Abstract: Methods and systems are provided for determining inter-region and intra-region traffic in one or more networks. To determine the network traffic flowing between a first region and a second region, one or more routers in the first and second regions are identified along with their respective links connecting the routers in the first region to those in second region. A processor located remotely from the routers in the first and second regions then determines discrete used bandwidth samples of the identified links over a period of time. Based on the used bandwidth samples, the processor determines the network traffic flowing between the first and second regions. To determine the network traffic within a region, a sample of hosts each including one or more web servers in the region are selected. The processor identifies the last pairs of identifiable routers on respective routes from the processor to the sample hosts and the respective links connecting the identified last pairs of routers.

Abstract: Methods and systems are provided for determining the geographical regions of hosts in a network. A plurality of sample hosts in the network are preselected. The sample hosts are preselected such that they are located in a plurality of geographical regions that are determinable using existing methods and systems or other means. A plurality of monitoring stations are provided in the network to determine first sets of information associated with each of the sample hosts and second sets of information associated with a host whose geographical region is to be determined. The geographical region of the host is then determined to be the same as the geographical region of those sample hosts whose respective mean of first sets of information has the shortest weighted vector (or euclidian) distance from the second sets of information.

Abstract: Methods and systems are provided for determining the geographical regions of hosts in a network. A plurality of sample hosts in the network are preselected. The sample hosts are preselected such that they are located in a plurality of geographical regions that are determinable using existing methods and systems or other means. A plurality of monitoring stations are provided in the network to determine first sets of information associated with each of the sample hosts and second sets of information associated with a host whose geographical region is to be determined. The geographical region of the host is then determined to be the same as the geographical region of those sample hosts whose respective mean of first sets of information has the shortest weighted vector (or euclidian) distance from the second sets of information.

Abstract: Methods and systems are provided for determining inter-region and intra-region traffic in one or more networks. To determine the network traffic flowing between a first region and a second region, one or more routers in the first and second regions are identified along with their respective links connecting the routers in the first region to those in second region. A processor located remotely from the routers in the first and second regions then determines discrete used bandwidth samples of the identified links over a period of time. Based on the used bandwidth samples, the processor determines the network traffic flowing between the first and second regions. To determine the network traffic within a region, a sample of hosts each including one or more web servers in the region are selected. The processor identifies the last pairs of identifiable routers on respective routes from the processor to the sample hosts and the respective links connecting the identified last pairs of routers.

Abstract: A novel packet switch architecture is disclosed. In the novel packet switch, inputs are dynamically allocated between newly arriving and recirculating packets to make maximum use of available bandwidth while simultaneously resolving output port conflicts.

Abstract: A method for routing data packets correctly and efficiently among a plurality of hosts in a multicast is disclosed. The hosts are connected to a packet switching network comprising a plurality of packet switches interconnected by trunks. A memory associated with the route controller in each switch maintains a table whose entries are tuples of the form <address, trunk identification number, time>. In the case of multicast routing, the address is a multicast address. The trunk identification number is a number which uniquely identifies each of the trunks connected to the switch. Time is a timing parameter which is initially set to a certain predetermined value (designated Maxtime herein) and decremented periodically. When time reaches zero, the associated table entry is cleared. When a host wishes to join a multicast conversation (even if it is the first participant), it simply transmits a packet from itself to the multicast adddress. This packet will be broadcast over the entire spanning tree network.

Abstract: A process is disclosed for effecting transmission over a generally cyclic communication system comprising numerous networks interconnected by gateway pairs. Each gateway of each pair implements a store-and-forward protocol whereby each gateway forwards message packets propagated over its associated network except for any packets that are destined for a device which has previously appeared in the sending address of another packet. In order to utilize this protocol as a basis for transmission, the system is covered with a set of spanning trees that satisfy capacity and reliability requirement. Each spanning tree is assigned a unique identifier and each packet traversing the system is assigned to and conveys the specified spanning tree. Each gateway passes the packet to determine the assigned spanning tree and forwards the packet accordingly. To mitigate system flooding by a newly connected device, the protocol may also incorporate a delay to allow the gateways to learn the location of the new device.