Showing papers on "Fast packet switching published in 1998"

Policy based quality of service

Abstract: A flexible, policy-based, mechanism for managing, monitoring, and prioritizing traffic within a network and allocating bandwidth to achieve true quality of service (QoS) is provided. According to one aspect of the present invention, a method is provided for managing bandwidth allocation in a network that employs a non-deterministic access protocol, such as an Ethernet network. A packet forwarding device receives information indicative of a set of traffic groups, such as: a MAC address, or IEEE 802.1p priority indicator or 802.1Q frame tag, if the QoS policy is based upon individual station applications; or a physical port if the QoS policy is based purely upon topology. The packet forwarding device additionally receives bandwidth parameters corresponding to the traffic groups. After receiving a packet associated with one of the traffic groups on a first port, the packet forwarding device schedules the packet for transmission from a second port based upon bandwidth parameters corresponding to the traffic group with which the packet is associated. According to another aspect of the present invention, a method is provided for managing bandwidth allocation in a packet forwarding device. The packet forwarding device receives information indicative of a set of traffic groups. The packet forwarding device additionally receives information defining a QoS policy for the traffic groups. After a packet is received by the packet forwarding device, a traffic group with which the packet is associated is identified. Subsequently, rather than relying on an end-to-end signaling protocol for scheduling, the packet is scheduled for transmission based upon the QoS policy for the identified traffic group.

Transparent optical packet switching: the European ACTS KEOPS project approach

Abstract: This paper reviews the work carried out under the European ACTS KEOPS (KEys to Optical Packet Switching) project, centering on the definition, development and assessment of optical packet switching and routing networks capable of providing transparency to the payload bit rate. The adopted approach uses optical packets of fixed duration with low bit rate headers to facilitate processing at the network/node interfaces. The paper concentrates on the networking concepts developed in the KEOPS project through a description of the implementation issues pertinent to optical packet switching nodes and network/node interfacing blocks, and consideration of the network functionalities provided within the optical packet layer. The implementation, from necessity, relies on advanced optoelectronic components specifically developed within the project, which are also briefly described.

Transparent optical packet switching: network architecture and demonstrators in the KEOPS project

Abstract: This paper reviews the work carried out in the ACTS KEOPS (Keys to Optical Packet Switching) project, describing the results obtained to date. The main objective of the project is the definition, development, and assessment of optical packet switching and routing networks, capable of providing transparency to the payload bit rate, using optical packets of fixed duration and low bit rate headers in order to enable easier processing at the network/node interfaces. The feasibility of the KEOPS concept is assessed by modeling, laboratory experiments, and testbed implementation of optical packet switching nodes and network/node interfacing blocks, including a fully equipped demonstrator. The demonstration relies on advanced optoelectronic components, developed within the project, which are described.

Mechanism for dispatching packets via a telecommunications network

Abstract: A mechanism for dispatching a sequence of packets via a telecommunications network includes a queue for packets for transmission and a queue controller responsive to receipt of a new packet for transmission to compare parameters of the new packet to parameters of any packet already in the queue, the queue controller determining whether to queue or drop the new packet depending on the result of the comparison(s). The queue can be implemented as a linked list of packet entries with individual pointers to the respective packets concerned. The queue entries can include details relating to the packet including data relating to the information flow and also the packet identity. In a TCP environment, the flow information can include the source IP address and the source TCP port, as well as the destination IP address and the destination TCP port. The identity information can include sequence numbers and acknowledgement numbers for the packet concerned. In order to optimize network usage, it can be useful to drop some packets at a routing node. A decision to drop a packet can be made if the new packet and a queued packet relate to the same information flow, the new packet sequence number equals the queued packet sequence number and the new packet acknowledgement number is less than the queued packet acknowledgement number. The new packet is dropped where the new packet is a retransmission of a queued packet and the length of the queued packet is greater than or equal to that of the new packet. A queued packet is replaced by a new packet when the new packet is determined to be a retransmission of the queued packet and the length of the new packet is greater than that of the queued packet.

Packet switching method with time-based routing

Abstract: This invention describes a method for transmitting and forwarding packets over a switching network using time information. The network switches maintain a common time reference, which is obtained either from an external source (such as GPS—Global Positioning System) or is generated and distributed internally. The time intervals are arranged in simple periodicity and complex periodicity (like seconds and minutes of a clock). A data packet that arrives to an input port is switched to an output port based on its order or time position in the time interval in which it arrives at the switch. The time interval duration can be longer than the time duration required for transmitting a data packet, in which case the exact position of a data packet in its forwarding time interval is predetermined. This invention provides congestion-free data packet switching for data packets for which capacity in their corresponding forwarding links and time intervals is reserved in advance. Furthermore, such data packets reach their destination, which can be one or more (i.e., multicast) in predefined time intervals, which guarantees that the delay jitter is smaller than or equal to one time interval

Wavelength routing-based photonic packet buffers and their applications in photonic packet switching systems

Abstract: Photonic packet buffers are essential components in photonic packet switching systems. We present a wavelength routing-based photonic packet buffer based on a state-of-the-art arrayed-waveguide grating (AWG) multiplexer. We show how this new packet buffer can be effectively used in the implementation of photonic packet switching systems. We also propose and examine two different photonic packet switch architectures.

Method and apparatus for time efficient retransmission using symbol accumulation

Abstract: An efficient retransmission of data using symbol accumulation wherein the packet received in error is retransmitted at a lower energy-per-bit level concurrently in the same frame with the new packet. The destination device receives the data transmission and retransmission, demodulate the signal, and separates the received data into the new and retransmitted packet. The destination device then accumulates the energy of the retransmitted packet with the energy already accumulated for the packet received in error and decodes the accumulated packet. The accumulation of the additional energy provided by the subsequent retransmissions improves the probability of a correct decoding. The throughput rate can be improved since the packet received in error is retransmitted concurrently with the transmission of the new data packet. The capacity is maximized since the retransmission of the packet received in error is at a lower energy level than that of the new packet.

Switching theory : architectures and performance in broadband ATM networks

Abstract: Broadband Integrated Services Digital Network. Interconnection Networks. Rearrangeable Networks. Non--Blocking Networks. ATM Switch Model. ATM Switching with Minimum--Depth Blocking Networks. ATM Switching with Non--Blocking Single--Queueing Networks. ATM Switching with Non--Blocking Multiple--Queueing Networks. ATM Switching with Arbitrary--Depth Blocking Networks. Appendix. Index.

Method for setting time stamp in syt field of packet headers for ieee-1394 devices

Abstract: Isochronous data packets transmitted within a digital network having a bus architecture that complies with the IEEE-1394 Standard for a High Performance Serial Bus are stamped with a presentation time stamp value determined according to a computed packet rate for the data For the case where the presentation time stamp field of a first packet of a second frame of data for transmission in the digital network is set with the presentation time value, the packet rate may be computed by measuring a difference between a desired presentation time value of a first packet in a first frame of the data and an actual transmission time of the first packet of the first frame The first frame preceding the second frame in time of transmission within the network

Implementing distributed packet fair queueing in a scalable switch architecture

Abstract: To support the Internet's growth, there is a need for cost effective switching technologies that can simultaneously provide high capacity switching and advanced QoS. Unfortunately, these two goals are largely believed to be contradictory in nature. To support QoS, sophisticated packet scheduling algorithms, such as fair queueing, are needed to manage queueing points. However, the bulk of current research in packet scheduling algorithms assumes an output buffered switch architecture, whereas most high performance switches are input buffered. While output buffered systems may have the desired QoS, they lack the necessary scalability. Input buffered systems, while scalable, lack the necessary QoS features. We propose the construction of switching systems that are both input and output buffered with the scalability of input buffered switches and the robust QoS of output buffered switches. We call the resulting architecture distributed packet fair queueing (D-PFQ) as it enables physically dispersed line cards to provide a service that closely approximates an output-buffered switch with fair queueing. By equalizing the growth of the virtual time functions across the switch system, most of the PFQ algorithms in the literature can be properly defined for distributed operation. We present our system using a cross bar for the switch core. Buffering techniques are used to enhance the system's latency tolerance, which enables the use of pipelining and variable packet sizes internally. We evaluate the delay and bandwidth sharing properties.

System for user-space network packet modification

Abstract: A system for user-space packet modification, including a set of kernel code and a user-level application programming interface (API). The system facilitates creation of a special socket for passing packets between kernel space and user space. The system in turn facilitates creation and application of a packet filter associated with the socket, in order to trap incoming or outgoing packets being processed in the kernel at a designated point in a protocol stack. Once a packet is trapped, it is moved through the socket into user space, thereby at least temporarily preventing the protocol stack from further processing the packet. In user space, an application may operate on the packet, for instance, modifying aspects of the packet or deleting the packet altogether. The system in turn facilitates injection of a packet from user space into kernel space, and into a designated point in the protocol stack for desired stack processing.

Method for assigning optimal packet lengths in a variable rate communication system

Abstract: A method for assigning optimal packet lengths in a variable rate communication system capable of data transmission at one of a plurality of data rates. The packet lengths for the data rates are selected such that the maximum throughput rate is achieved while conforming to a fairness criteria. The fairness criteria can be achieved by restricting the packet length assigned to each data rate to a range of value, or Limin ≤ L?i? ≤ Li?max?. The packet lengths for all data rates are first initialized to the maximum packet lengths for those data rates. Then, for each date rate, a determination is made whether another packet length assignment would result in improved throughput rate. If the answer is yes, the packet length for this data rate is reassigned and the throughput rate with the updates packet length assignments is recomputed. The process is repeated for each data rate until all data rates have been considered. The throughput rate can be calculated using a probabilistic model or a deterministic model.4

Software interface between switching module and operating system of a data packet switching and load balancing system

Abstract: A data packet switching system comprises a plurality of network interfaces each adapted to be coupled to respective external networks for receiving and sending data packets to and from the external networks via a particular communication protocol. The data packet switching system further includes a plurality of symmetrical processors, including a first processor providing a control processor and remaining ones of the processors each providing data packet switching processors. The data packet switching processors are coupled to the plurality of network interfaces. The control processor further includes a user portion and an operating system portion. The operating system portion is provided with a pseudo-network driver that appears to be a network interface to user application programs operating on the user portion of the control processor. A memory space is shared by the control processor and the data packet switching processors. The data packet switching processors route an incoming data packet directed to a user application program to the memory space. The pseudo-network driver retrieves the incoming data packet from the shared memory space and provides the data packet to the user application program.

Packet switching fabric using the segmented ring with resource reservation control

Abstract: A packet switching fabric includes a data ring, a control ring, a plurality of network links each coupled to at least one network node, and a plurality of switching devices coupled together by the data ring and the control ring so that the network links can be selectively communicatively coupled. Each of the switching devices includes: a data ring sub-system for transmitting and receiving bursts of data via data ring channels concurrently active on the data ring; a network interface coupled to the data ring sub-system and having at least one network port for transmitting and receiving data packets to and from one of the network links, the network interface also having a packet buffer for storing the data packets, the packet buffer providing bursts of packet data to the data ring sub-system via a plurality of concurrently active packet buffer channels; and a control ring sub-system coupled to the data ring sub-system and to the network interface and being responsive to control messages received from an adjacent one of the devices via the control ring, and operative to develop and transmit the control messages to an adjacent one of the devices via the control ring, the control messages for reserving bandwidth resources used in setting up and controlling the data ring channels and the packet buffer channels, the control ring sub-system also being operative to perform queuing operations for controlling the transfer of the bursts of packet data from the packet buffer to the data ring sub-system.

Forward error correction system for packet based data and real time media, using cross-wise parity calculation

Abstract: A computationally simple yet powerful forward error correction code scheme for transmission of real-time media signals, such as digitized voice, video or audio, in a packet switched network, such as the Internet. For each window of k data packets, the invention generates and transmits at least one cross-wise parity packet taken as an index-shifted function over the k data packets. The invention thereby enables a receiving end to recover from packet loss.

Method and apparatus for providing differentiated services using a multi-level queuing mechanism

Abstract: A method and apparatus for providing differentiated services using a multi-level queuing mechanism includes checking whether a packet of data is to receive a preferred level of service, and if the packet is not to receive the preferred level of service, then placing the packet in a first forwarding queue. However, if the packet of data is to receive the preferred level of service then checking whether it is permissible to forward the packet to a device in the network at the current time. If it is permissible to forward the packet to the device at the current time, then placing the packet in a second forwarding queue, otherwise temporarily placing the packet in a holding queue before placing the packet in the second forwarding queue.

Method and apparatus for tracking data packets in a packet data communication system

Abstract: A packet data communication system (100) employs a method and apparatus for tracking data packets in the packet data communication system. A sending communication device (e.g., 101) transmits a data packet (130) to a receiving communication device (e.g., 105), wherein the data packet includes a respective packet sequence number (131). Responsive to the transmission, the sending device increases a transmit tracking number (V(S)) that is used to indicate the packet sequence number of the data packet that is next in sequence to be transmitted. Upon receiving the data packet, the receiving device transmits an acknowledgment data packet (140) to the sending device, wherein the acknowledgment data packet includes a receive tracking number (139). The receive tracking number indicates the packet sequence number of the data packet that the receiving device expects to receive next. Upon receiving the acknowledgment data packet, the sending device determines whether a communication error occurred based on the receive tracking number and the transmit tracking number.

Method and apparatus for performing cut-through virtual circuit merging

Abstract: A method and apparatus for performing virtual connection merging in an output port of a network switch are disclosed. In the event no completely assembly packets have been received and scheduled for transmission (102), a partially received packet is selected for cut-through transmission prior to receipt of all cells comprising the packet (108). Transmission of the selected packet is initiated and a timer is started (110). If the timer expires prior to the receipt of an end of packet indication for the packet for which transmission has commenced (112), an end of packet signal is generated (114) and transmitted and the transmission of additional cells for the cut-through packet is aborted (116). In this manner, delays associated with packet reassembly may be avoided and sizes of reassembly buffers may be reduced.

Correlation of packet delay and loss in the internet

Abstract: In this paper we examine the correlation between packet delay and packet loss experienced by a continuous-media traffic source on the Internet. Our goal is to study the extent to which one performance measure can be used to predict of the future behavior of the other (e.g., whether observed increasing delay is a good predictor of future loss) so that an adaptive continuous media application might take {\it anticipatory\/} action based on observed performance. We ran numerous hour-long experiments in which continuous media traffic was sent from a source to a destination. We measured the per-packet delay and packet loss and then analyzed our measurements off-line. Our results provide a quantitative study of the extent to which such correlation exists. Interestingly, we observe periodic phenomena in the correlation that we had initially not expected. We discuss our results, speculate as to the reason for the observed behaviors, and discuss their implications for adaptive continuous media applications.

Performance analysis of CFDAMA-PB protocol for packet satellite communications

Abstract: Combined free/demand-assignment multiple-access (CFDAMA) schemes are suitable for broad-band packet satellite communications systems serving a finite number of bursty data sources. The performance analysis of the CFDAMA using piggy-backed (PB) reservation is presented. The probability generating function (PGF) of the packet delay is developed. The performance is evaluated in terms of three performance measures: average packet delay, variance of packet delay, and cumulative probability distribution of packet delay. Performance comparison with other pertinent schemes shows CFDAMA-PB to be superior for a wide range of user population sizes.

MPEG-2 Video Services over Packet Networks: Joint Effect of Encoding Rate and Data Loss on User-Oriented QoS

Abstract: We address the problem of video quality prediction and control for high resolution video transmitted over lossy packet networks. We analyze how the user-perceived quality is related to the average encoding bitrate for VBR MPEG2 video. We then show why simple distortion metrics (e.g., PSNR) may lead to inconsistent interpretations. Furthermore, for a given coder setup, we analyze the effect of packet loss on the user-level quality. We then demonstrate that, when jointly studying the impact of coding bit rate and packet loss, the reachable quality is upperbound and exhibits one optimal coding rate for a given packet loss ratio.

Apparatus and method for synthesizing management packets for transmission between a network switch and a host controller

Abstract: Management data is supplied to a management agent by a network switch by generating management packets having at least a portion of a received data packet, and management information specifying receive status and network switch response characteristics to the corresponding received data packet. The network switch includes a plurality of network ports, including network traffic ports and one management port for synthesizing the management frame. A network traffic port receiving a data packet generates receive status data specifying the reception status of the received data packet, including the presence of cyclic redundancy check (CRC) errors, frame alignment errors, and receive buffer overflow conditions. The received data packet and received status data are stored in a buffer memory, while switching logic generates port vectors specifying destination output ports and switching logic data specifying the switching logic response to the received data packet. The management port selectively compiles the received status data, the switching logic data, and at least a portion of the received data frame into a management frame, and outputs the synthesized management frame to a management agent according to a media access control (MAC) layer protocol. The generation of management frame provides detailed management information corresponding to characteristics of the received data packet and the corresponding network switch response, independent of the timing at which the data packet was received by the network switch.

Packet switching system, packet switching network and packet switching method

Abstract: Upon initialization, a VPC is set up between edge nodes. A control processor of each node creates an IP address/VPC mapping table using IP routing information and an address mapping table mapping correspondence between IP addresses and ATM addresses and supplied by a network management system. A gateway assigns a VCC to each packet input to the network. A sending-side edge node inputs the packet to the VPC corresponding to its destination by referring to the IP address/VPC mapping table. A transit node performs packet switching over VP. A receiving-side edge node transfers each packet to the gateway corresponding to its destination. If a series of packets meet a predetermined condition in a given edge node, its control processor sends VCC information to input interfaces of the edge node so that the packets are switched by an ATM switch in the edge node without intervention of the control processor.

Pseudo-interface between control and switching modules of a data packet switching and load balancing system

Abstract: A data packet switching system comprises a plurality of network interfaces each adapted to be coupled to respective external networks for receiving and sending data packets to and from the external networks via a particular communication protocol. The data packet switching system further includes a plurality of symmetrical processors, including a first processor providing a control processor and remaining ones of said processors each providing data packet switching processors. The control processor is coupled to a corresponding one of the plurality of network interfaces and the data packet switching processors are coupled to each remaining one of the plurality of network interfaces. A switch is coupled to the control processor through the corresponding one of the network interfaces and is coupled to at least one of the switching processors through at least one other one of the network interfaces. The switching processors thereby are capable of routing a received one of the data packets directed to the control processor through the switch to the control processor.

A deterministic approach to the end-to-end analysis of packet flows in connection-oriented networks

Abstract: We analyze the worst-case behavior of general connection-oriented networks, with first-in-first-out (FIFO) queueing policy, forwarding packets along an arbitrary system of routes. A worst-case bound is proven for the end-to-end queueing delay and buffer size needed to guarantee loss-free packet delivery, given that sources satisfy a given source rate condition. The results are based on a novel deterministic approach and help in reconciling the discrepancy between the unstable worst-case behavior of FIFO-based networks and their good practical performance.

Packet data flow control method and device

Abstract: In packet data flow control wherein point-to-multipoint communication or point-to-point communication is performed in a network comprising one or a plurality of packet exchanges that respectively accommodate a plurality of communication terminals, reduction of the transmission bandwidth or lowering of transmission rate caused by local congestion is absorbed by providing buffers for absorbing difference between the transmission bandwidths of the upstream link and downstream links of these packet exchanges at each packet exchange. At the packet exchanges the transmission rate of the downstream links from the packet exchange is controlled in accordance with a first load data that is notified thereto from the downstream links and, if the difference of transmission bandwidth of the upstream link and downstream links cannot be absorbed by the buffer, a second load data is generated based on the first load data and the free capacity of the buffer and the upstream link of the packet exchange is notified of the second load data.

System and method of transferring internet protocol packets using fast ATM cell transport

Abstract: In a method of transmitting an IP packet between a source and a destination through an ATM network which has a node formed by an ATM switch and a packet router, a reception packet or cell is transmitted to the node on an unused or undefined VC and is sent to the packet router in the node. In the packet router, an output port is selected by the use of the unused VC to establish a switched virtual channel in the ATM switch and to transfer each packet through the switched virtual channel after the switched virtual channel is established, as long as the reception packet is sent on the same VCI. Neither signaling nor protocol is needed between the nodes.

Support of multiple modulation levels for a cellular packet control channel

Abstract: A D-AMPS+ cellular communications air interface (50) is presented wherein a packet data control channel (40, 60) and packet data traffic channel (42, 62) are supported in addition to the conventional digital control channel (20) and digital traffic channel (22, 64). In particular, the packet data control channel and packet data traffic channel support multiple modulation level operation (high versus low). Procedures are provided for mobile station selection, as well as re-selection (102, 110), of either the high or low-level modulation for the packet channels. Procedures are further provided for facilitating a fall-forward (158, 164, 190, 194) to the high-level modulation packet data control channel, or a fall-backward (222, 232, 254) to the low-level modulation packet data control channel with respect to both uplink and downling packet data communications.

Physical and logical validation of a network based on all-optical packet switching systems

Abstract: The large growth of telecommunication traffic demand generated by multiple new applications and expected to last at least for the next decade will force telecom operators to consider offering more flexible transport services. All-optical packet switching is a powerful technique to provide this flexibility and to support in a cost-efficient way a wide range of bandwidth consuming applications. After a very brief introduction about the packet-switched network architecture studied in the framework of the ACTS KEOPS project, we describe the structure of the packet-switching node we have defined. We then move into physical and logical analysis of the network including more than 40 network sections based on 160 Gb/s throughput optical packet switching nodes could operate error free. In addition, logical simulations have proved that such networks could provide a quality of service (packet loss rate and packet transfer delay per node) compatible with a large variety of service classes. Both results validate the feasibility of the network concept and pace the way toward a flexible network based on all-optical switching techniques.

Optical packet switching without packet alignment

Abstract: Operation without packet alignment of an all-optical packet switch is proposed and predicted feasible through a detailed traffic analysis. Packet alignment units are eliminated resulting in a simple switch architecture while optimal traffic performance is maintained through the flexibility provided by WDM.