Abstract: A network system that uses a cluster of edge nodes to send and receive multicast traffic is provided. The network system is a network virtualization environment that includes one or more distributed routers, each distributed router implemented by virtualization software running on one or more host machines. The network system also includes a cluster of edge nodes for sending data from the one or more distributed routers to one or more uplink/upstream physical routers outside of a datacenter and for receiving data from the physical routers to the distributed routers. One of the edge nodes is a designated edge node that queries for membership information for one or more multicast groups to be received by at least two edge nodes of the cluster of edge nodes. The cluster of edge nodes forwards multicast traffic to and from the distributed routers according to the received membership information.

Abstract: A network system that uses a cluster of edge nodes to send and receive multicast traffic is provided. The network system is a network virtualization environment that includes one or more distributed routers, each distributed router implemented by virtualization software running on one or more host machines. The network system also includes a cluster of edge nodes for sending data from the one or more distributed routers to one or more uplink/upstream physical routers outside of a datacenter and for receiving data from the physical routers to the distributed routers. One of the edge nodes is a designated edge node that queries for membership information for one or more multicast groups to be received by at least two edge nodes of the cluster of edge nodes. The cluster of edge nodes forwards multicast traffic to and from the distributed routers according to the received membership information.

Abstract: A network system that uses a cluster of edge nodes to send and receive multicast traffic is provided. The network system is a network virtualization environment that includes one or more distributed routers, each distributed router implemented by virtualization software running on one or more host machines. The network system also includes a cluster of edge nodes for sending data from the one or more distributed routers to one or more uplink/upstream physical routers outside of a datacenter and for receiving data from the physical routers to the distributed routers. One of the edge nodes is a designated edge node that queries for membership information for one or more multicast groups to be received by at least two edge nodes of the cluster of edge nodes. The cluster of edge nodes forwards multicast traffic to and from the distributed routers according to the received membership information.

Abstract: A network system that uses a cluster of edge nodes to send and receive multicast traffic is provided. The network system is a network virtualization environment that includes one or more distributed routers, each distributed router implemented by virtualization software running on one or more host machines. The network system also includes a cluster of edge nodes for sending data from the one or more distributed routers to one or more uplink/upstream physical routers outside of a datacenter and for receiving data from the physical routers to the distributed routers. One of the edge nodes is a designated edge node that queries for membership information for one or more multicast groups to be received by at least two edge nodes of the cluster of edge nodes. The cluster of edge nodes forwards multicast traffic to and from the distributed routers according to the received membership information.

Abstract: Some embodiments provide a method for a host computer that executes a set of data compute nodes (DCNs), for identifying a designated router for a multicast group. The method receives a membership request message to join the multicast group from a particular DCN of the set of DCNs, the membership request message including an address associated with the multicast group. The method identifies a logical router gateway from a set of logical router gateways as the designated router for the multicast group, based at least in part on the address associated with the multicast group, where different logical router gateways from the set of logical router gateways are assigned as designated routers for different multicast groups with different associated addresses. The method forwards the membership request message to the identified logical router gateway.

Abstract: The technology disclosed herein enables multicast network traffic to pass an RPF check in a logical router having separated packet handlers. In a particular embodiment, a method includes, in a north/south packet handler of a first logical router, receiving first network traffic from an east/west packet handler of the logical router. The first network traffic is multicast network traffic and the logical router is a first hop router for the first network traffic from a source of the first network traffic. The method further includes identifying an entry for the source in unicast routing information for unicast network traffic. Identifying the entry indicates that the first network traffic passes a reverse path forwarding (RPF) check. In response to the first network traffic passing the RPF check, the method includes transferring the first network traffic from the north/south packet handler to a next hop for the first network traffic indicated by first multicast routing information.

Abstract: For a set of gateway devices at the edge of a logical network, some embodiments provide a method for ensuring that data messages from an external network requiring a stateful service are received at an active gateway device. The method advertises the availability of a set of internet protocol (IP) addresses from standby gateway devices with a higher cost than the cost advertised by an active gateway device. In some embodiments, the advertisement is made using a border gateway protocol. Data messages may be unexpectedly received on a standby node despite the higher advertised cost. This could happen due to asymmetric network failures. The method determines if a stateful service is needed for the data messages received on standby node. Based on the determination, the method forwards the received data message to the active gateway device for the active gateway device to provide the stateful service.

Abstract: The technology disclosed herein enables multicast network traffic to pass an RPF check in a logical router having separated packet handlers. In a particular embodiment, a method includes, in a north/south packet handler of a first logical router, receiving first network traffic from an east/west packet handler of the logical router. The first network traffic is multicast network traffic and the logical router is a first hop router for the first network traffic from a source of the first network traffic. The method further includes identifying an entry for the source in unicast routing information for unicast network traffic. Identifying the entry indicates that the first network traffic passes a reverse path forwarding (RPF) check. In response to the first network traffic passing the RPF check, the method includes transferring the first network traffic from the north/south packet handler to a next hop for the first network traffic indicated by first multicast routing information.

Abstract: Some embodiments provide a method for a host computer that executes a set of data compute nodes (DCNs), for identifying a designated router for a multicast group. The method receives a membership request message to join the multicast group from a particular DCN of the set of DCNs, the membership request message including an address associated with the multicast group. The method identifies a logical router gateway from a set of logical router gateways as the designated router for the multicast group, based at least in part on the address associated with the multicast group, where different logical router gateways from the set of logical router gateways are assigned as designated routers for different multicast groups with different associated addresses. The method forwards the membership request message to the identified logical router gateway.

Abstract: A method for identifying a first hop router (“FHR”) in a distributed virtualized network is presented. In an embodiment, a method comprises receiving a multicast message on an incoming interface. In response to receiving the multicast message, the router determines whether the router is a FHR for the multicast message, i.e., whether, in response to generating and transmitting a hello multicast message, the router does not receive a response on the incoming interface; or whether an interface to a RP for the multicast message is different than the incoming interface. If the router is the FHR for the multicast message, then the router determines an IP address of a RP for the multicast message. The router also encapsulates the multicast message into a unicast message, includes in the unicast message the IP address of the RP as a destination address, and transmits the unicast message to the RP.

Abstract: Some embodiments provide a method for a computing device that implements a centralized logical routing component of a logical network that handles data traffic between the logical network and an external physical network. The method receives a first routing protocol message from a machine that implements a route server for a distributed logical routing component of the logical network. The first routing protocol message includes (i) a parameter identifying the machine as a route server and (ii) a set of logical network addresses. The method receives a second routing protocol message from the route server that specifies a next hop address for the set of logical network addresses. The method adds a route specifying the next hop address for the set of logical network addresses to a routing table of the centralized logical routing component. The added route has a higher priority than routes received from the external physical network.

Abstract: Some embodiments provide a method for a computing device that implements a centralized logical routing component of a logical network that handles data traffic between the logical network and an external physical network. The method receives a first routing protocol message from a machine that implements a route server for a distributed logical routing component of the logical network. The first routing protocol message includes (i) a parameter identifying the machine as a route server and (ii) a set of logical network addresses. The method receives a second routing protocol message from the route server that specifies a next hop address for the set of logical network addresses. The method adds a route specifying the next hop address for the set of logical network addresses to a routing table of the centralized logical routing component. The added route has a higher priority than routes received from the external physical network.

Abstract: A method for identifying a first hop router (“FHR”) in a distributed virtualized network is presented. In an embodiment, a method comprises receiving a multicast message on an incoming interface. In response to receiving the multicast message, the router determines whether the router is a FHR for the multicast message, i.e., whether, in response to generating and transmitting a hello multicast message, the router does not receive a response on the incoming interface; or whether an interface to a RP for the multicast message is different than the incoming interface. If the router is the FHR for the multicast message, then the router determines an IP address of a RP for the multicast message. The router also encapsulates the multicast message into a unicast message, includes in the unicast message the IP address of the RP as a destination address, and transmits the unicast message to the RP.

Abstract: For a set of gateway devices at the edge of a logical network, some embodiments provide a method for ensuring that data messages from an external network requiring a stateful service are received at an active gateway device. The method advertises the availability of a set of internet protocol (IP) addresses from standby gateway devices with a higher cost than the cost advertised by an active gateway device. In some embodiments, the advertisement is made using a border gateway protocol. Data messages may be unexpectedly received on a standby node despite the higher advertised cost. This could happen due to asymmetric network failures. The method determines if a stateful service is needed for the data messages received on standby node. Based on the determination, the method forwards the received data message to the active gateway device for the active gateway device to provide the stateful service.