Abstract: According to the present invention, a technique for delivering content over a network having endpoints and nodes that store content is provided. In one embodiment, the technique is realized by first launching a request for content from a requesting end point. Then propagating the request over the network to the nodes. While the request propagates, it leaves a trail at the nodes and, when content matching the request is located, a copy of the content is returned to the requesting endpoint over the trail left by the request.

Abstract: A method and apparatus for controlling congestion at a node in a data network. The node includes an input for receiving traffic units from the network, an output for releasing traffic units to the network and a control unit. The control unit is responsible for estimating a level of data occupancy of at least a portion of the network by looking at the traffic units received at the input from a remote node in the network. When the data occupancy level reaches a certain threshold, the node issues a control signal to the remote node such that the remote node lowers its rate of traffic units input in the network. By estimating the network data occupancy level, congestion at the node can be effectively foreseen and controlled.

Abstract: An apparatus for implementing a collision avoidance mechanism in a bufferless network. The apparatus includes first and second inputs for receiving traffic units from first and second source points, respectively. The traffic units from the first source point arrive at the first input on a basis of a first traffic pattern. The traffic units from the second source point arrive at the second input on a basis of a second traffic pattern. The apparatus further includes a traffic detection unit and a notification unit. The traffic detection unit is operative to monitor the traffic units received at the first and second inputs for detecting the first and second traffic patterns. The notification unit is operative to generate a control signal for transmission to either one of the first and second source points on a basis of the detected first and second traffic patterns.

Abstract: An admission controller and its method of operation for controlling admission of data flows into an aggregate data flow are described. Admitted data flows are aggregated into an aggregate data flow for transmission by a router, for example, over a data network. The aggregate data flow typically follows a pre-established path through the network; for example a multi-protocol label switched (MPLS) path. The path has minimum and maximum bandwidth limits assigned to it. The admission controller controls admission of new data flows into the aggregate data flow by granting or denying new session requests for the new data flows. Congestion notifications received from the network and bandwidth limits of the path are considered in determining whether to grant or deny a new session request. In this way, the admission controller provides elastic sharing of network bandwidth among data flows without exacerbating network congestion and while remaining within the path's bandwidth limits.

Abstract: In accordance with a method and apparatus for managing aggregate flows of data in a differentiated services (DS) network, fairness issues are improved by enforcing congestion control at the edges of the DS in a TCP-friendly manner network, rather than relying on congestion to accumulate at the core. This is done by establishing a credit system for each aggregate flow through the edge. Only if credit is available for the flow, will data from an edge node be forwarded to the edge node's DS traffic conditioner where the data packets can be marked using DiffServ labels. When credit is not available, the data will be queued at the edge node. Each time data belonging to an aggregate flow is forwarded to the traffic conditioner, the amount of credit for the flow is decremented. Credit for the flow is incremented only when control packets can be inserted into the flow. Specifically, the present invention establishes control flows associated with each aggregate flow. A control packet (i.e.

Abstract: According to the present invention, a technique for delivering content over a network having endpoints and nodes that store content is provided. In one embodiment, the technique is realized by first launching a request for content from a requesting end point. Then propagating the request over the network to the nodes. While the request propagates, it leaves a trail at the nodes and, when content matching the request is located, a copy of the content is returned to the requesting endpoint over the trail left by the request.