Abstract: The present disclosure relates scaling out servers that performs Layer 3 (L3) termination. In particular, the server includes: a packet receiving unit which receives a packet from any load balancer performing L3 termination via an Layer 2 (L2) switch; an assignment unit which, by referencing a MAC/MARK number correspondence table in a storage unit, assigns to the received packet a MARK number that corresponds to a transmission source MAC address; a recording unit which records a MARK number for connection of the received packet in a connection/MARK number correspondence table; and a packet transmission unit which, when a reply packet to the received packet is to be transmitted, routes the reply packet via the L2 switch by acquiring, from a MARK number/GW correspondence table in the storage unit, an IP address of a load balancer corresponding to the MARK number associated with the connection of the received packet.

Abstract: A data flow classification device includes a forwarding circuit and a configuring circuit. The forwarding circuit looks the classification of an input flow up in a lookup table according to the information of the input flow, tags the packets of the input flow with the classification, and outputs the packets to a buffer circuit; but if the classification is not found in the lookup table, the forwarding circuit tags the packets with a predetermined classification, outputs the packets to the buffer circuit, and adds the information of the input flow to the lookup table. The configuring circuit determines a flow threshold according to a queue length of the buffer circuit and a target length, learns the traffic of multiple flows from the lookup table, determines the classifications of the multiple flows according to the comparison between the traffic and the flow threshold, and stores these classifications in the lookup table.

Abstract: Certain aspects of the present disclosure provide techniques for low power channel sensing for pedestrian user equipments (P-UEs). A method that may be performed by a UE (e.g., such as a P-UE) includes determining a level of channel congestion for a channel during a first sensing duration. The method includes determining a second sensing duration and a transmission duration based on the level of channel congestion for the channel determined from the first sensing duration. The method includes sensing the channel for the second sensing duration. The method includes transmitting on the channel for the transmission duration.

Abstract: Systems, methods, and computer-readable media storing instructions for handling wireless communication in wireless mesh networks are disclosed herein. The disclosed techniques virtualize a portion of the communications within the wireless mesh network by implementing a plurality of virtual routers on a server, each virtual router corresponding to a physical router device within the wireless mesh network. Each physical router device is connected to its corresponding virtual router by a deterministic connection, which may include a wired connection to the server. A virtual communication layer implemented by the server is configured to router inter-router communication between the virtual routers. In some embodiments, each virtual router is implemented within a separate container within a virtual mesh network environment running on the server.

Abstract: Systems and methods for controlling congestion in a data network are provided. A base target round-trip time (RTT) for packets of a network flow including packets transmitted from a source network device to destination network device is obtained. A number of hops packets associated with the network flow traverse between the source network device and the destination network device is determined. A topology scaled target RTT for the network flow is determined based on the base target RTT and the determined number of hops. A congestion window size for the network flow is managed based on the topology scaled target RTT.

Abstract: A method of managing communication services provided by a service provider, comprising obtaining a prediction of a portion of the quantity of the served devices that received substandard communication services from the service provider during a period of time, obtaining an acceptable deviation from the prediction of the portion of served devices, making a determination that a quantity of the served devices that received substandard communication services from the service provider during the period of time is outside of a range, and performing an action set to initiate remediation of the service provider. Specifically, the determination is based on the prediction of the portion of the quantity of served devices, and the acceptable deviation.

Abstract: A method performed by a user equipment (UE) wirelessly connected to a wireless communication network for facilitating transmission of a real-time media flow from a sender towards the UE is described. The method includes sending a request to the network, requesting a packet loss prediction and a throughput prediction for transmission of a real-time media flow from the network towards the UE. The method also includes receiving, from the network, information of a predicted UE throughput and a predicted UE packet loss predicted by the network in response to the sent request. In some embodiments, the method also includes sending the received information to a sender of the real-time media flow to the UE.

Abstract: A method of handling an overload situation of a Session Initiation Protocol, SIP, node in a telecommunication network, wherein the method includes the steps of receiving, by a proxy SIP node in the telecommunication network, an SIP request for requesting a service in the telecommunication network, forwarding, by the proxy SIP node, the SIP request to a service SIP node in the telecommunication network, receiving, by the proxy SIP node, from the service SIP node, a Server Failure Response message, wherein the Server Failure Response message includes an indication that the service node cannot fulfil the SIP request due to an overload situation at the service SIP node, and restricting, by the proxy SIP node, future SIP traffic to the service SIP node based on the received Server Failure Response message.

Abstract: A method, system, and computer program product to analyze data patterns in source workloads and predict the optimal parallel apply algorithms, where the method may include receiving source workload data and replication environment data, where the source workload data includes at least a stream of changes to a target DBMS. The method may also include analyzing characteristics of the source workload data and the replication environment data. The method may also include inputting, as input variables, the characteristics of the source workload data and the replication environment data into a machine learning algorithm. The method may also include obtaining, from the machine learning algorithm, an optimal parallel apply algorithm from a plurality of parallel apply algorithms. The method may also include applying the optimal parallel apply algorithm to the target database management system.

Abstract: Traffic for USB devices that are connected to a USB-C dock can be adaptively prioritized. When the consumption of the bandwidth of a connection between a computing device and a USB-C dock exceeds a threshold, a filter driver can notify a service. The service can update device priority values for the devices that are connected to the USB-C dock based on applications that are accessing the devices. The service can relay the updated device priority values to the filter driver. The filter driver can then attempt to reduce bandwidth consumption by changing device settings of any device with a lower priority value, and then, if changing device settings is insufficient, may attempt to reduce bandwidth consumption by lowering the priority of the device's traffic.

Abstract: A method of identifying a behavior direction recognition based service requester includes a plurality of intelligent robot devices arranged in the airport and a server controlling movement of one or more intelligent robot devices of the plurality of intelligent robot devices. The intelligent robot includes a communication unit transmitting movement information of an airport user moving in the airport from the server, and a processor configured to receive the movement information of the airport user from the communication unit, learn movement of the airport user, recognize a wandering state of the airport user based on the learned movement of the airport user, and move to the airport user based on a result of the recognition. The intelligent robot device may be associated with an artificial intelligence module, an unmanned aerial vehicle (UAV), a robot, an augmented reality (AR) device, a virtual reality (VR) device, and devices related to 5G services.

Abstract: A packet transmission system according to an example embodiment includes endpoint devices, a first transmission path connected to the endpoint devices and including a first radio section, a second transmission path connected to the endpoint devices and including a second radio section, bandwidth monitoring units that respectively monitor bandwidth information of the first radio section and the second radio section for each flow, a path switching unit that determines a packet transmission path for transmitting a packet on the basis of the bandwidth information of the first radio section and the second radio section, a bandwidth ratio calculation unit that calculates a bandwidth ratio of flows on the basis of the bandwidth information of the first radio section and the second radio section and determines a bandwidth control value, and a bandwidth control unit that controls a bandwidth of each flow on the basis of the bandwidth control value.

Abstract: A system and method for bot detection utilizing storage variables are presented. The storage variables generated is used to analyze user behavior and distinguish human traffic from bot traffic. The system for detecting bot traffic using storage variables includes a client application, a computer network, a bot detector, a bot computer, a storage variable generator, and a server. The client device enables a user to access information through the client application. The storage variable generator is configured to generates a plurality of storage variables including counter storage variable. The bot detector analyses the presence of bots in incoming traffic.

Abstract: This application provides data transmission methods and apparatuses. One method includes: processing, by a network device, a first optical data unit (ODU) to obtain a second ODU, wherein a bit rate of the second ODU is lower than a bit rate of the first ODU; and sending, by the network device, the second ODU.

Abstract: A method and system for prioritizing critical data object storage during backup operations. Specifically, the method and system disclosed herein entail reordering data objects, awaiting being written to storage and thus queued in one or more data object queues, in accordance with a nearest-critical based sequential order. The nearest-critical based sequential order may be derived through modified weight-based Euclidean distances calculated between adjacent data object pairs queued in any given data object queue. Further, the calculated modified weight-based Euclidean distances incorporate data criticality factors associated with the adjacent data object pairs. By reordering data objects in a nearest-critical based sequential order, critical data objects may be written into storage first, thereby avoiding possible critical data loss should a disaster occur during backup operations.

Abstract: Embodiments are provided for managing routes of data traffic within a network. The management may be performed via a graphical user interface that interacts with a Web server to update a configuration file. The configuration file can be converted to router management commands by a network management device (e.g., a BGP speaker). The commands can then be sent to border routers for controlling network traffic. Embodiments are also provided for capturing and logging routing updates made in a network.

Abstract: A method includes determining, during auxiliary power operating mode of an information handling system, devices connected to peripheral component interconnect (PCI) slots and identifying, via a controller of the IHS, when a Virtual Ethernet Bridge (VEB) is present on a first communication path, which indicates that a high power consuming device with sideband enablement is present in a corresponding one of the PCI slots. The method includes configuring, via the first communication path, an uplink port on the VEB, collecting information about the VEB, including a VEB MAC address and capabilities of the VEB, setting a VLAN/MAC filter rule for NC-SI pass-through in the VEB, and configuring an intermediate device within a second communication path between the controller and the VEB for NC-SI pass-through utilizing a management MAC address, enabling NC-SI pass-through of traffic routed between the controller and the VEB.

Abstract: The present disclosure is generally directed to media systems configured to receive and play media assets. In particular, methods and systems are provided for improved media asset session continuity across such media systems. Systems and methods are provided herein for continuing media asset sessions across media systems or media devices in a way designed to minimize manual intervention, for example, by determining a likelihood (e.g., a probability) of a user requesting media session continuation of an ongoing media asset or a segment thereof across two or more devices.

Abstract: This application provides a flow control method, including: sending, by a first device, a first data packet to a second device in a first sending mode; determining, by the first device, a network congestion status; and sending, by the first device, a second data packet to the second device in the second sending mode, where the second sending mode is a sending mode determined by the first device based on the network congestion status, and the second data packet is a data packet to be sent after the first data packet.

Abstract: Apparatus and methods are disclosed, including a memory device or a memory controller configured to determine that a condition has occurred that indicates a performance throttling operation, implement a performance throttling responsive to the determined condition, responsive to implementing the performance throttling, set a performance throttling status indicator in an exception event status attribute, receive a command from a host device across a memory device interface, perform the command, prepare a response to the command, the response including a flag indicating that the performance throttling status indicator is set in the exception event status attribute, and send the response to the host device. Methods of operation are disclosed, as well as machine-readable medium and other embodiments.

Abstract: Aspects of the disclosure relate to transmitting content throughout a distribution network. For example, a computing device may generate data packets and attribute packets, and transmit the data packets and attribute packets to one or more user devices. The attribute packets may comprise information that enables validation and storage after the data packet is reassembled by the receiving device.

Abstract: A system is described. The system includes a processing resource and a non-transitory computer-readable medium, coupled to the processing resource, having stored therein instructions that when executed by the processing resource cause the processing resource to collect quality of service (QoS) data of a distributed storage system associated with a client device, generate a plurality of QoS diagnostic views to facilitate display of visual representations of the collected QoS data and transmit the visual representations for display at the client device, wherein the data comprises telemetry data, statistics and performance health scores for nodes associated with the client device.

Abstract: A method and system for determining a path of data traffic based on a destination Internet Protocol (IP) address, the destination IP address being either private or public and belonging to any one of multiple organizations. In the method and system, a VPN is activated at a telecommunication endpoint. In response to activating the VPN, a service device on a private network interconnects with a VPN gateway. A subscriber identifier associated with the telecommunication endpoint and the destination IP address of one of the multiple organizations is received at a managed service platform. If both the subscriber identifier and the destination IP address of the one of the multiple organizations are not registered in a predetermined policy database at the telecommunication endpoint, a subscriber identified by the subscriber identifier is prompted to register the subscriber identifier associated with the telecommunication endpoint.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to reduce temperature of a networked device. An example apparatus includes, a temperature threshold monitor to identify a temperature condition associated with the device, a window information retriever to retrieve a current value of a network receive capacity parameter, and a window adjustor to reduce the temperature of the device by generating a modified network receive capacity parameter, the modified network receive capacity parameter based on a ratio of the current value of the network receive capacity parameter and a decrease factor.

Abstract: Automated dynamic load balancing across VNIC fast switchover devices including monitoring, by a resource monitor, resource utilization metrics of at least one resource used to send data along a first communications path through a hypervisor; in response to detecting, by the resource monitor based on the resource utilization metrics, that the at least one resource crosses a resource utilization threshold, sending a notification to a rebalancer, wherein the notification comprises a subset of the resource utilization metrics; generating, by the rebalancer based on the subset of the resource utilization metrics, a resource optimized mapping for the at least one resource used to send data along the first communications path; and issuing, by the rebalancer, the resource optimized mapping to the hypervisor, wherein the hypervisor reconfigures, based on the resource optimized mapping, the at least one resource used to send data along the first communications path.

Abstract: Some embodiments provide a novel method for receiving a plurality of attribute sets from a set of host computers, each attribute set associated with a group of one or more flows that is created by using a key to associate individual flows into the group of flows. The appliance, in some embodiments, identifies at least two received attribute sets from two different host computers that relate to a same set of flows between a same set of source machines and a same set of destination machines. The appliance merges the two identified attribute sets into one merged attribute set and analyzes the merged attribute set to identify a set of properties of the flows in the groups of flows associated with the two identified attribute sets, in some embodiments.

Abstract: A content distribution system is provided comprising a content client and a content server connected through a network. The content server is configured to send the content client a session description comprising at least one media description to enable it to receive over the network a content item. The system comprises a network monitor for monitoring network traffic going through the network, and a recommender for marking in dependency upon the network monitor, one or more of the media descriptions in the session description to obtain a marked session description. The marked media description recommends the content client to prefer certain content above others in view of the current network condition. For example, if available bandwidth is low, content is recommended which requires low bandwidth. Network congestion is thereby avoided for all users of the system.

Abstract: Techniques are described for using a supplemental data layer of a real time payments channel to send data. For example, an entity may request a transaction with another entity, and the transaction may be performed using a channel that includes a supplemental data layer. Supplemental data, in addition to the data for the transaction itself, may be communicated using the supplemental data layer to entities involved in the transaction, and/or to third parties. The supplemental data may be sent in real time with respect to the sending of the data for the transaction. In some instances, supplemental data layer is present in one or both directions in the two-way communications between the entities involved in the transaction. The supplemental data layer can be leveraged to provide additional information and/or services to a user involved in the transaction.

Abstract: The invention includes a method and apparatus for delivering content to one or more content destination nodes. A method includes receiving content utility prediction information for a content item, selecting a content distribution mode for the content item using the content utility prediction information, and propagating the content item toward at least one of the content destination nodes using the selected content distribution mode and, optionally, with a defined priority, sequence, or schedule. The content utility prediction information is associated with the content destination nodes, and is indicative of a level of utility of the content item to the content destination nodes. The content distribution mode may include any content distribution mode, such as broadcast, switched broadcast, multicast, unicast, and the like. The content utility prediction information is received from one or more content prediction nodes.

Abstract: Method of limiting offending messages communicated over a network, such as but not limited to messages associated with Spam and DoS attacks. The message limiting optionally including limiting bandwidth or other communication capabilities associated with an entity communicating or facilitating communication of the messages.

Abstract: A locationing platform may receive, via a mesh network, a locationing communication associated with a waypoint beacon. The locationing platform may identify an entry node of the locationing communication. The locationing platform may determine a waypoint location associated with the waypoint beacon based on locationing information in the locationing communication. The locationing platform may determine a node location of the entry node based on the waypoint location. The locationing platform may configure, based on the node location, a relay node of the mesh network to forward or drop subsequently received locationing communications that are associated with the waypoint beacon.

Abstract: In some implementations, a computer may during a repeating interval in a local GCAC node: (a) receiving and storing key data from one or more other GCAC nodes, the key data indicative of resource usage data for each key in the respective one or more other GCAC nodes wherein each key is associated with an allocation limit equal to a number of resources it may use at any given time across all GCAC nodes; (b) storing current key data, the current key data indicative of the resource usage data for each key in the local GCAC node; (c) determining estimated network-wide key data, the estimated network-wide key data indicative of an estimated network-wide resource usage across all GCAC nodes for each key, the estimated network-wide key data based on a combination of the current key data for the local GCAC node and a weighted multiple of the key data in each of the one or more other GCAC nodes; (d) storing the estimated network-wide key data for each key; and (e) repeating steps (a)-(d) during each interval.

Abstract: A network node and a method performed in a network node of a cellular communication network comprising the steps of obtaining an indicator to secure a traffic availability for a connection, wherein the traffic availability is related to how likely the connection will remain operable, determining whether the network node can secure resources to support the traffic availability, allocating resources for the connection when resources can support the traffic availability, and transmitting a positive response when the network node can secure resources to support the traffic availability.

Abstract: A link detection technology. In a link detection method, when a target event occurs on paths, in a plurality of paths, whose quantity exceeds a preset quantity, whether the paths on which the target event occurs share a bottleneck link is detected, and congestion control is performed. It can be learned that in the technical solution provided in the embodiments of this application, the target event is a trigger, and shared bottleneck detection is performed after the quantity of the paths on which the target event occurs exceeds the preset quantity. This implements purposeful link detection, reduces system overheads, and brings a fast convergence speed.

Abstract: A method and system for prioritizing critical data object recovery during restore operations. Specifically, the method and system disclosed herein entail reordering data objects, awaiting being read from backup storage and thus queued in one or more data object queues, in accordance with a nearest-critical based sequential order. The nearest-critical based sequential order may be derived through modified weight-based Euclidean distances calculated between adjacent data object pairs queued in any given data object queue. Further, the calculated modified weight-based Euclidean distances incorporate data criticality factors associated with the adjacent data object pairs. By reordering data objects in a nearest-critical based sequential order, critical data objects may be recovered first, thereby avoiding possible critical data loss should a disaster occur during restore operations.

Abstract: Embodiments of the disclosure provide a data write-in method and apparatus. The data write-in method includes: selecting a target replica server from a plurality of replica servers managed by a metadata server according to remaining capacity of the plurality of replica servers; selecting a write-in disk in the target replica server according to remaining capacity and load of disks managed by the target replica server; and storing write-in data into the write-in disk through the target replica server.

Abstract: A load balancer receives a sequence of requests for computing service and distributes the requests for computing service to a computing node in an ordered list of computing nodes until the computing node reaches its maximum allowable compute capability. Responsive to an indication that the computing node has reached its maximum allowable compute capability, the load balancer distributes subsequent requests for computing service to another computing node in the ordered list. If the computing node is the last computing node in the ordered list, the load balancer distributes a subsequent request for computing service to a computing node other than one of the computing nodes in the ordered list of computing nodes. If the computing node is not the last computing node in the ordered list, the load balancer distributes a subsequent request for computing service to another computing node in the ordered list of computing nodes.

Abstract: Managing traffic on a distributed system includes partitioning VCE load balancers of the distributed system among service virtual server instances (SVSIs) of the distributed system, each service virtual server instance (SVSI) being identical in function and deployed across two or more regions of the distributed system. The method further includes: periodically polling, by each SVSI each other of the SVSIs; monitoring, by each SVSI, a load on each VCE load balancer for which it is responsible, resulting in a monitored load; recovering, by an SVSI from a failure of VCE load balancer(s) for which it is responsible; automatically adjusting, by an SVSI a capacity of a VCE load balancer for which it is responsible based on a corresponding monitored load; and repartitioning, by the distributed system, in response to an SVSI failure, the VCE load balancers among remaining SVSIs.

Abstract: An information handling system includes a host, an intelligent peripheral connected to the host, and a baseboard management controller. The host includes a memory and a central processing unit. The intelligent peripheral is configured to communicate with the baseboard management controller to obtain configuration information, and provide networking services to the host after configuration. The baseboard management controller is configured to communicate with a management network; establish a communication channel with the intelligent peripheral independent of the host system; obtain configuration information for the intelligent peripheral from the management network; and provide the configuration information to the intelligent peripheral.

Abstract: Storage-based slow drain detecting and automated resolution is provided herein. A data storage system as described herein can include a memory that stores computer executable components and a processor that executes computer executable components stored in the memory. The computer executable components can include a switch query component that obtains a host transfer rate negotiated between a host device and a network switch from a host-connected port of the network switch; a comparison component that compares the host transfer rate to an array transfer rate negotiated between the network switch and a storage array; and a rate limiter component that limits a data transfer from the storage array to the host device to the host transfer rate in response to the host transfer rate being less than the array transfer rate.

Abstract: Distributed machine learning systems and other distributed computing systems are improved by embedding compute logic at the network switch level to perform collective actions, such as reduction operations, on gradients or other data processed by the nodes of the system. The switch is configured to recognize data units that carry data associated with a collective action that needs to be performed by the distributed system, referred to herein as “compute data,” and process that data using a compute subsystem within the switch. The compute subsystem includes a compute engine that is configured to perform various operations on the compute data, such as “reduction” operations, and forward the results back to the compute nodes. The reduction operations may include, for instance, summation, averaging, bitwise operations, and so forth. In this manner, the network switch may take over some or all of the processing of the distributed system during the collective phase.

Abstract: A multimedia file transmission method, which relates to the fields of data transmission and image processing, includes acquiring network information; determining a transmission version of a multimedia file according to the network information, where different transmission versions are corresponding to different multimedia file quality; and transmitting the multimedia file according to the transmission version. The network information is automatically acquired, and the transmission version of the multimedia file is automatically determined according to the network information, which not only saves an operation of manually selecting, by a user, a transmission version and reduces operation complexity, but also improves appropriateness of determining a transmission version.

Abstract: Systems and methods for monitoring and steering service requests to acceleration components are provided. An example method in a system comprising a host configured to send messages, corresponding to a service, to any of a set of acceleration components may include tracking an input buffer occupancy corresponding to each of the set of acceleration components. The method may further include, upon a first input buffer occupancy of a first acceleration component from among the set of the acceleration components meeting a predetermined criterion, generating a first penalty value for the first acceleration component based at least on the first input buffer occupancy associated with the first acceleration component, and comparing a second value indicative of an availability of the second acceleration component to receive messages from the host to a second dynamically updated value to determine whether the host can send a next message to the second acceleration component.

Abstract: An exemplary system according to the present disclosure comprises at least one mobility networking module implemented in at least one lower tier radiolet. The mobility network module performs one or more traffic data plane functions. The exemplary system further comprises at least one mobility application module implemented in at least one upper tier radiolet. The mobility application module performs one or more application data plane functions. The exemplary system further comprises at least one mobility cloud module implemented in the at least one lower tier radiolet and the at least one upper tier radiolet. The mobility cloud module performs computing execution functions of a cloud session.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that facilitate near real-time dynamic generation of graphical content based on audio, text, or gesture input. An example method includes receiving an input from an input device such as a microphone, a keyboard, or a camera. As such, the input may include text, speech, and/or a gesture. The method includes determining, based on the input, one or more command instructions. The method also includes determining, based on the one or more command instructions, a scene layout. Each of the one or more command instructions correspond to at least one element of the scene layout. The method also includes providing a rendered scene based on the determined scene layout.

Abstract: In one embodiment, a computing device determines a window of time that is available after each operation to fill a frame buffer. The device determines a first number indicating a number of elements of a requested content that fit into a display region, an increment based on an application-specific rule, and a second number indicating a number of elements to pre-generate based on the first number and the increment. The device selects one or more pre-generation tasks for pre-generating content for the second number of select elements of the requested content. During the window of time, the device executes at least a subset of the pre-generation tasks to pre-generate content for a corresponding subset of the select elements. The pre-generated subset of the select elements is stored in an application-tailored recycler.

Abstract: A base station can select orthogonal frequency-division multiplexing (OFDM) numerologies that define subcarrier spacing values based on attributes associated with one or more services that a user equipment (UE) is using. The base station can use the selected OFDM numerologies for transmission associated with the services. When the UE is using multiple services simultaneously, the base station can select the same or different OFDM numerologies for the multiple services.

Abstract: A device, method, and computer readable storage medium for routing in a computational grid arranged in a fat-tree, including performing preprocessing, performing base path routing to determine rules for routing in switches based on stored switch level, switch group, links between switches, links between switches and hosts, and fat-tree type, and performing run-time processing, including detecting a path fault, and determining at least one redundant path for the path fault. The base path routing determines a first routing rule that specifies a certain destination in every switch. The run-time processing further performs searching for a first match between a packet and a source-destination pair, and when the first match is not found, performs searching for a second match between the packet and the first routing rule that specifies the certain destination. By using routing rules that specific a destination, the routing in a computational grid can significantly scale up.

Abstract: A plurality of agents is instantiated on a corresponding plurality of servers that are running concurrent instances of a control plane application. A first agent is selected from the plurality of agents to mediate communication between a client and the concurrent instances of the control plane application running on the plurality of servers. The first agent replicates uplink packets received from the client for transmission to the plurality of agents. A first connection is used to convey the uplink packets. The first agent transmits a single copy of redundant downlink packets received from the plurality of agents to the client via the first connection. In some cases, the first connection is a stateful transmission control protocol (TCP) connection and the first agent replicates the uplink packets received in a single TCP stream. The first agent merges the redundant downlink packets into the single TCP stream.

Abstract: A packet splicer receives, from a first server of a plurality of servers, a first request for retransmission of an uplink packet that was received from a client and replicated to the plurality of servers. The plurality of servers is running concurrent instances of a control plane application configured to process information in the uplink packet. The packet splicer transmits the first request for retransmission of the uplink packet to the client. The packet splicer bypasses transmission of a second request for retransmission of the uplink packet received from a second server of the plurality of servers. In some cases, the packet splicer receives, from the client, a request for retransmission of a downlink packet that was redundantly transmitted by the plurality of servers. The packet splicer replicates the request for retransmission of the downlink packet to the plurality of servers.