摘要:
In the packet scheduling device, the output class selection part is made to conduct the selective output of priority class, based on the weight count value calculated from a weight value corresponding to minimum guarantee bandwidth for each priority class and the amount of packet accumulated in each queue. The output class selection part operates so that if the weight count value is “0” or more, then the packet length of variable-length packet data to be output is subtracted from the current weight count value to give a renewed weight count value. On the other hand, if the weight count value is less than “0”, until reaching a predetermined repeat count (Cmax−1), the weight value of each priority class is added to the weight count value of all priority classes, and, when reaching the repeat count (Cmax−1), the weight value is added to the weight count value of all priority classes or replaced by half the weight value. Thereby, the weight count value is made to be more than “0”, and the values thus renewed are subject to the rule of round robin so as to select the priority class.
摘要:
In a high-speed routing control system, plural path candidates each leading to a destination node are selected on the basis of physical connection information of links connecting respective nodes in a path candidate selection unit 11, and a path candidate containing no congestion-occurring link is specified as the optimum path from the plural path candidates selected by the path candidate selection unit 11 in an optimizing unit 12. Accordingly, even when the destination node is nearer to the self node, the optimum path selection can be performed, and a switching operation to a bypass path can be performed at high speed. In addition, the traffic amount based on topology information can be suppressed, and the large-scaling of the network can be supported.
摘要:
In an ATM connectionless communication system for delivering connectionless datagram in the form of sessions through an ATM network, each one of multiple edge units is connected to a user network where quality-of-service (QoS) sessions and No-QoS sessions are mixed. Each of the edge units accommodates an IPv4 packet from the user network into an intermediate frame and accommodates the intermediate frame into ATM cells. The edge units are connected to core units by permanent virtual paths to form a connectionless network. Each of the edge units has a session supervising section for determining whether the IPv4 packet is a QoS session or a No-QoS session, to allocate a control flow number to the IPv4 packet when the IP4 packet is a QoS session.
摘要:
A transfer destination determining apparatus and method performing a load distribution in a network, where a plurality of transfer destination subjects are present as a transfer destination of a datagram, while the same collected flows are transferred to the same transfer destinations, the transfer destinations are allocated with respect to each of the collected flows. This apparatus has an extracting part extracting address and flow discrimination information, in the IP datagram; route solving part determining a transfer path based on the address information. When the determined path is a single path, the transfer destination is uniquely determined and the path is outputted to a post-staged apparatus, whereas when the path is a multi-path having a plurality of transfer destination subjects a multi-path number is outputted to a flow managing means which determines a transfer path corresponding, to a collected flow based on the received flow discrimination information and the multi-path number.
摘要:
A loose source routing method is provided to transfer an IP packet from a transmission source gateway to a transfer destination gateway by way of ATM nodes, which are freely designated. At the transmission source gateway, the IP packet given from a user LAN is dissolved into ATM cells containing a BOM cell whose destination address designates the transfer destination gateway. In addition, at least one pseudo BOM cell whose destination address designates an ATM node in the ATM network is added and is located at a top place of a cell stream constructed by the dissolved ATM cells. Thus, the cell stream is transferred from the transmission source gateway to the designated ATM node in accordance with the destination address of the pseudo BOM cell. The designated ATM node discards the pseudo BOM cell so that the original BOM cell is now located at the top place of the cell stream. Then, the cell stream is transferred from the ATM node to the transfer destination gateway in accordance with the destination address of the BOM cell. The transfer destination gateway assembles the ATM cells into the IP packet, which is then sent to a user LAN.
摘要:
A lightwave network data communications system having such an architecture that routing operation within a lightwave network is simplified, and a high-speed transfer process is attained in a large-scale basic network utilizing a wavelength division multiplexing (WDM) technology which accommodates internet traffics from a plurality of subscriber networks. The lightwave network data communications system includes a unit for giving a lightwave router address to an edge router and a core router in the lightwave network, a unit for resolving a destination lightwave router address and an aggregated flow identifier about an IP packet received from a subscriber, and encapsulating the IP packet into a lightwave adaptation frame with these pieces of data contained in header information in the edge router, a unit for encapsulating a plurality of packets each having a length under a fixed length into a superframe, a unit for executing a share-ride scheme in which the packets of a plurality of IP flows can be shared as a superframe, and a unit for monitoring the traffic on the superframe basis and regulating an excessive traffic.
摘要:
The present invention provides a gateway unit which comprises a detecting unit for detecting the congestion from an RM cell from the ATM cell, for discriminating a TCP data gram (TCPDG) from a packet received during the congestion and for acquiring a sequence number of the TCPDG, a sending unit for sending an ACK having a window-controlled value to a transmitting terminal so as to limit a data transmission, and a discarding unit for discarding the packet from the transmitting terminal during the congestion after the confirmation of packet. During the congestion in the ATM network, a pseudo acknowledgment is sent from the gateway unit so as to indicate that the receiving terminal cannot receive data. This restrains the transmission of TCPDG from the transmitting terminal.
摘要:
In a connectionless communication system, a number of interworking units are provided for accommodating local area networks. At the internetworking units, a packet from the local area networks are segmented into cells for transmission to an ATM network. The cells contain a connectionless address designating a designation local area network, a virtual path identifier (VPI) designating a permanent virtual connection, and a virtual channel identifier (VCI) identifying the packet. The ATM network includes a number of cross-connect nodes interconnected by communication links. The cross-connect nodes establish a plurality of permanent virtual connections between the interworking units and the nodes and between the nodes in pairs, receive the cells from the interworking units and route the cells having VCIs of the same value through the permanent virtual connections to a destination interworking unit according to the connectionless address and the VPI contained in the cells. The destination interworking unit assembles the cells having VCIs of the same value into a packet for routine the packet to the destination local area network according to the connectionless address transmitted by the packet.
摘要:
A layer 1 network frame is disclosed that includes data of a layer 2 frame. A header of the layer 1 frame header includes: a packet length field to indicate a size of a payload portion of the layer 1 frame, a priority field to indicate a priority of the layer 1 frame, a protocol field to identify a protocol of the data in the layer 2 frame, a frame mode field to indicate a correspondence between the layer 1 frame and the layer 2 frame included within the payload, a stuff field to indicate whether stuff data is contained in the layer 1 frame, and a cyclic redundancy check (CRC) field to indicate a CRC result.
摘要:
A layer 1 frame, capable of accommodating data of any protocol selected from an STM (Synchronous Transfer Mode) signal, ATM (Asynchronous Transfer Mode) cells, a primary IP (Internet Protocol) packet and a best effort IP packet in a common frame format, is constructed by an edge node and transmitted to a data transfer system composed of edge nodes, core nodes, etc. The layer 1 frame includes a layer 1 frame header, a layer 1 frame payload and a payload CRC (Cyclic Redundancy Check) field. The layer 1 frame header includes a “Packet Length” identifier, a “Priority” identifier, a “Protocol” identifier indicating the type of the data, a “Frame Mode” identifier, a “Stuff” identifier indicating whether or not stuff data is contained in the layer 1 frame, and a “Header CRC” identifier. The layer 1 frame payload contains a layer 2 frame having a layer 2 frame header and a layer 2 frame payload in which the data is packed. STM layer 1 frames containing the STM signals are transferred with the highest priority at fixed intervals (125 μsec). Best effort IP layer 1 frames containing the best effort IP packets are transferred with the lowest priority. Partitioning of the best effort IP layer 1 frame, stuffing, dummy frames, etc. are employed for implementing the periodical transmission of the STM layer 1 frames.