Multiplex communications – Pathfinding or routing – Switching a message which includes an address header
Reexamination Certificate
1998-08-21
2001-12-18
Chin, Wellington (Department: 2664)
Multiplex communications
Pathfinding or routing
Switching a message which includes an address header
C370S404000, C370S405000, C370S406000
Reexamination Certificate
active
06331985
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention relates generally to the field of communications and, in particular, to circuits and methods for a telecommunication network with variable address learning, switching and routing over ring networks without the use of a token or encapsulation.
BACKGROUND OF THE INVENTION
Computer networks have become commonplace in large and small businesses, universities, and other organizations. Such networks allow a number of users to share data and resources, such as data storage systems, file servers, switches, routers, printers, modems, and other peripherals. There are three basic types of protocols for transmitting data in these networks exemplified by: Ethernet, Token Ring or Fiber Distributed Data Interface (FDDI) and encapsulation. Each of these network protocols have advantages and disadvantages which are presented briefly here in Section I of the Background of the Invention in order to better understand the teachings of the present invention. Further, conventional Ethernet protocols are described in Section II. Finally, aspects of conventional Ethernet switches that limit the ability of such switches to be configured in a ring network are described in part III.
I. Network Protocols
Ethernet, described more fully below, is basically a broadcast protocol. Its main advantage is its simplicity. This allows Ethernet to be implemented with less costly hardware and software. The main drawback with conventional Ethernet is that there are significant limitations on the physical distance that the network can cover. Further, conventionally, the bandwidth of an Ethernet network is greater closer to the center of the network.
Token Ring and FDDI, as described in computer industry standards IEEE-802.5 and ANSI XT3.9, respectively, provide the distinct advantage that data can be transmitted over much greater distances when compared to conventional Ethernet. Further, Token Ring and FDDI provide virtually equal bandwidth throughout the network. The main disadvantage of Token Ring and FDDI is their complex protocols which make the equipment significantly more costly than Ethernet equipment. These complex protocols are necessitated by the manner in which packets are transmitted in Token Ring and FDDI networks. The protocols are dependent on the use of a “token.” This token is passed around the network such that only the network entity, or entities, in possession of a token can transmit data on the network. When a token is corrupted, the network elements are unable to transmit packets on the network. In some instances, this can last for several seconds. To compensate for this problem, complex protocols have been developed that allow the network elements to determine when a token has been lost and to create a new token. The enormous amount of logic circuitry required to implement these protocols makes Token Ring and FDDI networks expensive to implement and maintain when compared to Ethernet networks.
Finally, encapsulation protocols have been developed to allow Ethernet packets to be transmitted over longer distances. In such protocols, the entire Ethernet packet is placed within another type of packet with its own header including additional addressing information, protocol information, etc. These protocols typically also suffer from the problem that they may require special higher level protocol information to be included in the data field of the Ethernet packets for purposes of directing routers in the network, thereby limiting the types of data packets that can be handled and putting significant processing burden on both the network devices generating the packets and the routers used to transmit and receive the packets between the various Ethernet networks. These additional protocol elements and restrictions typically require expensive hardware and software to be added to an otherwise inexpensive Ethernet network. Further, such protocols typically require the use of manually created address tables for the routers.
II. Ethernet
Ethernet has become a common protocol for local area networks and is in wide use due to the advantages described above. For purposes of this specification, the term “Ethernet” includes the entire class of Carrier Sense Multiple Access/Collision Detection (CSMA/CD) protocols covered by the family of computer industry standards known variously as IEEE-802.3 and ISO 8802/3. This includes but is not limited to 1-megabit Ethernet, known as “StarLAN”, 10-megabit Ethernet, 100-Megabit Ethernet, known as “Fast Ethernet”, 1-gigabit Ethernet, known as “Gigabit Ethernet” and any future CSMA/CD protocols at any other data rates.
Originally, Ethernet was designed as a half-duplex broadcast system with a data bus that carries data packets at a rate of approximately 10 Megabits per second to and from terminals. Each terminal connected to an Ethernet can either transmit to or receive from all other terminals on the network (“Multiple Access”, the “MA” in CSMA/CD), but, in the original Ethernet, may not transmit and receive at the same time. Further, Ethernet was designed as a network with no central control over which terminal has access to the data bus at a given time. Ethernet was based on the probabilistic principle that two terminals rarely will transmit at the same time and that each terminal first “listens” to the bus to see if another terminal is already transmitting (“Carrier Sense”, the “CS” in CSMA/CD). This is in contrast to Token Ring and FDDI systems where deterministic control is administered by Tokens and ATM (Asynchronous Transfer Mode) networks and routers where central deterministic control is handled by either an ATM switch or the routers through special inter-router protocols.
When two terminals attempt to transmit at the same time, there is a collision. The terminals that are involved detect the collision (“Collision Detection”, the “CD” in CSMA/CD) by monitoring the data bus for a collision signal or corrupted data packets on the bus after a transmission. In order for all the terminals that have transmitted to realize there is a collision, all the terminals must receive all the packets and collision signals involved. Therefore, the network cannot be any larger than half the distance that the smallest packet will cover from start to finish. At 10 Megabits per second, a 64-byte packet, the minimum Ethernet packet, takes 51.2 microseconds from start to finish. Therefore, a local area network can be no larger than the distance a packet will travel in 25.6 microseconds, including any propagation delays from equipment in the network. At 100 Megabits per second, a 64-byte packet takes 5.12 microseconds from start to finish. Therefore, a local area network can be no larger than the distance the packet will travel in 2.56 microseconds, including any propagation delays from the equipment. When the collision is detected, each of the terminals will wait a random amount of time before attempting to retransmit its packet so as to avoid further collisions on the network. This is in contrast to Token Ring, FDDI, ATM and routers, which because of the centralized deterministic control administered through the use of Tokens and additional protocols do not allow collisions and can therefore transmit data over much longer distances.
The Ethernet, as do all the other network protocols, transmits in packets. These data packets include a source address, a destination address, the data being transmitted, and a series of data integrity bits commonly referred to as a cyclical redundancy check or CRC. The source address identifies the device that originated the packet and the destination address identifies the device to which the packet is to be transmitted over the network.
Full-duplex Ethernet was developed more recently to eliminate the timing restrictions of half-duplex Ethernet by having separate transmit and receive channels between two terminals. In this manner, since the transmit channel is only transmitting to a single receiver, which never receives transmissions from any other transmitter, there can never be a collision. Full-duplex
ADC Telecommunications Inc.
Chin Wellington
Fogg Slifer & Polglaze, P.A.
Pham Brenda H.
LandOfFree
Telecommunication network with variable address learning,... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Telecommunication network with variable address learning,..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Telecommunication network with variable address learning,... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2578869