Abstract: High accuracy timing over packet networks is achieved by generating correction factors from multiple separation intervals and timing information contained in packets in both directions between a master and a slave. The methods are based on evaluating the weighted average of short-term, medium-term, and long-term measurements of local clock offset. Weighted averages are used to develop robust correction terms that are modified with an ?-shaping factor to provide additional immunity to packet network instabilities.

Abstract: A clock at a first network element that is connected to a second network element over an optical fiber link is aligned using bursts of timing information exchanged between the two network elements. According to one method, the bursts from the first network element to the second network element and the bursts from the second network element to the first network element are transmitted over the same wavelength channel of the optical fiber link, in which case zero asymmetry in the transit delays can be assumed during the alignment procedure. According to another method, the bursts from the first network element to the second network element and the bursts from the second network element to the first network element are transmitted over different wavelength channels of the optical fiber link, in which case the asymmetry in the transit delays can be quantified and applied during the alignment procedure.

Abstract: Embodiments of the present invention set forth a method and system for reducing uncertainty in receive and transmit timestamps created by an NTP server. The uncertainty in the receive timestamps is removed by recording the time-of-arrival in the hardware clock of the NTP server before the incoming packets may be delayed by traversing the various layers of software in a timestamping system. The uncertainty in the transmit timestamps is removed by giving the outgoing packets a timestamp in the future using an estimate of the transmission latency calculated by the latency estimator filter. Subsequently, the actual time-of-departure is used to re-calculate and update the estimate of the transmission latency. In this fashion, superior control of the timestamping function may be implemented in existing NTP servers in a manner that retains interworking compatibility with the current NTP standards.

Abstract: Routes of a packet network are analyzed according to various transit delay metrics. Preferred packet network routes are selected between source and destination based on these metrics. In packet networks employing boundary clocks and transparent clocks, faulty boundary clocks and faulty transparent clocks are identified using the metrics.

Abstract: A clock at a first network element that is connected to a second network element over first and second optical links that are physically distinct from each other is aligned using optical timing signals having different wavelengths. Transit delays between the first and second network elements may be determined using the same optical timing signals.

Abstract: A latency floor between two nodes of a packet-switched network is estimated using transit times of a group of packets traversing the two nodes. In particular, a periodically generated histogram of packet transit times is used to estimate the latency floor. In some packet-switched networks, the behavior of some network elements changes drastically when the network is congested. Because latency floor cannot be accurately estimated under such conditions, packet transit times collected during a congested state of the network are discarded.

Abstract: Packet network performance is assessed using transit delay metrics and compliance masks generated at various evaluation nodes of the network. The evaluation nodes may employ network probes that make precise measurements of transit delays and thereby of transit delay variations. Based on the assessments, a master may be added to the network or relocated within the network, rate of timing packets generated by the master may be adjusted up or down, or oscillators used at the slaves may be upgraded.

Abstract: A packet probe for a packet network accurately generates and monitors packets within the network. The packet probe supports packet generation and packet transmission. When a packet is ready for transmission, a hardware-based time stamp unit affixes a time stamp to the packet reflecting an actual transmission time. The packet probe also supports receiving, filtering, and time stamping received packets. When a packet is received, a packet filter determines whether the received packet should be stored in memory along with a time stamp reflecting an actual reception time.

Abstract: One embodiment of the present invention sets forth a method for autonomously validating the time and frequency data obtained from multiple sources, and generating a suitable estimate of the frequency difference between the client clock and the source. The method includes the steps of protocol data unit validation, offset measurement, minimum offset filtering, and frequency filtering. With these steps, the negative effects of packet delay variation may be mitigated and a frequency estimate is determined for the source in question, together with an associated validity status. Consequently, quality control of the local clock is achieved in packet networks at significantly reduced cost and decreased level of complexity relative to prior art approaches.

Abstract: Methods and apparatus for a play-out buffer that may adjust offsets between clocks of two ends of a network link with an adaptive play-out buffer and adaptive clock control. The play-out buffer is a circular jitter buffer that permits the absorption of a frequency offset using controlled slips between two nodes of a network. The play-out buffer also accommodates some wander introduced by the time-delay variation across the network. The adaptive clock control reduces the frequency offset between the clocks of the two nodes. In this manner, even though some offsets between two nodes would render communication inefficient, embodiments of the present invention allow the effects of these offsets to be mitigated, thus providing for a better quality coupling.

Abstract: The invention comprises a method and apparatus for reducing uncertainty in timing on the network. The uncertainty in receive buffers is removed by time stamping the arriving packets before sending the packets to the receive buffer. The uncertainty in the transmission buffer is removed by giving the packets a timestamp in the future, and holding the packets until precisely that time. Time precision is ensured by only releasing time packets at the host physical layer to network boundary at the time specified within the packet.

Abstract: Systems and methods are described for supporting synchronization status messages (SSM) on building integrated timing system synchronization supply unit (BITS-SSU) remote shelves. A method includes providing a composite clock reference to a remote synchronization supply unit from a master synchronization supply unit; providing a synchronization status message to the remote synchronization supply unit; relaying the composite clock reference from the remote synchronization supply unit; and relaying the synchronization status message from the remote synchronization supply unit.

Abstract: Methods and apparatus are described for a play-out buffer. A method includes writing a data packet into a jitter buffer at a write address specified by a write address generator; incrementing the write address generator; generating the difference between the write address and a current read address specified by a read address generator; reading a data packet from the jitter buffer from the current read address specified by the read address generator; generating a new read address based on the difference between the write address and the current read address by the read address generator. An apparatus includes a jitter buffer; a write address generator for storing a write address; a read address generator for storing a current read address; a read address increment control; wherein the read address increment control sets the future read address based on the difference between the write address and the current read address.

Abstract: The variable latency associated with standard network forwarding devices is eliminated by forwarding timing packets through a network forwarding device with a constant delay. The network forwarding device of the invention time stamps timing packets that are received by the input ports with a predefined future time, and then outputs the timing packets from the output ports at the predefined time so that the transmission delay through the network forwarding device is the same from any port to any other port.

Abstract: A method to determine real time image complexity in video streaming, IPTV and broadcast applications using a statistical model representing channel bandwidth variation and image complexity that considers scene content changes. Available channel bandwidth is distributed unevenly among multiple video streams in proportion to bandwidth variation and image complexity of the broadcast video stream. The distribution of available channel bandwidth is determined based upon an image complexity factor of each video stream as determined from probability matrices considering bandwidth variations and image complexity.

Abstract: Methods and apparatus are described for providing multiple transmission bandwidth streams with differentiated quality of service on a inter-machine trunk. One approach includes time-division multiplexing. Another approach includes statistical multiplexing. Another approach includes packet segmentation. The approaches are commercially important because they significantly reduce time-delay variation.

Abstract: A two-way time transfer protocol includes: sending a signal from a first node including a first clock and a first time interval counter coupled to the first clock over a transport physical layer coupled to the first node to a second node coupled to the transport physical layer, the second node including a second clock and a second time interval counter coupled to the second clock; then sending a last second node time interval counter value from the second time interval counter of the second node over the transport physical layer to the first node; and then comparing at the first node a last first node time interval counter value to the last second node time interval counter value. A first-way path latency from the first node to the second node is substantially equal to a second-way path latency from the second node to the first node, and all of a first node transmit delay, a first node receive delay, a second node transmit delay and a second node receive delay are substantially constant.

Abstract: A method to determine real time image complexity in video streaming, IPTV and broadcast applications using a statistical model representing channel bandwidth variation and image complexity that considers scene content changes. Available channel bandwidth is distributed unevenly among multiple video streams in proportion to bandwidth variation and image complexity of the broadcast video stream. The distribution of available channel bandwidth is determined based upon an image complexity factor of each video stream as determined from probability matrices considering bandwidth variations and image complexity.

Abstract: A method for determining spatial and temporal loss in a packet based video broadcast system in an encrypted environment involves measuring video coding layer information at an unencrypted head end of a video stream and network layer information at an encrypted downstream end of the same video stream. Video coding layer information is correlated with network layer information having a corresponding time stamp to compute the spatial and temporal loss. The video coding layer and network layer information is taken from discrete segments of the video stream including access units, slices or macroblocks. Impairments in the video stream are determined using the computed spatial and temporal loss.

Abstract: Systems and methods are described for distribution of synchronization in a packet switched local area network environment. A method for extracting network synchronization timing from a data transmission burst includes: recovering a clock during the data transmission burst; and then holding over the clock after the data transmission burst ceases. A method for inserting network synchronization timing into a data transmission burst includes encoding data using a time-base reference signal governed clock.