Locality of networks

From Citizendium, the Citizens' Compendium
Jump to: navigation, search
This article is a stub and thus not approved.
Main Article
Talk
Related Articles  [?]
Bibliography  [?]
External Links  [?]
Citable Version  [?]
 
This editable Main Article is under development and not meant to be cited; by editing it you can help to improve it towards a future approved, citable version. These unapproved articles are subject to a disclaimer.



An underlying concept, locality of networks, underlies a great many terms ranging from "personal area network" to "interplanetary network". Many physical, and some logical, networking technologies make an assumption of what area they serve, or, in other terms, what is local to them.

New technologies, such as wearable computers, ubiquitous computing, virtual reality, and automated checkout & logistics are enabled by new short-range networks.

Range Generic type of network Representative technologies Application
First meter Personal area network, RFID Bluetooth Replace computer-to-peripheral cable; physical object inventory; assistive devices in or on the body
First 10 meters Line-of-sight wireless Infrared wireless Control local devices (TV remote control, lights)
First 100 meters LAN star cabling, femtocell 10BaseT, 100BaseT Desktop to wire closet
First kilometer Local loop Digital subscriber loop, intracell cellular telephony, T1 carrier, E1 carrier Subscriber building to end office of telephone system
Second kilometer Drop from metropolitan area network, interoffice trunk, pair gain SONET, SDH, T3 carrier, short-haul optical, CWDM Interconnections among end office and higher-level switching centers; distant subscriber buildings; backhaul to cellular towers
First Tens of kilometers Metropolitan area network SONET, SDH, DWDM Metropolitan interconnection
Planetary Internet, Public Switched Telephone Network Logical packet switching, HF radio, very long haul optical One world
Interplanetary Interplanetary Internet project http://www.ipnsig.org/aboutstudy.htm One solar system, or users with very long delay time (e.g., reindeer herders who communicate weekly)

Scope of information

In dealing with networking technologies of different locality, the idea of scope or locality of reference applies to addresses and other identifiers. The levels of locality listed below are general concepts and have a variety of protocol- and implementation-specific variants.

  • link-local information is unique only on a shared transmission medium
  • area-local information is unique only to a certain part of the hierarchy of a logical addressing structure
  • domain-local information is unique only to the local logical structure. For example, if the logical structure used Internet Protocol version 4 in the private address space [1] and used Network Address Translation (NAT) to communicate with the public Internet, the private address space used on the "inside" of the NAT would be domain-local.
  • global information is unique to all reachable points of the logical or physical network, such as IEEE 802 address

Locality of general protocol information

Locality may be a matter of addresses needing to be unique only in a particular locality, or, instead, a limitation on certain protocol activity. For example, in the Open Shortest Path First (OSPF) protocol, certain information used in building the routing "map" may be distributed only on a link, only within an area, or only within a single OSPF domain.

Locality of addresses

It is a separate issue if the addresses, to which that raw routing information applies, need be unique to administrative entities, such as enterprises, extranets, or the public Internet. In a 1998 book, John Moy, the main architect of Open Shortest Path First (OSPF), summarized some thinking abut locality:

IPv6 can be thought of as an attempt to capture current IPv4 implementation in protocol specifications. Those IPv4 features that are either unused (for example, TOS-based routing) or discouraged (such as fragmentation by intermediate routers) have been deleted from the protocol specification.... In addition, address scoping has been made an initial part of the IPv6 addressing structure, building on IPv4 experience with private internet addresses and proposals for IPv4 multicast scope. Address scoping is a way of dropping the global uniqueness requirement for certain addresses.[2]

It is a decade after 1998, and there is now enough IPv6 experience to suggest that not all the features captured were necessarily a good idea; address scoping is much more complex than it might seem. The idea of a "private" address space that can be used by any enterprise has become less attractive, and the IPv6 work has come up with a method of declaring unique local addresses that will not be Internet-routable, but cannot be duplicated by different organizations. [3]

Interconnecting scopes

All too often, people have been taught oversimplifications that suggest one interconnects "layers", but that rapidly breaks down. The usual convention has been to say that routers link things at the network layer, gateways interconnect things at the application layer except when the term "gateway" is used for network layer functions. Older but generally accurate terminology had bridges as the interconnection device at the data link layer, and repeaters at the physical layer. The term "hub" also was used for physical layer interconnection devices.

In the eighties, there were indeed cases where bridges were faster than routers. The classical bridge, however, had certain limitations to which some reasonably elegant hardware solutions were applied. At first, these improved bridges were called switches, and "switch", for a time, had a fairly specific meaning. For Ethernet/IEEE 802.3 technology, they had microsegmentation and could operate in full duplex; some used cut-through forwarding.

Unfortunately, one networking vendor created the slogan "switch when you can, route when you must". Vendors that had specialized in routers began to call their equipment "switches" or "layer 3 switches", even when the layer 2 techniques ran into scalability problems and could be outperformed by true routers.

Yet other vendors, including Synoptics, decided to call all of their devices "hubs", for they were the hub of the network &mdash although their products were modular units that mixed physical, data link, and network layer devices in a single chassis. A cherished Cisco Internetworking Glossary reflected a planned merger between Cisco, who called their routers "routers", and Synoptics, in which router and hub technology would join in a device to be called the "Rubsystem", which, presumably, made sense to the marketing staff in California.

Perhaps Rubsystem was not worthy of life, but there came a time when almost every device was called a "switch", such that the term became meaningless. Things became even more complicated with the advent of middleboxes, such as firewalls and network address translators, that broke the end-to-end routing paradigm, potentially with multiple instances of the same addresses.

User perspectives

From the end user perspective, locality starts from the person, and moves outward. It may even begin inside the person, as with a cardiac pacemaker that electromagnetically couples to an interrogation probe connected to a monitoring control computer.

Perceptions of the network

While much of this article deals with transmission technology, evolution of human interfaces should not be forgotten. Ubiquitous computing creates an environment where the computer interfaces are everywhere, but taking actions invisibly to a user. A very basic use of ubiquitous computing, for example, turns on a room light when it senses a person has entered. A much more advanced version, demonstrated at Xerox Palo Alto Research Center, when a room was occupied by one person, it would sense the direction in which the person is looking, and turn lights on and off so the light was never in the occupant's eyes. When the system detected an additional person or persons in the room, it would change to diffuse lighting.

The other extreme of human interface is virtual reality, where, rather than the ubiquitous computing paradigm that senses a human in the real environment, the alternate paradigm creates the reality.

Both of these technologies will involve short-range, high-speed communications, for functions such as detecting a human in a room, or, in a virtual reality system, detecting body movement and changing the generated visual perspective to match.

Carrier perspectives

Some writers treat these in reverse, as, for example, the "last mile" problem, but there is little question that one of the most challenging, capital- and labor-intensive problems in communications is reaching the individual user, or small office.

FTTX

Carrier perspectives indeed often focus on the "last" increment of media, the subscriber end of fiber optic networks, generically Fiber to the X (FTTX), where:

  • Fiber to the home (FTTH) describes installation where optical fiber physically enters the subscriber premises and connects to a router or other end equipment there;
  • Fiber to the curb (FTTC) can refer to telephone installations where the fiber terminates in a curbside weatherproof pedestal, from which copper pairs connect to the subscriber locations. This is also seen in cable television networks where the main distribution is optical, but, at each telephone pole from which a customer will receive service, there is a converter box that picks off the appropriate signal and enters the premises via coaxial cable or, rarely, high-grade copper pair using Very High Speed Digital Subscriber Loop (VHDSL) technology.
  • Fiber to the building (FTTB) uses fiber to deliver a high-speed connection from the provider's point of presence to a multi-unit office or apartment building, where it is distributed using LAN technology. Depending on the speeds and distances involved, the in-building distribution may use fiber or high-grade (Cat5 or better) twisted pair. Coaxial cable LANs essentially are obsolete.

Digital Subscriber Loop

Wireless Local Loop

References

  1. Y. Rekhter, B. Moskowitz, D. Karrenberg, G. J. de Groot, E. Lear. (February 1996), Address Allocation for Private Internets, RFC1918
  2. Moy, John (1998), OSPF: Anatomy of an Internet Routing Protocol, Addison-Wesley, p. 23
  3. R. Hinden, B. Haberman (October 2005), Unique Local IPv6 Unicast Addresses, RFC4193