Multiplex communications – Pathfinding or routing – Switching a message which includes an address header
Reexamination Certificate
2000-05-12
2001-07-17
Olms, Douglas (Department: 2661)
Multiplex communications
Pathfinding or routing
Switching a message which includes an address header
C370S469000
Reexamination Certificate
active
06262988
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates in general to computer communication networks, and in particular, to systems and methods for transforming an existing router-based network to a switch-based network.
Current internet-based networks are built on TCP/IP (Transmission Control Protocol/Internet Protocol) and related standards. TCP provides application services but the network infrastructure is implemented with the IP portion of TCP/IP. Some of the more notable features of the Internet protocols are that they:
a. can be implemented on many different underlying hardware technologies from different vendors;
b. hide the underlying hardware;
c. allow autonomous networks to be built and managed independently;
d. allow autonomous networks to be easily interconnected with routers to build larger networks which can be further interconnected to build very large networks like the global Internet;
e. are open standards.
The flexibility, scalability and openness of the Internet protocols has led to their widespread use even inside private networks, which are commonly known as Intranets. The IP protocol essentially overlays the interconnected networks which are implemented with different hardware technologies over widespread geographic distances and make it all look like a single homogenous network.
Most organizations design, implement and manage their own autonomous IP networks or Intranets and connect them using routers to other networks or to the global Internet. Large organizations generally further divide their IP network into subnetworks using subnet technology for distributed administration and traffic control. Again routers are used to interconnect these subnetworks.
Most networks on customer premises are built with IEEE 802 Local Area Network (LAN) hardware technologies (e.g., 10 Mbps Ethernet and 16 Mbps token-ring). These are generally referred to as layer-1/2 technologies. Layer-1 refers to the wiring and signalling characteristics of the network and layer-2 refers to the data link protocols, also called Media Access Control (MAC) and Logical Link Control (LLC) protocols. Normally, each LAN is defined as a separate IP subnetwork and then interconnected using routers.
Geographically distributed LANs are normally interconnected with routers containing a LAN connection and a wide area network (WAN) connection. Examples of WAN connections are leased lines, frame-relay and satellite links. The LANs and WANs are then overlayed with a layer-3 protocol like IP. The IP protocol hides the underlying physical networks and gives a single homogenous logical view of the entire network to the attached hosts.
Recently, LANs have undergone a dramatic transformation due to the introduction of layer-2 switches for LANs (also known as LAN switches). Shared media LAN hubs like 10 Mbps Ethernet and 16 Mbps token-ring are being replaced by Ethernet and token-ring switches. Switches are fast, inexpensive and can provide full and dedicated bandwidth to each attached user. For example, a single 10 Mbps Ethernet hub being shared by 20 users can now become a 200 Mbps switched network (10 Mbps×20 ports). Some switches even allow attached hosts to transmit and receive simultaneously at full LAN speed which effectively doubles the throughput on each switch port, for example, each Ethernet switch port could be 20 Mbps in full duplex mode.
To build a larger switched network, users can connect LAN switches to higher speed backbone switches using high speed interfaces like Fast Ethernet (100 Mbps) or ATM (155 or 622 Mbps). Even with this high speed switched network, routers are still needed to route packets between the logical IP subnetworks. Routers are therefore still being used to route between subnets that are implemented on switches instead of LAN hubs.
The switching phenomenon has tremendously increased much needed network capacity. While traditional routers could easily handle 10 Mbps of traffic coming out of LAN hubs, they cannot handle the hundreds of Mbps of traffic that switches can generate. One partial solution to this problem involves connecting routers to switches with high speed interfaces (e.g. 155 Mbps). But the internal routing capacity of the router must also be increased to forward the high volume of traffic coming in to the router. This is very expensive. Moreover, unlike switches, routers still introduce significant latency during data transfer. Another similar solution is to remove the router from the network and embed it inside switches. This might reduce the cost of the router and hide it from the user, but it still suffers from the same throughput and latency limitations as the previous solution. Yet another solution is to create a flat network, i.e., single IP subnet or IP network. This will let hosts communicate with each other without an intermediate router. However, this defeats the original reasons for subnetting. In this scenario, broadcast floods which are common on LANs, but are normally contained within an IP subnet, will be propagated all over a large switched network and seriously degrade network performance.
In the era of high-speed switching, routers have become a bottleneck in networks. When compared to switches, routers are slow, expensive, difficult to administer and restrict host mobility between IP networks. But they are still needed to interconnect layer-3 IP networks and IP subnets. Thus there is a need for a system and method that enables network administrators and end users to get the benefits of high-speed switching and the benefits of IP subnetting without the use of routers. This invention specifically solves the problem of creating subnets in a switched network without IP routers; separating subnet broadcast traffic but allowing hosts on different subnetworks to communicate directly at full switching speed without involving a router. Furthermore, the invention greatly simplifies the subnetting of an IP network by allowing centralized control of subnets.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide an improved data communication and networking system.
It is another object of the present invention to provide a method for distributing IP network layer routing functions to a switch device.
It is a further object of the present invention to provide a method that enables host devices on different subnetworks to communicate at full switching speed without the use of router devices.
It is a still further object of the present invention to provide a method that enables centralized control of subnets in an IP network.
The foregoing objects are achieved by the system and method of the present invention in which hosts (end stations) are moved from router-attached legacy LANs to a switched network composed of interconnected layer-2 switches. A large IP subnetwork is established by setting the subnet mask on the IP hosts in the switched network to a smaller value. Enlarging the IP subnetwork enables all hosts to communicate directly with each other without an intermediate router. Subnets are then implemented in the switches within the IP subnetwork, effectively moving subnet control from the hosts to the centrally controlled switches. The switch CPU learns subnet information by analyzing layer-2 multicast packets, specifically the layer-3 header to determine if the packet is an IP ARP request packet. The switch builds a subnet to port mapping table based on packets received from each source host and selectively forwards the multicast packet to all ports on which the destination subnet is active. Once the destination host responds to the IP ARP request, a direct communication path is established in the switch between the port on which the source host is attached and the port on which the destination host is attached.
REFERENCES:
patent: 4821265 (1989-04-01), Albal et al.
patent: 4864559 (1989-09-01), Perlman
patent: 4893302 (1990-01-01), Hemmady et al.
patent: 4933937 (1990-06-01), Konishi
patent: 4942518 (1990-07-01), Weatherford et al.
patent: 5018137 (1991-05-01), Backes et al.
patent: 5027350 (1991-06-
Cesari and McKenna LLP
Cisco Technology Inc.
Olms Douglas
Vanderpuye Ken
LandOfFree
Method and system for subnetting in a switched IP network does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for subnetting in a switched IP network, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for subnetting in a switched IP network will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2466325