Technical English - Spanish Vocabulary

Topology, topología ( Informática y Computación ) En una red de comunicaciones, patrón de interconexión entre nodos, por ejemplo una configuración de bus, de anillo o de estrella.

Topologies
The topology of a network refers to the 'structure' of the network, ie how all the machines, termed participants or users, are connected.
The most simple topology is point to point - a single link between two machines (Figure 1a).

This generally works well in very small installations. When the installation grows and communication is required between all the 'participants' in the system, the configuration becomes very messy, see Figure 1b. This is commonly known as a mesh topology. As seen here, to connect eight users will require 28 lines therefore 56 interfaces. A ninth user is an additional 8 lines and 16 interfaces. This is clearly very expensive in hardware and installation.

As sites got bigger, so the bus or local area network (LAN) was developed. The concept here is to have one communication interface per user, and a single cable (or medium) connecting all users. Physically this is normally achieved in a tree (Figure 2a) or daisy chain (Figure 2b) structure. The tree topology uses taps or splitters to separate information from the main bus (trunk) and transmit it down the branches to the users. The daisy chain topology is very similar but has the main bus cable running into and out of the communication interfaces of the users. This method requires isolation between the electronics of the interface and the bus itself to prevent a failure of the interface from affecting the bus.
It is also necessary in some applications usually restricted by geographical layout to configure a network in a star topology (Figure 2c). This works a little like a tree network with a very short trunk and long branches.
After the development of bus type communication there immediately arose the problem of control of the bus. In point to point communication the control is a master-slave type control. This works well, as if either user fails, no communication can take place anyway. In a bus configuration, however, this is not always the case. For installations where the 'master' device is always in control and the slaves are 'dumb' devices which need only to communicate with the master, this topology is sufficient. In a distributed control environment however, where all users need access to the bus, the failure of a master station and subsequent loss of the communications network is not acceptable.
This leads to the concept of 'peer to peer' communication. In this format, no single user has control, but a protocol is developed to allow control of the bus to be shared between all participants. In PLC communications this normally takes the form of a 'token passing' network or 'carrier sense multiple access/collision detect' (CSMA/CD) network.
Token passing means that all participants on the network have a list of all the participants on the network including itself, usually in the form of an address or node number in ascending order. At any time, one of the participants has the token for an amount of time equal to or less than a pre-defined maximum time.
During this time it may send data to or request data from any other node. When it is finished, or its maximum time has elapsed, it will 'pass-on' the token to the next node in the list of participants and listen, as though it were a slave until it receives the token again.
CSMA/CD networks, or what is more commonly known as Ethernet, work on the principal of there being no absolute control of the network. Each user on the network detects for itself whether it is connected to the network. This is known as carrier sense. Once it detects a carrier on the network it sees the network as alive and accesses the network, sending to or requesting data from another user. Clearly as there is more
than one user on the network, more than one user may try to access the network at the same time, multiple access. Electrically these two messages will corrupt each other, so when this occurs, both users which have transmitted data will detect data on the network other than what it sent, collision detect. Both users then stop communicating for a random amount of time and then try again.

Topologías
La topología de una red se refiere a la “estructura” de la red, o sea, cómo todas las máquinas, denominadas participantes o usuarios, son conectadas .
La topología mas simple es punto a punto – un único lazo entre dos máquina ( figura 1a). Esta generalmente trabaja bien en instalaciones muy pequeñas. Cuando la instalación crece y la comunicación es requerida entre todos los “participantes” del sistema, la configuración se vuelve muy desordenada, ver figura 1b. Esta es comúnmente conocida como topología en malla. Como se ve aquí, para conectar 8 usuarios se requerirían 28 líneas y por lo tanto 56 interfases. Un noveno usuario agrega 8 líneas adicionales y 16 interfases. Esta es claramente muy cara en equipamiento e instalación.
A medida que los sitios se volvieron mas grandes, de igual manera se desarrolló el bus o red de área local (LAN , por sus siglas en Inglés ) . El concepto aquí es disponer de una interfase de comunicación por usuario, y un único cable ( o medio ) conectando todos los usuarios. Físicamente esto es normalmente alcanzado en una estructura en árbol ( figura 2a) o guirnalda ( nota de traducción: sinónimo técnico para estructura en árbol es topología de bus lineal y la topología de guirnalda también se conoce como topología de margarita ). La topología de árbol usa derivadores o divisores para separar la información del bus principal ( troncal ) y trasmitirla a los ramales de los usuarios.
La topología de guirnalda es muy similar pero tiene el cable de bus principal entrando y saliendo de las interfases de los usuarios. Este método requiere aislamiento entre la electrónica de la interfase y el bus mismo para evitar que una falla de la interfase afecte al bus.
Es además necesario en algunas aplicaciones usualmente restringidas a distribución geográfica configurar una red en topología en estrella ( figura 2c) . Esta trabaja un poco como una red en árbol con un tronco muy corto y largas ramas.
Luego del desarrollo de las comunicaciones tipo bus surgió inmediatamente el problema del control del bus. En comunicaciones punto a punto el control es del tipo maestro-esclavo. Esto trabaja bien, ya que si cualquier usuario falla, ninguna comunicación tendrá lugar de todos modos. En una configuración de bus, sin embargo, esto no es siempre el caso. Para instalaciones donde el dispositivo “maestro” está siempre en control y los esclavos son dispositivos “tontos” que sólo se necesitan comunicar con el maestro, esta topología es suficiente. En un ambiente de control distribuido si embargo, donde todos los usuarios necesitan acceso al bus, la falla de una estación maestra y la subsiguiente pérdida de comunicación en la red no es aceptable. Esto lleva al concepto de comunicaciones “par a par” ( Nota de traducción: en inglés peer-to-peer -que se traduciría de par a par- o de punto a punto, y más conocida como P2P ). En este formato, ningún usuario único tiene el control, pero el protocolo está desarrollado para permitir que el control del bus sea compartido entre todos los participantes. En comunicaciones de PLC esto normalmente toma la forma de una red “paso de testigo” ( token passing) o red “acceso múltiple con escucha de portadora y detección de colisión” ('carrier sense multiple access/collision detect' - CSMA/CD). Paso de testigo significa que todos los participantes en la red tienen una lista de todas las direcciones o números de nodos en orden ascendente. En cualquier momento, uno de los participantes tiene un testigo por una cantidad de tiempo igual o menor que un tiempo máximo predefinido.
Durante este tiempo el mismo puede enviar o requerir datos desde cualquier otro nodo. Cuando esto se termina, o su tiempo máximo ha transcurrido, este “hará entrega” del testigo al próximo nodo en la lista de participantes y escuchará, como si fuera un esclavo hasta recibir el testigo nuevamente.
Las redes CSMA/CD, o lo que se conoce comúnmente como Ethernet, trabajan sobre la idea central de que no haya control absoluto sobre la red. Cada usuario de la red detecta por sí mismo si está conectado a la misma. Esto es conocido como escucha de portadora. Una vez que el usuario detecta una portadora en la red, el mismo ve si la red está viva y accede a dicha red, enviando o requiriendo datos de otro usuario. Claramente ya que existe más de un usuario sobre la red, más de un usuario tratará de acceder a la red al mismo tiempo, creando accesos múltiples. Eléctricamente estos dos mensajes se corromperán uno al otro, por lo tanto cuando esto ocurre, ambos usuarios que han transmitido datos detectaran otros datos en la red, distintos a los que cada uno envió, detectando colisiones. Ambos usuarios luego paran de comunicarse por un período aleatorio de tiempo y luego tratan nuevamente.

 

 

 

 

 

 

 

Custom Search

 


 

<< Previous - T - Home - Next >>

A - B - C - D - E - F - G - H - I - J - K - L - M - N - O - P - Q - R - S - T - U - V - W - X - Y - Z


 


 

Last updated: January 29, 2012