摘要:
A method is intended for distributing a frequency reference to network clocks (NC 1 -NC 5 ), comprising a synchronization card (SC) and an interface card (IC), and belonging to a mobile network (MN) having a physical-layer-based protocol for distributing the frequency reference to the network clocks according to a chosen topology, and a packet-based protocol for distributing this frequency reference to the network clocks respecting this chosen topology. This method comprises a step (i) during which one detects a synchronization failure in a synchronization card (SC) of a failed network clock (NC 2 ) connected to neighbour network clocks (NC 1 ,NC 3 -NC 5 ), a step (ii) during which one activates a mechanism of the packet-based protocol between a master neighbour network clock (NC 1 ) and slave neighbour network clocks (NC 3 -NC 5 ), and a step (iii) during which one uses the activated protocol to transmit the frequency reference recovered by the master (NC 1 ) to the slaves (NC 3 -NC 5 ) by following the topology.
摘要:
The invention is directed to a clock module and method for distributing a time reference to at least one clock in a packet-switched network. The clock module includes a slave port, a master port and a local clock. The method comprises the steps of receiving a first synchronization packet at the slave port, the first synchronization packet comprising a first master clock timestamp and generating at least one internal signal comprising the first master clock timestamp. The method also includes the steps of transmitting the at least one internal signal to the master port and receiving the at least one internal signal at the master port. Then a method includes determining the internal propagation time of the signal through the clock module and generating a second synchronization packet at the master port comprising a second master clock timestamp, the second master clock timestamp comprising the sum of the first master clock timestamp and the internal propagation time. Finally, the second synchronization packet is sent to at least one other clock in the packet-switched network.
摘要:
The invention relates to a method for enabling Service Level Agreement, SLA, negotiation, in a data center (1) comprising, at a first data center site, overlay and underlay network elements, the overlay network elements including a plurality of virtual machines (2), the underlay network elements including a gateway node (4). The data center (1) is connected to a provider edge node (7) of a carrier network (5) via the gateway node (4) through a User-Network Interface, UNI, (6) between the gateway node (4) and the provider edge node (7). The negotiation of the SLA between the first data center site and the carrier network (5) is controlled by a UNI control function. Egress traffic flows in the overlay network from at least one of the plurality of virtual machines (2) are encapsulated into a virtual network tunnel (8) by a virtual network control function, the virtual network tunnel (8) crossing the gateway node (4) and the UNI (6). The method comprises: - dynamically communicating required Quality of Service, QoS, parameters detained by the virtual network control function to the UNI control function for an egress traffic flow within the virtual network tunnel (8); and - applying the required QoS parameters in the SLA negotiation for the egress traffic flow.
摘要:
A method for centralized address resolution in a data center comprising at least one processing resource, at least one processing resource controller, at least one networking resource, and at least one networking resource software-defined network controller, comprises the steps of: - maintaining a local address resolution database (LDB1) associated to a processing resource controller, for storing all the MAC address - IP address associations of the processing resources controlled by this controller; - maintaining a local address resolution database (LDB2) associated to a networking resource software-defined network controller, for storing all the MAC address - IP address associations of the networking resources controlled by this controller; - maintaining a centralized address resolution database (CARDB) for storing all the MAC address - IP address associations within the data center; - when a resource requests the resolution of an IP address, sending a unicast read message from this resource to its controller, this controller checking whether the associated local database (LDB1, LDB2) stores a MAC address associated to this IP address, and: -- if a MAC address, associated to this IP address MAC address, is stored in this local database, then supplying the MAC address to the resource that requested the resolution of this IP address; -- if no MAC address, associated to this IP address MAC address, is stored in this local database, then forwarding the unicast read message to the centralized address resolution database (CARDB).
摘要:
The method comprises the steps of: - sending a Sync message (SYNC) from a first peer-to-peer transparent clock (P2P TC B) to a second peer-to-peer transparent clock (P2P TC C), - estimating the path delay of the transmission path travelled by the synchronization message from the first to the second peer-to-peer transparent clocks, - and then taking this path delay into account for updating the time information carried by a synchronization message. For estimating said path delay, it comprises the steps of: - creating a list of the network addresses of the network interfaces (IPRB, IPRI1, IPRI2, IPRC) traversed by the synchronization message (SYNC); - ordering the first list into the order in which the network interfaces have been traversed by the Sync message (SYNC); - creating a second list by reversing the order of the first list; - communicating the second list to the second peer-to-peer transparent clock (P2P TC C); - and using the mechanism available at the transport protocol level, to constrain the respective paths of Pdelay_Req and Pdelay_Resp messages so that their respective paths map to the second and first ordered lists of traversed interfaces.
摘要:
La présente invention concerne un procédé de sélection de ressources réseaux de bordure inter-domaines d'un réseau de communication, un domaine étant sous la responsabilité d'un opérateur unique et comprenant une pluralité de noeuds dont au moins un noeud de bordure, les domaines étant divisés en au moins trois niveaux dont un niveau de service, un niveau de supervision et un niveau réseau. Le procédé comprenant les étapes de : - au niveau de service : - sélectionner au moins une séquence d'éléments de service, ladite séquence satisfaisant une demande de service d'un utilisateur d'un domaine, un élément de service étant associé à un domaine et comprenant des données de qualité de service et une référence vers une pluralité de labels de service - pour chaque domaine participant à ladite séquence : - sélectionner des labels de service correspondant aux éléments de service de ladite séquence, un label de service comprenant les données primaires suivantes : - un identifiant de label de service ; - au moins un identifiant d'un noeud de bordure qui supporte le label de service ; - au moins un identifiant d'un lien d'adjacence entre ledit noeud de bordure et un noeud de bordure adjacent ;
- effectuer une corrélation entre chacun des labels de service du domaine et les labels de service d'au moins un domaine adjacent interconnecté en fonction des données primaires des labels de service de sorte à obtenir une liste de labels de service ; - sélectionner un label de service parmi ladite liste ; - à un niveau différent du niveau de service, pour chaque domaine participant à ladite séquence, sélectionner un identifiant de ressources correspondant, ledit identifiant de ressources comprenant des données réseaux relatives à des ressources réseaux de bordure inter-domaines supportées par le noeud de bordure qui supporte le label de service sélectionné ; et - au niveau réseau, pour chaque domaine participant à ladite séquence, vérifier la disponibilité des ressources réseaux de bordure inter-domaines définies dans l'identifiant de ressources sélectionné et réserver lesdites ressources réseaux de bordure inter-domaines si elles sont disponibles.