higher layer intelligence (applications, services, and content) exist and is created. The computer uses applications to create content and injects this content as packets into the network. The network itself does not interact with the content.  The content is created and processed at the end.  Thus it is inaccurate to say that, for example, an ISP gives and end user access to The Internet, as if The Internet were some far off and remote thing.  Rather, the ISP provisions Internet connectivity.    Every device and end user that has Internet connectivity is "On Net" and is a part of The Internet.

 Finally, note that while the FNC definition gives recognition to the higher layer applications that The Internet supports, it goes no recognition to the physical networks over which The Internet is provisioned.  This is important as The Internet is regularly confused with the underlying physical networks over which it is provisioned.

Historical

Of course The Internet is not just any network; it is a unique network in history.  Unlike, for example, cable, where there is a wide assortment of different cable networks out there, there is one unique network in history with the name The Internet. The Internet was born on an October day in 1969. On that day, a pack of geeks located at UCLA sent the first transmission over The ARPANet to another pack of geeks at Stanford.  By the end of that year, there were four nodes on The ARPANet. By 1971, there were 15 nodes. [52] This unique networked community continued to grow, adding networks, devices, and users.

 In 1983, the US Government decided that The Network would migrate from the existing NCP network protocol to the TCP/IP protocol.   This unique network came to have the name The Internet. [53] And The Internet continued to grow.

 The National Science Foundation recognized a need for a new backbone to service the community and thus established the NSFNet. [54]   The NSFNet however had an Acceptable Use Policy the prohibited commercial traffic (driving the creation of alternative commercial backbones).  As demand for commercial use of The Internet grew, commercial networks, which could not use NSFNet to exchange traffic, set up the Commercial Internet Exchange (CIX). [55]   Eventually NSF altered the NSFNet AUP to permit commercial traffic, established the Network Access Points (NAPs) for traffic exchange, and privatized the NSFNet. 

 This Internet continued to grow but was nevertheless largely hidden from popular sight.  Then, in 1990, Tim Berners-Lee invented the World Wide Web and in 1993, Marc Andreesen invented the Mosaic Web Browser, [56] and The Internet broke into popular use.

 The Internet is a unique networked community that traces its history back to a particular day in history.  It would be hard to identify this network with a specific technology as the technically has evolved over the years and is used in other networks.  What has been unique and identifiable has been the networked community.  It grew from a small community of users experimenting with novel packet switched communications, to a global community of millions of users engaged in an information revolution. 

In Conclusion, It's The Wrong Question

So enough of making fun of other people's definitions.  What is The Internet?

 After hammering on this question long enough, one realization is that it is not so much a definition as it is a proper name (this is why The Internet is capitalized).  The Internet is the proper name of a unique network. 

 Definitions of The Internet generally collapse into circular statements. The Internet is defined as the network with the name The Internet. Consider that one defines [57] terms [58] but not necessarily proper names. [59]   A term is defined by a set of differentiating criteria; things that meet those criteria are identified with that term. If a flying disc (aka Frisbee) is a "a plastic disk several inches in diameter sailed between players by a flip of the wrist," [60] and if that thing there on the ground is a "a plastic disk several inches in diameter sailed between players by a flip of the wrist," then we can joyfully conclude that the thing on the ground (and all other things that meet the definition) is a flying disc. Definitions provide us with differentiating criteria that allow us to determine whether a thing in question is properly labeled with a term.

 Now consider proper names and the question, "who is Bob."  The answer is that "Bob is the person with the name Bob."  Things with proper names are not so much defined as they are described.  "Bob" is that particular tall person over there with a beard and blonde hair who generally wastes too much time on Internet law issues.  But it is not true that all tall blonde people with beards are Bobs.  And if Bob's description changes, if Bob shaved his beard and dyed his hair green, he would still, in fact, be Bob.  

 There is only one computer network with the name The Internet.  Other networks that meet the technical description of The Internet nevertheless are not The Internet (these may properly fall under the definition of an internet, a computer network that is logically linked using IP but which is not The Internet).  And the technical description of The Internet can change over time (for example, from NCP to TCP to TCP/IP to IPv4 to IPv6) and still be The Internet.  This is true because The Internet is a name given to a particular network rather than networks that meet certain technical specifications.

 

Even if "what is the definition of The Internet" is the wrong question, when confronted with a bunch of wires and boxes and blinking lights, one would like to have the ability to understand whether these things are The Internet and how this Internet thing is distinguished from anything else. 

 The Internet is the name of a unique networked community that saw its birth in October 1969.

·         On October 25, 1969, the network comprised of two end points.  At the end of the year it had 4 end points.  Slowly this networked community grew and in the year 2002 it is estimated that there are 580 million individuals on the network. [61]

·         In 1983, this Network migrated to the Internet Protocol, from which it got its name.

·         The Network is logically linked using the unique address space based on IP, currently managed by IANA.

·         The Network is a computer network overlayed over physical networks. 

·         The Network supports higher layer applications following the end-to-end principle.

 This is not intended to be a final word.  Instead it is intended to be a first step, or perhaps a shove, in the right direction. It is not far from the FNC definition - only now it is a description.  The primary revision is to change the perspective from a focus on technology to a focus on the unique networked community. Also, this describes the network relationally both down and up.  It describes The Internet in relationship to the physical networks over which The Internet is provisioned (something which the FNC did not do).  It then describes The Internet in relationship to the applications and services in the higher layers which it supports (the end to end principle).  In this way it clarifies that the Internet is a particular network that is neither the underlying infrastructure nor the overlying applications. It is a particular logical network, created over existing physical networks, creating a global network of interconnected devices and users.

Post Script: Why It Doesn't Matter

 It is prudent to struggle for a proper understanding of The Internet so that when Internet legal issues are addressed, things do not get mucked up.  But in struggling so hard for a definition, and then realizing that perhaps seeking a definition was the wrong endeavor, perhaps there is another realization; perhaps the definition does not or should not matter.

 The Internet is a specific network. The network can change.  The technology can evolve.  Use of the network can change.  Users can become fickle and prefer the next pet-rock fad in communications.  There are other networks.  Is it prudent to create law particularly based on one specific medium?  The struggle to define what The Internet is, begs the question, "why are you asking." 

 Consider this.  Two guy break the law playing three card Monty.  One is running a scam on a webserver based on some off-shore island.  The other has a beat on the corner of Bleeker and Sullivan. Would it make sense for the law that applies to be different for guys A and guy B?  Should one spend two years in jail and another spend four?

 The argument here is for neutrality.  The application of law should probably not vary depending upon technology. [62]   Suggestions of the Internet as a unique and mythical medium where business plans defy the need for making a profit, where criminal activity defies jurisdictional boundary, and where The Network for some exceptional reason defies definition, are (hopefully and thankfully) gone.  The mythology of the Internet having been dispelled, perhaps we can return to the conclusion that a policy goal is a policy goal online or off.  Online and offline activity should generally be treated equally.

 Returning to the example of Internet gambling, here is a phenomena that has created the opportunity for anyone with a computer and a credit card to easily waste away vast sums of money, engaging the offerings of off-shore gambling companies.  The cyber casinos are not regulated.  There is no assurance that the game is not rigged.  There are no background checks on the operators.  Scams can fold up shop at a night fall.  There is no money going into a state's coffers.  There are no state citizens being employed.  Removed from the highly social glitz and glamour of Las Vegas or Atlantic City, Internet gambling is a solitary between individual and computer screen, where compulsive gamblers can become entranced and minors can act like adults. It is a disruptive innovation to the business plans of established incumbents gambling businesses. 

 Every Congress since the 104th Congress has introduced legislation to combat Internet gambling. [63]   And they have come close to success.  Congress has generally sought to meddle with The Wire Act which prohibits the business of gambling on sports over wire communications. [64]   While the Department of Justice has supported clarification that The Wire Act applies to all situations including Internet gambling, it has opposed Congressional attempts to solve the problem by creating separate and inconsistent legislation.  Get nailed under the Wire Act and get treated one way; get nailed under the new Internet gambling law and get treated a different way.  DOJ repeatedly advocated that Congress lose the technology bias:

[The Internet Gambling Prohibition Act of 1999] is not technology-neutral, but applies only to Internet gambling while leaving the existing prohibition on gambling over "wire communication facilities" in general unchanged. While the Department is generally concerned about legislation designed for particular technologies such as the Internet, it is specifically troubled here by the creation of two inconsistent gambling prohibitions - one expressly for the Internet and a different one for the use of wire communication facilities (which includes the Internet).

 Indeed, any effort to distinguish Internet transmission from other methods of communication is likely to create artificial and unworkable distinctions. For example, we expect digital Internet telephony to grow in popularity over the next few years. How would we deal with gambling that occurred over this technology, which would use the Internet or other packet-switched networks for pure voice communications? Would it be under the proposed section 1085, which is designed specifically for the Internet, or under section 1084, which deals with wire communications in general (but also includes the Internet)? This is especially problematic, as section 1084 and the new section 1085 proposed by H.R. 3125 would have different standards and punishments.

 The Department urges Congress to identify the conduct that it is trying to prohibit and then to prohibit that conduct in technology-neutral terms. The fact that gambling, an age-old crime, has gone high-tech and can now be done through the Internet, is no reason to pass new laws that specifically target the Internet for regulation. Passing laws that are technology-specific can create overlapping and conflicting laws prohibiting the same activity, but with different legal standards and punishments. This will be the result if H.R. 3125 is enacted in its current form. We will have both section 1084, which we’ve used to prosecute Internet gambling, and a new section 1085 which would prohibit some, but not all, types of Internet gambling. This overlap in the statutes can only complicate law enforcement’s efforts on the Internet gambling front. [65]

In response to the endeavor to define The Internet comes the begged question, "why are you asking."  Perhaps the answer is, "you should not be asking."  If the legal concern is for activity in the application or content layers (gambling), then focus the legal solution on the layer where the problem exists, the application or content layer.  The underlying network layers can be swapped arbitrarily, and yet the policy concern with the activity (the application) likely remains the same.  Laws inconsistently applied in different spaces ought to be greeted with a raised eyebrow. 

When does it matter?

But wait, having said that The Internet should does not matter, I back peddle to suggest that there are times when it does. Circumstances where a proper understanding of The Internet is necessary include (1) First Amendment law, and (2) communications law.  There are certainly other occasions where the unique environment of online communications leads to unique legal concerns (but even here, the law may wish to consider general online concerns as opposed to specific Internet concerns).

In different circumstances, the federal government has greater ability to restrict speech than in others.  On television and radio, the government has been able to restrict the broadcast of "The Seven Dirty Words," a monologue by George Carlin. [66]   Conversely with printed matter such as newspapers and books, the government has significantly reduced ability to restrict the dissemination of adult content.  According to the Supreme Court as stated in Sable, [67] the First Amendment analysis depends upon a proper understanding of the medium.  Broadcast spectrum is scarce and held in the public trust, therefore the FCC has authority to restrict the broadcast of mature content to later hours.  Printed matter is a highly competitive medium, with great diversity, which is not invasive in your home.  While with a TV, you set the channel and passively absorb whatever pours from the set, reading material must be affirmatively acquired by the reader.  Likewise, with the Internet, the Supreme Court in Reno v. ACLU stated that the nature of the medium is crucial. [68]   The Internet is not scarce.  It is not held in the public trust.  It is not invasive.  Rather the Internet is a broad and diverse medium where the user goes out and selects the content of that user's choosing.  Given the nature of the medium, censorship by the government would be inappropriate.

Communications law is not uniform but is rather broken into silos.  Title II of the Communications Act [69] regulates the telephone network.  Title III regulates wireless services.  Title VI regulates cable.  While convergence of the mediums is making these silos arcane, nevertheless for the time, regulation applicable to a given communications medium is dependent on what that communications medium is.  In communications terms, it is important that The Internet is The Internet, an information service, and not a telephone service or a cable service.  This is at the center of several FCC proceedings. [70]