Telefonaktiebolaget Lm Ericsson (Publ)

Ericsson Ab

Google Inc.

Sriganesh Kini

Wenhu Lu

Rishi Mehta

Attila Takács

Maurilio Cometto

Manoj Sharma

Srikanth Narayanan

Marc Rapoport

Mohammad Y Hajjat

Stefano Ruffini

András Császár

Balázs Peter Gerö

Gregory Mirsky

Eric Ward Gray

Hua Autumn Liu

Andrew Wilkinson

David Sinicrope

Hector Avalos

Faisal Andre Khan

Mate Ferenczy

Panagiotis Saltsidis

János Farkas

Balázs Varga

Peter Arberg

Gabor Sandor Enyedi

Albert Jining Tian

Pramodh D'Souza

Shahryar Khan

Venkatesan Pradeep

Bhaskara Chinni

Michael Chen

Máté Ferenczy

Gábor Sandor Enyedi

Attila Zoltan Takacs

Autumn Liu

In one embodiment, a system includes a plurality of first host machines implementing a public-cloud computing environment, wherein at least one of the first host machines comprises at least one public-cloud virtual machine (VM) that performs network address translation; and a plurality of second host machines implementing a private-cloud computing environment, wherein at least one of the second host machines comprises one or more private-cloud virtual machines, wherein the public-cloud VM is configured to receive, via a network tunnel from the private-cloud VM, one or more first packets to be sent to a public Internet Protocol (IP) address of a public network host, translate, using a NAT mapping, a source address of each first packet from a private IP address of the private-cloud VM to an IP address of the public-cloud VM, and send the first packet to the IP address of the public-cloud VM.

Virtual private cloud network switching

Network architecture for cloud computing environments

A method of operation of a Multiprotocol Label Switching network involves, in an active node of the network, receiving a first data packet from a source node and forwarding the first data packet to a destination node. At the same time, the active node measures a residence time of the first data packet in the active node. The active node then sends a further data packet containing residence time information.

Transmitting residence time information in a network

According to one aspect of the present disclosure, a method of operating a network node of a communication network is provided, the network node being connected to a downstream network node, and to a non-downstream loop free alternate (LFA) network node comprising a load balancing identifier indicating whether it is allowed to load balance the data packet via a non-downstream LFA network node or not; changing, if the load balancing identifier indicates that it is allowed to load balance the data packet via a non-downstream LFA network node, and if the network node decides to load balance the data packet via the non-downstream LFA network node, the load balancing identifier of the data packet such that it indicates that a further load balancing of the data packet via a further non-downstream LFA network node is not allowed; and forwarding the thus modified data packet to the non-downstream LFA network node.

Technique of operating a network node for load balancing

Exemplary methods for creating label stacks include creating and sending a first SR label stack for a data packet, wherein the first SR label stack causes the data packet to be forwarded through the SR network using a first set of links, and wherein the first SR label stack includes a first service label that identifies a first service to be applied to the data packet by a second network device. In one embodiment, the methods include creating and sending a second SR label stack for an operations administration and maintenance (OAM) packet, wherein the second SR label stack causes the OAM packet to be forwarded through the SR network using the first set of links, and wherein the second SR label stack prevents the second network device from applying the first service to the OAM packet.

Label stack encoding and processing to enable OAM procedures for service segments in segment routed (SR) networks

Exemplary methods for allocating multiple local sub-blocks (LsBs) of segment identifiers (IDs) include allocating a first set of LsBs, each LsB associated with a start index and a range, wherein the start index indicates a starting SID of a corresponding LsB and the range indicates a total number of SIDs included in the corresponding LsB, each LsB identified by a sub-block number. In one embodiment, the methods include mapping a plurality of segment routing global block (SRGB) indexes to a plurality of corresponding SIDs of the first set of LsBs, the mapping performed based on the SRGB indexes, start indexes of each LsB, and ranges of SIDs of each LsB. In one embodiment, the methods include advertising the first set of LsBs by transmitting a first advertisement message that includes the start indexes, ranges of SIDs, and sub-block numbers of all LsBs of the first set of LsBs.

Scalable segment identifier allocation in segment routing

An edge router runs a Multipath Transmission Control Protocol (MPTCP) proxy to allow for a host that implements TCP (Transmission Control Protocol) to operate normally yet reap the benefits of an MPTCP connection. An upgrade of a TCPIP stack on the host is not necessary. The edge router demultiplexes packets received from the host over a TCP connection to an MPTCP connection and multiplexes packets sent to the host over an MPTCP connection to a TCP connection. As a result, higher throughput of packet communication can be realized, for example, for improved video support.

Multipath transmission control protocol proxy

A method and apparatus for using entropy labels in segment routed networks is disclosed. A single ELI and a single EL are provided for a label stack. The ELI and EL are inserted directly below a top-most label in the label stack. A method and apparatus for using entropy labels in segment routed networks. A top-most label of a label stack of a packet is examined. A NHLFE for the packet is determined. A determination is made as to whether or not an ELI is below the top-most label. An EL is re-used when the ELI is determined to be present . The ELI and the re-used EL are inserted directly below a label associated with the NHLFE in a new label stack.

Method and apparatus for using entropy labels in segment routed networks

There is provided a method for use by a router in a communications network. Forwarding information is maintained (S) which specifies the next hop node for each of a plurality of possible destination nodes. Update information is maintained (S) which specifies how, if at all, the next hop nodes specified in the forwarding information are to be updated for a plurality of possible network resource failures. Known failure information is maintained (S) which relates to a known network resource failure or which specifies that there is no known network resource failure. A failure notification is received (S) relating to a network resource failure. In response to receipt of the failure notification, it is determined (S) from the failure notification how, if at all, the known failure information is to be updated, and the known failure information is updated, if required, based on the determination (S). A communications packet is received (S). Various steps (Sto S) are carried out in response to receipt of the communications packet. The destination node for the received packet is determined (S). It is determined (S) whether the next hop node specified by the forwarding information for the determined destination node is potentially invalid in view of the known failure information. If so, then the next hop node for the determined destination node is updated if necessary in the forwarding information using the update information and the known failure information (S). The next hop node is determined for the received packet using the forwarding information and the determined destination node (S). The received packet is forwarded towards the determined next hop node (S).

Method and apparatus for handling network resource failures in a router

A primary border node (BN) and a standby BN are provided for internetworking two network domains, such that connectivity between the two network domains is maintained when a failure occurs in one of the network domains. The two network domains include an access network that implements MPLS-TP and a core network that implements IP, MPLS, or a combination of both. The primary BN establishes a tunnel from itself to the standby BN, and re-directs network data traffic from itself to the standby BN via the tunnel when it detects that an access node has switched connection from the primary VN to the standby BN. The primary BN also monitors its connections to the core network, and signals access nodes to switch to the standby BN if a failure is detected in these connections.

Internetworking and failure recovery in unified MPLS and IP networks

A network element that has a broadcast interface to a broadcast network becoming operational determines whether at least one alternate path exists to the broadcast network. The broadcast interface is to carry traffic on a label switched path. If an alternate path exists, the broadcast interface will not be advertised to the broadcast network until label distribution protocol (LDP) is operational with all neighbors on the broadcast interface.

LDP IGP synchronization for broadcast networks

IP addresses are efficiently assigned and advertised across the boundary of two network domains. The two network domains include an access network that implements MPLS-TP and a core network that implements IP, MPLS, or a combination of both. A primary BN at the domain boundary receives an assignment of an IP prefix for a VPN that connects a set of CEs to one or more SNs. The IP prefix represents a pool of IP addresses. The primary BN assigns only one IP address to both itself and a standby BN for the VPN to minimize the use of assigned IP addresses. The primary BN assigns the other IP addresses to the CEs in the VPN. The primary BN then advertises the IP prefix into the core network to enable routing of network data traffic across the network domain boundary.

Internetworking and IP address management in unified MPLS and IP networks

An hierarchical LSP is established to transport packets belonging to a FEC attached to an egress LSR and includes an egress LSR LSP that is common for each of the FECs attached to the egress LSR and forms a path from the ingress LSR through intermediate LSR(s) to the egress LSR. The egress LSR LSP is used when label switching packets destined for the FECs attached to the egress LSR. The hierarchical LSP also includes a unique FEC LSP for each FEC that is used by the egress LSR to identify and forward packets to that FEC. Responsive to a topology change that changes a next-hop of the ingress LSR to reach the egress LSR, the ingress LSR modifies an entry in a forwarding structure to change the next-hop for the egress LSR LSP and does not modify substantially any forwarding structure entities for the FEC LSPs.

Method and apparatus to improve LDP convergence using hierarchical label stacking

Growing community of inventors

Fremont, CA, United States of America

Sriganesh Kini