Ring configuration for network switches

Details Ring configuration for network switches Ring configuration for network switches

1999-09-30

2003-09-16

Rao, Seema S. (Department: 2666)

Multiplex communications

Pathfinding or routing

Switching a message which includes an address header

C370S393000

Reexamination Certificate

active

06621818

ABSTRACT:

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
Not applicable.
BACKGROUND OF THE INVENTION
This invention is in the field of data communications networks, and is more specifically directed to switch architectures for packet-based communications, such as Ethernet.
In recent years, the deployment of local-area networks (LANs) of personal computers and workstations in business and educational enterprises has become widespread. These modem networks have provided significant productivity gains, not only by providing individuals with the ability to electronically communicate with one another, but also by enabling shared access of data bases and documents stored on a central server and on the networked client workstations. With the ever-improving performance of modem communications and network elements, both large and small enterprises now utilize networks among their personal computers and workstations. Conventional LANs can also be interconnected with one another by way of bridges or routers into a Wide Area Network (WAN). Typically, each network segment includes a number of client workstations and a bridge or router; interconnection of the bridges and routers in a ring or tree fashion permits communication between clients located on different segments of the WAN. Further, the availability of Internet access to such networked computers extends desktop access beyond local, proprietary, data sources to worldwide public and secure data source sites.
The rapid improvement in functionality and performance of modern personal computers and workstations has, in turn, created a desire for higher data rate network communications, as the quantity of data that can be processed by a client workstation has increased accordingly. Similarly, high bandwidth communications can enable the exchange of data types such as 3-D images, full-motion video, sound files, and the like, each of which are represented by large blocks of data. Another trend is from half-duplex communications over the network to full duplex network communications, permitting the simultaneous transmission and receipt of data, and thus effectively doubling the bandwidth while eliminating bandwidth loss due to collisions.
Packet-based data communications is a common approach to data communications in LANs and WANs, particularly in high data rate connections over busy-networks. As opposed to connection-oriented networks, in which a dedicated connection is established between the source and destination nodes in the network, packet switching divides each message (i.e., data block to be communicated) into small packets. Each packet carries identifiers of the source and destination nodes of its corresponding message, along with an identifier of its location in the sequence of packets that comprise the overall message. This permits the multiplexed communication of packets from multiple messages, between various source and destination nodes, over the network. The identifiers in each packet allows the destination nodes to retain those packets addressed thereto, and to resequence the received packets into the communicated message. As a result, packet-switched networks permit more nodes to communicate with one another at a given time, with each communication utilizing a portion of the overall network bandwidth without blocking other channels. Examples of conventional packet-based networks include Ethernet, token ring, and FDDI.
Traditionally, many Ethernet LANs have been constructed so as to use repeaters, or hubs, to which each client workstation connects; the hubs provide flexibility in the addition and deletion of clients from the network, and also enables the use of twisted-pair wire for network communications. Most existing hubs and repeaters provide only half-duplex communications, however, and as such some amount of conflict among clients on the same network “segment” (i.e., associated with the same hub) may arise.
Recently, LAN switches have become available for use in Ethernet networks, as well as in FDDI and token ring LANs. A LAN switch has multiple ports, each of which may connect to a client workstation (or a network hub, as the case may be), or provide an uplink to another switch or a server. A LAN switch differs from a bridge or hub, in that the switch allows simultaneous switching of packets between multiple pairs of its ports. As a result, the switch provides higher bandwidth for its clients, in the aggregate, with reduced latency. In addition, ports of an Ethernet switch can be connected both to a network segment via a hub, and to a client workstation over a dedicated segment.
FIG. 1
illustrates a conventional switched Ethernet network of client workstations C
0
through C
7
with server SVR. In this example, client workstations C
0
through C
3
comprise a network segment attached to half-duplex hub H. As such, only one of client workstations C
0
through C
3
can be transmitting or receiving at any given time. Hub H is connected to one port of Ethernet switch SW. Client workstations C
4
through C
6
are connected to other ports of switch SW, by way of dedicated full-duplex connections. Switch SW is also connected to server SVR by way of a full-duplex connection. The use of full-duplex-capable switch SW permits client workstations C
4
through C
6
to communicate with one another, with server SVR, or with one of client locations C
0
through C
3
via hub H, in multiple pairs, thus improving the overall bandwidth of the network. As such, Ethernet switches such as switch SW in
FIG. 1
are becoming popular in high-traffic network applications.
Heretofore, however, the number of ports supportable by conventional switches have been relatively limited. For example, conventional Ethernet switches can support on the order of twenty-four full-duplex 10-Mbit/sec ports and three full-duplex 100-Mbit/ sec ports. With the advent of gigabit Ethernet communications, conventional Ethernet switches for supporting eight full-duplex 100-Mbit/sec ports and one 1000-Mbit/ sec (“gigabit”) port are now entering the marketplace.
However, network designers and users are desirous of supporting large numbers of ports, particularly if dedicated full-duplex connections of client workstations (e.g., client workstations C
4
through C
6
of
FIG. 1
) are desired. Because of the limitation of the number of ports supported by conventional switch architectures, these conventional designs must insert additional levels of hierarchy into the switching architecture to provide the desired number of ports for a given network. Not only does the management of the network become complex in such conventional arrangements, but the reconfiguration of ports in the network is made quite cumbersome.
BRIEF SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a network switch that can be flexibly connected with other network switches to support a high port density from a single switch location.
It is a further object of the present invention to provide such a network switch system which includes high-speed uplink capability.
It is a further object of the present invention to provide such a network switch system in which additional ports may be easily added.
It is a further object of the present invention to provide such a network switch system in which new communications from a port is readily established.
Other objects and advantages of the present invention will be apparent to those of ordinary skill in the art having reference to the following specification together with its drawings.
The present invention may be implemented by way of a plurality of network switch devices that are configurable into a ring arrangement. The ring of network switches are capable of “learning” address information for packet communications, such address information being added as a tag to each packet by the receiving ring switch. Communications among ports on the switch devices in the ring may then be handled by each switch determining whether the packet was sourced by itself (in which case the packet is discarded), and if not, for determini

Affiliated with

Also associated with

Rating

Ring configuration for network switches does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Ring configuration for network switches, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Ring configuration for network switches will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3040628