Abstract: Updating a machine learning model for semantic segmentation by receiving an image and a segment ID mask, the image having a plurality of pixels, and the segment ID mask containing for each pixel of the image one of a plurality of segment identifiers identifying a segment to which the pixel belongs, feeding the image to a machine learning model and computing an outcome, computing a per-segment value for each one of the plurality of segment identifiers from the outcome and from pixels associated with the one of the plurality of segment identifiers, aggregating a plurality of per-segment values to create a segment-aware value and using the segment-aware value for training, validating and testing the machine learning model.

Abstract: A system for identifying one or more sub-fields within a field of a given field type, the field comprising a plurality of field bits, and each of the sub-fields comprising a distinct sub-group of the field bits having one or more fixed consecutive bit positions within each instance of a plurality of instances of the field, the system including a processing resource configured to: obtain one or more input time-series, each input time-series of the one or more input time-series including a plurality of input instances of the field; generate one or more new time-series, each new time-series of the generated new time-series being associated with at least one input time-series of the one or more input time-series and including a plurality of new instances of the field; generate a bit-field based on selected instances of the new instances in each new time-series of the generated new time-series; and analyze the bit-field to identify the fixed consecutive bit positions of at least one of the sub-fields within the fi

Abstract: A key having a gripping portion, the key is left protruding so that torque can be applied by the key user. The gripping portion includes a gripping portion structure. The gripping portion structure further includes an electronic device having a plug connectable to a plug receptacle of electronic appliance. The electronic device integrated in the gripping portion structure. Along at least one side of the gripping portion structure disposed an elongated slot which is wide enough to be used as a key ring receptacle where the key ring, can pass through. The key gripping portion have a freedom of movement for the key user to move the key around the key ring and also a freedom of movement to cross the key ring to either direction, inwardly or outwardly for enabling comfortable access of the plug of the electronic device to be connected to the plug receptacle of electronic appliance.

Abstract: A key having a gripping portion, the key is left protruding so that torque can be applied by the key user. The gripping portion includes a gripping portion structure. The gripping portion structure further includes an electronic device having a plug connectable to a plug receptacle of electronic appliance. The electronic device integrated in the gripping portion structure. Along at least one side of the gripping portion structure disposed an elongated slot which is wide enough to be used as a key ring receptacle where the key ring, can pass through. The key gripping portion have a freedom of movement for the key user to move the key around the key ring and also a freedom of movement to cross the key ring to either direction, inwardly or outwardly for enabling comfortable access of the plug of the electronic device to be connected to the plug receptacle of electronic appliance.

Abstract: A novel and useful hitless node insertion mechanism for configuring a new node to be inserted into an Ethernet network. The mechanism comprises pre-configuring the new node into a partially configured state, disconnecting a network link where the new node is to be inserted and connecting the new node to at least one end of the disconnected link. All data traffic received from neighboring nodes is then blocked and fake signal failure (SF) messages are generated and sent if Ethernet Ring Protection (ERP) messages are detected. An in-band management port is then selected and a configuration from a Network Management System (NMS) is received. Once configuration of the said new node is complete, the new node is set to a fully configured state.

Abstract: A novel and useful one-to-many OAM/protection inter-work function (IWF). The mechanism stitches a single OAM session (e.g., a CCM session) on a multi-traffic class (TC) M-domain to multiple OAM-sessions (which monitor different transport entities or connections) on a single-TC S-domain using duplication and filtering of OAM messages at the ENNI. Using the OAM sessions, protection decision manipulation is performed to coordinate the protection decisions of the different connections serving the different TCs of the same S-domain endpoint. The manipulation includes manipulating fields in one OAM-session as a function of information gained from an OAM-session monitoring a different transport entity such that the first connection is forced to follow protection decisions based on the second OAM-session. Bundled single-TC connections are forced to take protection decisions that steer user traffic through transport entities provisioned in advance to share the same fate in the event of link or equipment failures.

Abstract: A method of connecting a client to a server by a load balancer associated with a plurality of servers. The method includes establishing a first connection for transmission of packets between the load balancer and the client, selecting a server to service the client, and splicing the first connection with a second connection between the load balancer and the selected server, the second connection being established before the first connection.

Abstract: A novel and useful hitless node insertion mechanism for configuring a new node to be inserted into an Ethernet network. The mechanism comprises pre-configuring the new node into a partially configured state, disconnecting a network link where the new node is to be inserted and connecting the new node to at least one end of the disconnected link. All data traffic received from neighboring nodes is then blocked and fake signal failure (SF) messages are generated and sent if Ethernet Ring Protection (ERP) messages are detected. An in-band management port is then selected and a configuration from a Network Management System (NMS) is received. Once configuration of the said new node is complete, the new node is set to a fully configured state.

Abstract: An E-Tree service interoperability mechanism between VPLS domain devices (e.g., MPLS domain devices) and Ethernet domain devices. E-Tree interoperability functionality is provided whereby the E-domain device directly connected to the VPLS device is modified to perform an asymmetric VLAN tag manipulation on traffic forwarded between the VPLS device and itself. The capabilities of VPLS are used to divide between roots and leaves, even if both exist in the same E-domain, so that they do not share VLANs resulting in preventing roots and leaves in the same E-domain from communicating directly, but rather through the VPLS devices to which the E-domain connects. Traffic on the E-domain is segregated into a root VLAN to which roots are connected, a root-to-leaf VLAN for forwarding root-originated traffic from the VPLS-domain to the leafs, and a leaf-to-root VLAN for handling traffic originated by the leafs destined to roots.

Abstract: A novel and useful system and method for fast protection of dual homed Virtual Private LAN Service (VPLS) spokes. Fast protection is achieved by forwarding traffic to the dual homed VPLS spoke from the VSI that is not currently the one currently serving it, as long as the traffic is known unicast. Egress traffic at the VSI that does not currently serve the VPLS-spoke is re-routed over a protection transport entity to the VPLS-spoke through the device of the VSI currently serving it. Broadcast, multicast and unknown (BMU) traffic is sent to the VPLS spoke only by the BMU-primary VSI. BMU frames received at the BMU-secondary VSI are dropped. A fast election process serves to elect the BMU-primary for coordinating between the two VSIs serving the same VPLS spoke. In addition, a flag is inserted in each frame at an ingress VSI to indicate whether the frame is known or unknown unicast. Unknown unicast frames are forwarded to the VPLS spoke only by the BMU-primary.

Abstract: A method of determining local multicast information of a local area network (LAN), comprising dividing the LAN to a number of segments larger than the number of virtual LANs (VLANs) in the network and creating a layer-3 multicast routing table, which relates to each of the segments separately.

Abstract: A method of accelerating the operation of a load balancer by an accelerator switch. The method includes receiving, by the accelerator switch, packets directed to the load balancer, determining, for at least some of the received packets, whether the packets match an entry of a list of packet groups, by comparing fewer than five parameters of the packets to respective fields of entries of the list, and forwarding, by the accelerator switch, at least some of the received packets, directly to their destination, responsive to the determining.

Abstract: A novel mechanism for performing policy aware frame loss measurements that overcomes the problems of the prior art. The mechanism permits OAM frames to travel through the network in the same way data frames travel marked as green or yellow. The frame loss measurement is policy aware (i.e. SLA aware), taking CIR and EIR policy as well as policy in the middle of the network into consideration. In one embodiment, the value of the policer at the ingress to the client side is read just before sending an OAM frame and a potential decision value is calculated which reflects the decision that the policer would have taken if given the OAM frame at that time. If the calculated decision is to send the OAM frame as green or yellow, the frame is sent as green or yellow, accordingly. If, however, the calculated decision is to drop the frame, the frame is randomly marked as either green or yellow in accordance with the ratio of CIR to EIR. In another embodiment, red traffic (i.e.

Abstract: An apparatus for and method of packet loss measurement for TLS connections in an MEN that overcomes the problems of the prior art. The mechanism is operative to ensure that only one copy of each received packet is counted when exiting the TLS connection (unless dropped along the path due to congestion, etc.), so that the number of egress packets counted can be accurately compared with the number of ingress packets counted. Only a single copy of each packet is marked, i.e. flagged, at ingress. Any bridging along the path that needs to duplicate the packet forwards only a single marked copy of the packet. All other copies are forwarded unmarked. At the egress from the TLS, only marked packets are counted by the egress counter. In a second embodiment, a duplication field is inserted into each packet to better track all the duplicate copies of a packet.

Abstract: A novel fast and scalable protection mechanism for protecting hierarchical multicast service in ring based networks. The mechanism of the present invention is especially suitable for use in Multi-Protocol Label Switching (MPLS) ring based networks such as Metro Ethernet Networks (MENs). The mechanism provides fast protection for MPLS based point-to-multipoint (P2MP) Label Switched Paths (LSPs) in a scalable manner. Each multicast connection on each ring in the network is split into two sub-LSPs: a primary P2MP sub-LSP originating on a primary node and a secondary P2MP sub-LSP originating on a secondary node traveling opposite to the primary path. For each node to be protected, a point-to-point protection tunnel is provisioned from that node to a secondary node that forwards the packets to the secondary path on all child rings connected to that parent ring through the protected node and that are provisioned to receive the specific multicast connection.

Abstract: An apparatus for and method of enforcing precedence of committed traffic over excess traffic that overcomes the difficulty in supporting committed over excess traffic and has particular application to distributed queuing systems. Queue control messages are generated and transmitted to other line cards to guarantee that committed traffic is not dropped when sent from multiple sources such as line interface cards to a single destination. When the level of a queue exceeds a threshold and flow control is received for the corresponding destination, the queue starts dropping excess traffic and messages indicating the queue status are sent to all other lines cards. When a line card receives a queue control message that another line card started dropping excess traffic to a specific destination, it also starts dropping excess traffic to that destination as well.

Abstract: A subport forwarding and provisioning mechanism whereby a plurality of subports implemented using slower speed processors are used to perform the packet processing for a higher speed packet stream. Outbound packets are assigned a subport based on their MPLS labels. Once assigned, the packets are forwarded to a particular subport. The output of all the subports is aggregated and combined into a single high speed packet stream and transmitted over the link. Inbound packets are received by the high speed PHY and forwarded to one of a plurality of subports. The subport chosen is the one having the same subport index as the one through which the packet was sent out over at the transmitting node. Conventional layer 2 switches can also be used whereby reserved source MAC addresses are assigned to subports. The switch forwards the inbound packets to one of the subports in accordance with the packet's source MAC address.

Abstract: A protection mechanism capable of providing both local and end-to-end connection protection for dual homed access/aggregation devices or customer-edge devices in a network. The protection mechanism provides end-to-end and fast local protection for off the shelf access devices that do not have any built in per-connection protection capabilities. The access device is connected via two separate physical uplinks to two edge switches of the network. For each connection to be protected, a main path is provisioned from one edge switch and an alternative path is provisioned from the other edge switch. The edge switches are adapted to comprise means for switching traffic from the main path to the alternative path in the event a failure along the main path is detected. Failures both in the stack portion, the core portion and in the access device uplinks are protected against.

Abstract: A novel mechanism for performing policy aware frame loss measurements that overcomes the problems of the prior art. The mechanism permits OAM frames to travel through the network in the same way data frames travel marked as green or yellow. The frame loss measurement is policy aware (i.e. SLA aware), taking CIR and EIR policy as well as policy in the middle of the network into consideration. In one embodiment, the value of the policer at the ingress to the client side is read just before sending an OAM frame and a potential decision value is calculated which reflects the decision that the policer would have taken if given the OAM frame at that time. If the calculated decision is to send the OAM frame as green or yellow, the frame is sent as green or yellow, accordingly. If, however, the calculated decision to drop the frame, the frame is randomly marked as either green or yellow in accordance with the ratio of CIR to EIR. In another embodiment, red traffic (i.e.

Abstract: An end-to-end monitoring mechanism whereby edge nodes are notified that a particular path flows through a local protection tunnel along the way. The notification mechanism utilizes OAM packets which include a link protection indication that are transmitted on the Ethernet service layer. The link protection indication may be represented by one or more flags to indicate that local protection is in place along the path. The mechanism enables fast local protection to be used with slower end-to-end protection by informing the latter when local protection is activated. In an implementation in which local-protection does not preserve the CIR of the protected traffic, the invention provides minimal interruption of Committed Information Rate (CIR) connections by minimizing the time traffic is diverted through local protection tunnels. Notification of the edge node that local protection is in use occurs very quickly, thus allowing the edge node to rapidly switch to a backup path.