Abstract: A computer-implemented method may comprise receiving proposals to mutate a data stored in a distributed and replicated file system coupled to a network, the distributed and replicated data system comprising a plurality of nodes, each comprising a server. A metadata service maintains and updates a replica of a namespace of the distributed and replicated file system and coordinates updates to the data by generating an ordered set of agreements corresponding to the received proposals, the ordered set of agreements specifying an order in which the nodes are to mutate data stored in data nodes and cause corresponding changes to the state of the namespace. For each agreement in the generated ordered set of agreements, a corresponding writers list may be provided that comprises an ordered list of nodes to execute the agreement and make corresponding changes to the namespace.

Abstract: Methods and systems for managing deployments are disclosed. A deployment may include one or more devices. The devices may include hardware and/or software components. The operation of the deployment may depend on the operation of these devices and components. To manage the operation of the deployment, a system may include a deployment manager. The deployment manager may obtain logs for components of the deployment reflecting the historical operation of these components and use the log to predict the future operation of the deployment. Based on the predictions, the deployment manager may take proactive action to reduce the likelihood of the deployment becoming impaired.

Abstract: Data units can be relocated in scale-out storage systems. For example, a computing device can receive, at a first node of a scale-out storage system, a request for a data unit. The first node can include a metadata entry associated with the data unit. The computing device can determine, based on the metadata entry, that a second node of the scale-out storage system includes the data unit. The computing device can determine, from the metadata entry, that a number of versions of the data unit in the scale-out storage system meets or exceeds a threshold. The computing device can output a command to cause the data unit to be relocated to the first node with the metadata entry.

Abstract: Systems, apparatuses, methods, and computer-readable media are provided for device interface management. A device includes a device interface, a virtual machine (VM) includes a device driver, both to facilitate assignment of the device to the VM, access of the device by the VM, or removal of the device from being assigned to the VM. The VM is managed by a hypervisor of a computing platform coupled to the device by a computer bus. The device interface includes logic in support of a device management protocol to place the device interface in an unlocked state, a locked state to prevent changes to be made to the device interface, or an operational state to enable access to device registers of the device by the VM or direct memory access to memory address spaces of the VM, or an error state. Other embodiments may be described and/or claimed.

Abstract: A method and system for license server synchronization are disclosed. According to one embodiment, a computer-implemented method comprises receiving a first capability request from a licensing server, transmitting a first capability response to the licensing server. The licensing server provides a second capability response to a device, and the second capability response is in response to a second capability request sent by the device to the licensing server. The second capability response includes one or more capabilities included in the first capability response. A server synchronization message is received from the licensing server and a request for payment is transmitted to a service provider, wherein the request for payment is generated based on the server synchronization message.

Abstract: This disclosure discloses an address translation method, apparatus, and system. The method includes: receiving, by a first address translation unit (ATU), an access request sent by a first virtual machine (VM), where the access request carries a first address and parameter information of the first address, and the parameter information of the first address includes an identifier of a target virtual function (VF); determining, by the first ATU, a matching relationship based on the parameter information of the first address and VF range description information in an address mapping table; and when an identifier of a target VF and the VF range description information are successfully matched, translating, by the first ATU, the first address into a second address based on the address mapping table.

Abstract: The DSI protocol (distributed system interface) that relates to the bus communication in sensor arrangements is known from the prior art. Methods are proposed for operating a sensor arrangement (12) on the basis of the DSI protocol in order to consequently be able to operate, in particular in an advantageous manner, active sensor units (20A-20F) having a transmitter (21) and a receiver (22). In particular, the use of three part phases is proposed for the communication between the processing unit (14) and the active sensor unit (20A-20F), namely a CRM phase (40) for the bidirectional communication between the processing unit (14) and the active sensor unit (20A-20F), a power phase (43) for the transmission of energy from the processing unit (14) to the active sensor units (20A-20F) and a PDCM phase for the unidirectional transmission of data from the active sensor unit (20A-20F) to the processing unit (14).

Abstract: Embodiments of this application provide a link group configuration method and an apparatus. The link group configuration method includes: obtaining, by a first network device, candidate groups to which M physical ports of the first network device belong respectively; obtaining, by the first network device from a second network device, candidate groups to which M physical ports of the second network device belong respectively; and selecting N physical links from M physical links, as a link group between the first network device and the second network device, based on the candidate groups to which the M physical ports of the first network device belong respectively and the candidate groups to which the M physical ports of the second network device belong respectively.

Abstract: The present teaching relates to request management and data recovery in a data system. In one example, a failure in connection with first data is detected at a first node in a data system. Information associated with a most recent transaction related to the first node is obtained from the persistent storage. Each of other nodes in the data system is requested to transmit one or more transaction requests previously sent to the first node after the most recent transaction. The one or more transaction requests are received from at least one of the other nodes. A sequence of one or more transactions associated with the one or more transaction requests is determined. The one or more transactions are executed according to the sequence in order to recover the first data at the first node.

Abstract: An example operation may include one or more of receiving a data packet from a network controller, where the data packet is marked as valid based on a checksum calculated by the network controller, determining that the checksum generated by the network controller is in error based on a recalculation of the checksum of the data packet via a network layer, and transmitting the data packet from the network layer to the network controller with a notification that the checksum is in error.

Abstract: Provided are a computer program product, system, and method for using a machine learning module to determine when to replace a storage device. Input on attributes of the storage device is provided to a machine learning module to produce an output value. A determination is made whether the output value indicates to replace the storage device. Indication is made to replace the storage device in response to determining that the output value indicates to replace the storage device.

Abstract: Provided are a computer program product, system, and method for using a machine learning module to determine when to replace a storage device. Input on attributes of the storage device is provided to a machine learning module to produce an output value. A determination is made whether the output value indicates to replace the storage device. Indication is made to replace the storage device in response to determining that the output value indicates to replace the storage device.

Abstract: Disclosed herein are methods, systems, and processes to perform input/output (I/O) fencing without dedicated arbitrators. A coordination storage identifier is stored in a storage device as metadata. The coordination storage identifier is associated with a node of a cluster coupled to the storage device, and an I/O fencing operation is performed using the coordination storage identifier.

Abstract: Example peer storage systems, storage devices, and methods provide peer data recovery across a peer communication channel. Peer storage devices establish peer communication channels that communicate data among the peer storage devices. A storage device may identify storage media segments from their storage medium for recovery of failed data units. A peer storage device may be identified that contains recovery data for the failed data units. The recovery data may be received over the peer communication channel and the storage media segments may be recovered using the recovery data.

Abstract: A high availability environment of resource adapters implements processes to manage and to distribute work among the adapters or adapter instances. An input resource, such as a file, is received and tasks are created to distribute the content to the different instances of the adapters that are configured in the cluster. A resource adapter instance switches to manage the creation of the task based on task-definitions of the adapter. The task-definitions are rules specified in the adapter on chunks of data. The tasks are created such that chunks of data are independently locked and processed without duplication. In order to distribute the work, the tasks are persisted into a table/xml on a persistent disk. The remaining instances interact with the table to access the tasks specified by the entries in the table, thus executing the tasks.

Abstract: A database cluster architecture based on dual port solid state disk (SSD) is provided. The database cluster includes a plurality of nodes and a plurality of dual port solid state disks (SSDs). Each dual port SSD of the plurality of dual port SSDs is connected to a unique pair of nodes and each node in the unique pair of nodes is connected to a unique pair of dual port SSDs. A first node of the plurality of nodes writes a transaction log to a first dual port SSD, a second node connected to the first dual port SSD acquires the transaction log from the first dual port SSD and operates a database process to recover data according to the transaction log when the first node crashes.

Abstract: Disaggregated computing architectures, platforms, and systems are provided herein. In one example, a data system is provided. The data system includes a first assembly comprising a plurality of modular storage bays populated with one or more graphics processing modules each including a graphics processing unit (GPU), wherein the plurality of modular storage bays each comprise a bay connector that includes a bay Peripheral Component Interconnect Express (PCIe) connection. The first assembly further comprises PCIe switch circuitry configured to communicatively couple the bay PCIe connections to a PCIe fabric over one or more external PCIe links.

Abstract: An intelligent log gap detection to prevent unnecessary backup promotion. Specifically, the method and system disclosed herein entail determining whether to pursue a requested database backup type or to promote the requested database backup type to another database backup type, in order to preclude data loss across high availability databases. When a decision is made to pursue the requested database backup type, storage space, intended for backup consolidation on a backup system or media, is saved for future backup requests rather than being consumed as would be the case would the requested database backup type had been promoted.

Abstract: A computer-implemented method includes receiving, from a first log agent, a first log collection. The computer-implemented method further includes receiving a first policy, wherein the first policy includes a definition of a first pattern and a definition of a procedure. The computer-implemented method further includes scanning the first log collection against the first policy to determine a match between a portion of the first log collection and the first pattern, with the matching portion of the first log collection being identified as a first data artefact. The computer-implemented method further includes, responsive to identifying the first data artefact, executing the procedure defined by the first policy, wherein the procedure includes: filtering the first log collection to yield a first group of filtered log entries, receiving a first data collection, and sending the first group of filtered log entries and the first data collection to a recipient system.

Abstract: Embodiments of a sensor network system provide surveillance capabilities in multiple contexts/environments (e.g., military, commercial, scientific, civic, urban, wilderness, etc.). Network nodes may include devices such as sensors, network routers, network controllers, etc. Network sensors may be configured so that power management objectives are maximized. Network sensors (both individually and as a group) may be capable of intelligent and cooperative information gathering, so that the output of the sensor network does not contain high levels of irrelevant information. The network nodes may communicate among one another via one or more communication links, and in some cases, multiple routes between any two network nodes may be available. The sensor network may include aspects of both high data rate and low data rate network features.

Abstract: A computer-implemented method includes receiving, from a first log agent, a first log collection. The computer-implemented method further includes receiving a first policy, wherein the first policy includes a definition of a first pattern and a definition of a procedure. The computer-implemented method further includes scanning the first log collection against the first policy to determine a match between a portion of the first log collection and the first pattern, with the matching portion of the first log collection being identified as a first data artefact. The computer-implemented method further includes, responsive to identifying the first data artefact, executing the procedure defined by the first policy, wherein the procedure includes: filtering the first log collection to yield a first group of filtered log entries, receiving a first data collection, and sending the first group of filtered log entries and the first data collection to a recipient system.

Abstract: Examples disclosed herein relate to identification of an alternate principal member port by a target device in a Storage Area Network (SAN). In some examples, a target device in a SAN may determine that a principal member port of a target driven peer zone on the target device is likely to fail based on diagnostic information related to the principal member port. The target driven peer zone may be configured in the SAN via the target device. In response to determining, the target device may identify an alternate principal member port on the target device. The target device may perform an action to indicate the alternate principal member port as the principal member port.

Abstract: Provided are a computer program product for managing bus interface errors in a storage system coupled to a host and storage. A determination is made as to whether a first number of correctable errors on a first bus interface, connecting a first processing unit to the storage, exceeds a second number of correctable errors on a second bus interface, connecting a second processing unit to the storage, by a difference threshold. The correctable errors in the first and second bus interfaces are detected and corrected in the first and second bus interfaces by first hardware and second hardware, respectively. In response to determining that the first number of correctable errors exceeds the second number of correctable errors by the difference threshold, at least a portion of Input/Output (I/O) requests are redirected to a second processing unit using the second bus interface to connect to the storage.

Abstract: An adaptive interface high availability storage device. In some embodiments, the adaptive interface high availability storage device includes: a rear storage interface connector; a rear multiplexer, connected to the rear storage interface connector; an adaptable circuit connected to the rear multiplexer; a front multiplexer, connected to the adaptable circuit; and a front storage interface connector, connected to the front multiplexer. The adaptive interface high availability storage device may be configured to operate in a single-port state or in a dual-port state. The adaptive interface high availability storage device may be configured: in the single-port state, to present a single-port host side storage interface according to a first storage protocol at the rear storage interface connector, and in the dual-port state, to present a dual-port host side storage interface according to the first storage protocol at the rear storage interface connector.

Abstract: Various examples of techniques for identifying a corrupt data lane and using a spare data lane are described herein. Some examples include a system of coordinating spare lane usage between link partners. One such example comprises analyzing data from a link partner to identify a corrupt lane, and communicating the corrupt lane to the link partner, wherein the communication does not require sideband communication channel. In some embodiments, communicating the corrupt lane to the link partner comprises identifying a transmit lane corresponding to the corrupt lane, transmitting a set of data intended for a corresponding transmit lane using a spare data lane, and transmitting bad data to the link partner using the corresponding transmit lane.

Abstract: Machines, systems and methods for enhancing data recovery in a data storage system, the method comprising determining whether one or more data storage mediums in a data storage system are unavailable; determining data that are at a risk of loss, due to said one or more data storage mediums being unavailable; from among the data that is determined to be at the risk of loss, identifying data that is highly vulnerable to loss; and creating one or more temporary replicas of the data that is highly vulnerable to loss.

Abstract: Techniques to achieve high availability in a multi-controller system are disclosed. In various embodiments, an indication is received that a network ping-based heartbeat mechanism has failed. In response to said indication, an alternative heartbeat is provided at least in part by periodically updating a heartbeat data stored in an on-disk heartbeat data structure, in a portion of a disk that is available to both an active controller and a standby controller of the multi-controller system.

Abstract: A ring network system using peripheral component interconnect express (PCIe) is disclosed. The ring network system includes N PCIe bridges. Each of the N PCIe bridges is connected to an individual server and includes a first port and a second port. The second port of an ith PCIe bridge of the N PCIe bridges is connected to the first port of an ((i mod N)+1)th PCIe bridge of the N PCIe bridges. Each of the N PCIe bridges includes an address mapping chip. The address mapping chip of each of the N PCIe bridges configurably maps to a system address of each of at least portion of N servers connected by the N PCIe bridges and configurably maps to an address of each of at least portion of the N PCIe bridges for setting up a mapping relationship between the N PCIe bridges.

Abstract: An apparatus is described. The apparatus includes a point-to-point link interface circuit. The point-to-point link interface circuit is to support communication with a level of a multi-level system memory. The point-to-point link interface circuit includes a circuit to interlace payload data with cyclic redundancy check (CRC) values, where, different data segments of the payload are each appended with its own respective CRC value.

Abstract: A computer-implemented method for preventing split-brain scenarios in high-availability clusters may include (1) detecting, at a first node of a high-availability cluster, a partitioning event that isolates the first node from a second node of the high-availability cluster, (2) broadcasting, from a health-status server and after the partitioning event has occurred, a cluster-health message to the first node that includes at least a health status of the second node that is based on whether the health-status server received a node-health message from the second node, and (3) reacting, at the first node and based at least in part on whether the first node received the cluster-health message, to the partitioning event such that the partitioning event does not result in a split-brain scenario within the high-availability cluster. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A semiconductor device includes a data output selection circuit suitable for outputting first pattern data as selection data in the case where a training operation is performed, outputting information data as the selection data in the case where a mode register read operation is performed, and outputting second pattern data in the case where the training operation is performed; and a data pad circuit including a first data pad and a second data pad, wherein the first data pad outputs the selection data and the second data pad outputs the second pattern data.

Abstract: For a network control system that receives, from a user, logical datapath sets that logically express desired forwarding behaviors that are to be implemented by a set of managed switching elements, a controller for managing several managed switching elements that forward data in a network that includes the managed switching elements is described. The controller includes a set of modules for detecting a change in one or more managed switching elements and for updating logical datapath set based on the detected change. The logical datapath set is for subsequent translation into a set of physical forwarding behaviors of the managed switching elements.

Abstract: A method is disclosed for performing actions in a computing device based on sensor data from remote devices. While in a power-savings mode, the computing device monitors for activity associated with a mobile device. Based on the monitoring for activity data, the computing device receives an indication that a user of the mobile device intends to use the computing device. Based on the indication, the computing device switches from the power-saving mode to a warming mode, the warming mode comprising restoring power to the one or more components and initiating loading the operational state of the computing device before the user physically interacts with the computing device.

Abstract: A computer-implemented method includes receiving, from a first log agent, a first log collection. The computer-implemented method further includes receiving a first policy, wherein the first policy includes a definition of a first pattern and a definition of a procedure. The computer-implemented method further includes scanning the first log collection against the first policy to determine a match between a portion of the first log collection and the first pattern, with the matching portion of the first log collection being identified as a first data artifact. The computer-implemented method further includes, responsive to identifying the first data artifact, executing the procedure defined by the first policy, wherein the procedure includes: filtering the first log collection to yield a first group of filtered log entries, receiving a first data collection, and sending the first group of filtered log entries and the first data collection to a recipient system.

Abstract: A network system of the present invention includes a computer and a device connected via a network, and a system management device. The computer and the device include, respectively, bridges that encapsulate transmission/reception data transmitted and received to and from each other and transmit and receive the data to and from each other via the network. Each of the bridges includes a control data transmitting means for generating control data for controlling the state of the system based on control auxiliary data issued from the computer or the device and used for controlling the state of the system, and transmitting the control data to the system management device via the network. The system management device includes a system controlling means for controlling the state of the system in accordance with the control data received thereby.

Abstract: Apparatus adapted for exascale computers are disclosed. The apparatus includes, but is not limited to at least one of: a system, data processor chip (DPC), Landing module (LM), chips including LM, anticipator chips, simultaneous multi-processor (SMP) cores, SMP channel (SMPC) cores, channels, bundles of channels, printed circuit boards (PCB) including bundles, floating point adders, accumulation managers, QUAD Link Anticipating Memory (QUADLAM), communication networks extended by coupling links of QUADLAM, log 2 calculators, exp2 calculators, log ALU, Non-Linear Accelerator (NLA), and stairways. Methods of algorithm and program development, verification and debugging are also disclosed. Collectively, embodiments of these elements disclose a class of supercomputers that obsolete Amdahl's Law, providing cabinets of petaflop performance and systems that may meet or exceed an exaflop of performance for Block LU Decomposition (Linpack).

Abstract: Techniques are described for recommending updates to a customer instance in a software as a service (SaaS) model. A request can be received to upgrade features belonging to the customer instance. In response to receiving the request, a feature configuration that corresponds to the customer instance can be identified. The feature configuration can a plurality of features from the SaaS model that are available to the customer interface. Once the feature configuration has been identified, a feature upgrade from a plurality of feature upgrades in an upgrade library can be determined to be applicable to the customer instance. The determination can be made by evaluating the upgrade library and the feature configuration.

Abstract: A fleet rebuild service examines hosts in a fleet to determine whether any of the hosts in the fleet are to be rebuilt. If a host is to be rebuilt, the fleet rebuild service moves the host to a vetting pool. The fleet rebuild service, or another service, may cause automated testing to be performed on the hardware and/or software of hosts in the vetting pool. If a host passes the automated testing, the fleet rebuild service may move the host from the vetting pool to a provisioning pool. If a host does not pass the automated testing, the host may be moved from the vetting pool to a diagnostics pool for further testing. The fleet rebuild service may select hosts from the provisioning pool for automated configuration and deployment to the fleet.

Abstract: To select a master controller from a plurality of controllers, a computing system may consider whether connections between components within field replaceable units (FRUs) are functional as well as the data dependencies between the FRUs. For example, in addition to identifying whether the connections between the FRUs and the controllers are functional, the computing system sends instructions to each of the FRUs coupled to the controllers to determine if inter- and intra-FRU connections required for booting a compute node hosting the FRUs are functional. For instance, one FRU may depend on data from another FRU in order to boot. If a communication link between the FRUs is non-functional, the compute node may fail to boot. The computer system uses these evaluations to determine which of the controllers to select as the master.

Abstract: A method for improving a global lifespan of a storage system of a computing system is provided. The method includes: providing a prediction engine associated with the storage system including at least one storage device, the prediction engine being initiated at an installation time of the at least one storage device; automatically determining lifespan values of the at least one storage device to assign the global lifespan of the storage system, the global lifespan being dependent on the installation time of the at least one storage device; replacing, responsive to a time-based failure event, a storage device of the at least one storage device; and subsequently monitoring, using the prediction engine, the global lifespan of the storage system based on the time-based failure event to define, at least in part, an optimized lifespan of the storage system.

Abstract: A processing system includes a memory module that includes a register space for storing a plurality of register data in a plurality of registers and secure access data corresponding to the register space. A register arbitration module operates to receive a request to access one of the registers from a client module; retrieve secure access data corresponding to the client to determine if the client is trusted; and to grant the request to access the register if the client is trusted. If the client is not trusted, the register arbitration module retrieves secure access data to determine if the register is non-secured for the client. The register arbitration module grants the request to access the register when the register is non-secured for the client.

Abstract: A communication management apparatus includes a token-circulation-order-information storing unit that stores token circulation order, a token-frame processing unit that determines whether transmission right acquiring apparatus information of the token frame indicates the own apparatus and transmits a token frame in which first transmission right acquisition determination information indicating a sequence number of a communication apparatus that can acquire a transmission right in the token circulation order, second transmission right acquisition determination information indicating the number of frames that can be transmitted during one token frame circulation, and transmission right acquiring apparatus information indicating the next transmission destination of the token frame after the own apparatus acquired from the token circulation order are set, and a data-frame-communication processing unit that transmits and receives the data frame, thereby to reduce time required until a communication node that acqui

Abstract: A method, apparatus and computer-implemented method, the method comprising: receiving a statechart comprising a complex feature; and replacing the complex feature with a transformed feature, thereby transforming the statechart to a second statechart, wherein replacing the complex feature comprises: creating an auxiliary variable or a default state; changing a value of the auxiliary variable at the beginning of the transformed feature and changing the value of the auxiliary variable again at its end; and taking a transition from the default state, such that the transition occurs in accordance with the value of the auxiliary variable.

Abstract: A content management server comprises: a first connection detection unit that detects whether or not the content management server is connected with the content-using device; a content management unit that manages the content according to a result of the detection by the first connection detection unit and holds information on how the content-using device controls the use of the content. A content-using device comprises: a second connection detection unit that detects whether or not the content-using device is connected with the content management server; a second selection unit that specifies a method for controlling the use of the content according to a result of the detection by the second connection detection unit; and a content control unit that controls the use of the content according to the method specified by the second selection unit.

Abstract: A technique may improve LDP-IGP synchronization in a MPLS network when an interface of a router coupled to a communications link along a shortest path between a source router and a destination router becomes newly active. Specifically, an IGP adjacency may be held down over the newly active link while a LDP adjacency is established over the link. The router may inject a temporary, non-redistributable route having a large administrative weight into its routing table utilizing addresses learned during establishment of the LDP adjacency. The injected route ensures that there is a reachable route to a peer router so that LDP may converge. After LDP has converged and all needed label bindings have been distributed in the MPLS network utilizing the injected route, IGP executing on the router may advertise a normal cost metric for the link associated with the newly active interface, and IGP convergence may be completed.

Abstract: Methods, apparatus, and systems for implementing a link layer retry protocol utilizing implicit ACKnowledgements (ACKs). Peer link interfaces are configured to facilitate confirmed error-free delivery of link-layer packets through use of implicit ACKs, while also providing retransmission of packets for which errors are detected and guaranteeing the link control data is either successfully received or data transfer over the link is prevented. In conjunction with transmitting packets, reliable packets are copied into sequential slots in a replay buffer. Each link interface tracks the slot at which each reliable packet is buffered, and in response to detection of an error, a retry request is sent to the transmit-side to retransmit the errant packet. The previously buffered copy of the errant packet is retrieved from the replay buffer and retransmitted.

Abstract: Methods, computer systems, and computer program products are provided for managing a storage network system is provided. The storage network system includes a plurality of zones. Each of the plurality of zones includes at least one storage network device. A link-down event associated with one of the storage network devices is detected. One of the plurality of zones is identified is identified as being unused if the zone is not accessed by another of the storage network devices in another of the plurality of zones and if the zone does not access another of the storage network devices in another of the plurality of zones.

Abstract: A method and an apparatus for data transfer in a network having at least two data processing units that exchange messages via the network are described, the exchanged messages having a logical structure in accordance with the CAN specification ISO 11898-1, coding of the bits for at least one first predefinable region within the exchanged messages being accomplished according to the method in accordance with the CAN standard ISO-11898-1, and such that when a switchover condition exists, coding of the bits for at least one second predefinable region within the exchanged messages is accomplished according to a method departing from the CAN standard ISO 11898-1.

Abstract: A device and method for routing separate data flows to subscribers on a content distribution network are provided. The device includes a group of access modules having signal processing circuitry each coupled to a corresponding interface unit. The interface units are coupled in a daisy chain configuration. The access modules include at least first and second active access modules and a spare access module. The group being configured for normal mode operation in which the first active access module processes a first data flow and the second active access module processes a second data flow, and the group being configured for sparing event mode operation in which the first data flow is automatically switched to the second active access module and the second data flow is automatically switched to the spare access module.

Abstract: A hybrid STP can be implemented in a hybrid communication network to prevent communication loops without disabling network interfaces of hybrid devices and redundant paths of the hybrid communication network, while preserving the advantages of the hybrid communication network. Each network interface of a hybrid device can be designated as a relay interface or a non-relay interface. Each of the relay interfaces is independent of the other network interfaces of the hybrid device. The non-relay interfaces operate as a single virtual interface and do not forward packets between each other. Each network interface of the hybrid device may also be associated with a forwarding bit that indicates whether broadcast packets should be forwarded via the network interface. The hybrid device can determine whether/how to forward a packet depending on whether an ingress interface is a relay or non-relay interface and a forwarding bit value of the ingress interface.