摘要:
In one embodiment, a network device determines a path from itself to a source device in a computer network, where the source device utilizes the path in reverse to reach the network device. Based on determining a reliability of the path in reverse, the network device may dynamically adjust one or more keepalive parameters for keepalive messages sent on the path. Accordingly, the network device may then send keepalive messages on the path based on the dynamically adjusted keepalive parameters.
摘要:
In one embodiment, a network device may receive an indication of a particular future message time, and determines a path validation time that is prior to the particular future message time by an amount at least long enough to detect and report a route change of a path from the network device to a source of the particular future message, wherein the source utilizes the path in reverse to reach the network device for the particular future message. Accordingly, the network device sends, at the path validation time, a keepalive message on the path, where in response to a failure of the keepalive message on the path, the network device repairs the path to the source with a particular route change, and reports the particular route change to the source, e.g., such that in response, the source may transmit the particular future message on the changed path in reverse.
摘要:
In one embodiment, an expected transmission count (ETX) link metric is computed for a link between a transmitter and a receiver in a communication network, the ETX representative of an expected number of transmissions necessary for a message to be successfully received by the receiver over the link, and a data rate of the link at which the ETX is computed is also determined. From these, an estimated transmission overhead (ETO) link metric for the link may be computed by dividing the ETX by the data rate. In one embodiment, the data rate of the link may be adjusted based on the ETO (e.g., to minimize the ETO). In another embodiment, routes through the communication network may be selected based on ETO values along the route.
摘要:
In one embodiment, a communication device in a frequency hopping communication network determines an intention to forward a first packet in a particular timeslot of a frequency hopping sequence. As such, the device scans in receive mode for an initial portion of the particular timeslot on a particular frequency known to neighbors of the communication device for reaching the communication device. In response to determining that the communication device is receiving a second packet during the initial portion, the device remains in receive mode to receive a remainder of the second packet. Conversely, in response to not receiving the second packet during the initial portion, the device proceeds to transmit the first packet during a remainder of the particular timeslot.
摘要:
In one embodiment, a node joins a communication network, and in response to joining the network, operates in a rapid startup mode, wherein the node in rapid startup mode establishes network configurations rapidly by deemphasizing quality (optimality) of the network configurations. Subsequent to operating in the rapid startup mode (e.g., after some timer or explicit command), the node then operates in a robust mode, wherein the node in robust mode iteratively refines the network configurations to increase the quality of the network configurations.
摘要:
In one embodiment, a node in a computer network detects a misbehaving node in the computer network based on the misbehaving node acting in violation of one or more rules. As such, the node communicates information regarding the misbehaving node to a network management system (NMS), and then may receive isolation instructions from the NMS regarding how to isolate the misbehaving node from the computer network. Accordingly, the node may perform the isolation instructions. In another embodiment, the NMS receives the communicated information regarding the misbehaving node, and determines whether the misbehaving node should be isolated based on the communicated information. If so, then the NMS determines isolation instructions regarding how to isolate the misbehaving node from the computer network, and transmits them to one or more nodes in the computer network, accordingly.
摘要:
In one embodiment, a device in a computer network monitors an alternating-current (AC) waveform of an electrical power source at the device, where the power source is part of a polyphase power source system. Once the device determines a particular phase of the polyphase power source system at the device, then the device joins a directed acyclic graph (DAG) specific to the particular phase. In another embodiment, a device detects a time of a zero crossing of the AC waveform, and may then determine a particular phase of the polyphase power source system at the device based on the time of the zero crossing relative to a corresponding location within a frequency hopping superframe of the computer network.
摘要:
In one embodiment, a particular node in a computer network receives an indication of a number of child nodes of one or more potential parent nodes to the particular node in a primary directed acyclic graph (DAG). From this, the particular node selects a particular potential parent node with the highest number of child nodes as a secondary DAG parent for the particular node, and joins the secondary DAG at the selected secondary DAG parent (e.g., for multicast and/or broadcast message distribution). This may recursively continue, such that nodes gravitate toward parents with more children, potentially allowing parents with fewer children to relinquish their parental responsibilities.
摘要:
In one embodiment, in response to a trigger condition being detected at a particular location in a primary directed acyclic graph (DAG) in a computer network, a particular node in the primary DAG at the particular location may be determined to act as a remote stitched (RS)-DAG root for an RS-DAG at the particular location. The determined RS-DAG root may then be instructed to initiate the RS-DAG, the instructing indicating one or more properties for the RS-DAG that are based on the trigger condition and that are different from properties of the primary DAG. In another embodiment, a particular node receives instructions to initiate an RS-DAG as its RS-DAG root, initiates the RS-DAG, and relays messages of the RS-DAG with a primary root of the primary DAG.
摘要:
In one embodiment, a probing technique allows a root node to determine whether to trigger reoptimization of a computer network represented by a directed acyclic graph (DAG) without injecting unnecessary traffic into the network. The root node may store and maintain information indicative of an ideal shape or topology of the DAG. During a normal DAG maintenance operation, the root node may transmit a DAG discovery request (probe request) that is configured to probe each node within the DAG for information used to determine a current topology of the DAG. In response, each node may record the information, e.g., routing and non-routing metrics, in a DAG discovery reply (probe reply) that is propagated to the root node. Upon receiving one or more replies, the root node may analyze the metrics to determine whether the current topology of the DAG deviates from the ideal DAG topology. The root node may thus determine DAG topology deviation upon probing at minimal cost. A number of algorithms may then be used to determine whether reoptimization, i.e., global repair, of the DAG is is required and, if so, the root node may dynamically trigger the global repair.