COMMON TIME REFERENCE FOR PACKET SWITCHES
BACKGROUND OF THE INVENTION:
This invention relates generally to a method and apparatus for transmitting of data on a communications network. More specifically, this invention relates to timely forwarding and delivery of data over the network and to their destination nodes. The timely forwarding is possible by time information that is globally available from a global positioning system (GPS). Consequently, the end-to-end performance parameters, such as, loss, delay and jitter, have either deterministic or probablistic guarantees.
This invention further facilitates a method and apparatus for integrating the transfer of two traffic types over a data packet communications network. More specifically, this invention provides timely forwarding and delivery of data packets from sources with constant bit rate (CBR) and variable bit rate (VBR) over the network and to their destination nodes.
This invention further relates a method and apparatus for transmitting of data on a communications network via communication links with variable delays. More specifically, this invention relates to timely forwarding and delivery of data over networks with links with variable delays to their destination nodes, while ensuring the end-to-end performance parameters, such as, loss, delay and jitter, have either deterministic or probablistic guarantees.
This invention also provides a method and apparatus for monitoring, policing, and billing of the transmission of data packet on a communications network. More specifically, this invention provides the monitoring, policing, and billing in networks with timely forwarding and delivery of data packets to their destination nodes. Consequently, the end-to-end performance parameters, such as, loss, delay and jitter, are predictable, and therefore, it is possible to measure them. Consequently, such measurements are used in the monitoring, policing and billing.
This invention also provides for a method and apparatus for transmitting of data on an heterogeneous communications network, which has two types of switching nodes: (i) asynchronous and (ii) synchronous with common time reference. More specifically, this invention provides timely forwarding and delivery of data over the network and to their destination nodes. Consequently, the end-to-end performance parameters, such as, loss, delay and jitter, have either deterministic or probablistic guarantees.
This invention furthermore facilitates a routing decision by using both timing information and position information of packets within time frames. In this case, there
is no need to decode the address in the packet header, it is feasible to encrypt the entire data packet (including the header) as it is transferred through a public backbone network, which is an important security feature. Consequendy, over this novel communications network it is possible to transport wide variety of data packets, such as, IP (Internet protocol) and ATM (asynchronous transfer mode).
The proliferation of high-speed communications links, fast processors, and affordable, multimedia-ready personal computers brings about the need for wide area networks that can carry real time data, like telephony and video. However, the end-to- end transport requirements of real-time multimedia applications present a major challenge that cannot be solved satisfactorily by current networking technologies. Such applications as video teleconferencing, and audio and video multicasting generate data at a wide range of bit rates and require predictable, stable performance and strict limits on loss rates, average delay, and delay variations ("jitter"). These characteristics and performance requirements are incompatible with the services that current circuit and packet switching networks can offer.
Packet switching networks like LP (Internet Protocol)- based Internet and Intranets [see, for example, ATannebaum, "Computer Networks" (3rd Ed) Prentice Hall, 1996] and ATM (Asynchronous Transfer Mode) [see, for example, Handel et al., "ATM Networks: Concepts, Protocols, and Applications", (2nd Ed.) Addison- Wesley, 1994] handle bursty data more efficiently than circuit switching, due to their statistical multiplexing of the packet streams. However, current packet switches and routers operate asynchronously and provide best effort service only, in which end-to-end delay and jitter are neither guaranteed nor bounded. Furthermore, statistical variations of traffic intensity often lead to congestion that results in excessive delays and loss of packets, thereby significantly reducing the fidelity of real-time streams at their points of reception. In fact, under best effort service, the performance characteristics of a given connection are not even predictable at the time of connection establishment.
Efforts to define advanced services for both IP and ATM have been conducted in two levels: (1) definition of service, and (2) specification of methods for providing different services to different packet streams. The former defines interfaces, data formats, and performance objectives. The latter specifies procedures for processing packets by hosts and switches/routers. The types of services that defined for ATM include constant bit rate (CBR), variable bit rate (VBR) and available bit rate (ABR). For EP, the defined services include guaranteed performance (bit rate, delay), controlled flow, and best effort [J. Wroclawski, "Specification of the Controlled-Load Network
Element Service", IETF RFC 2211, September 1997] [Shenker et al., "Specification of Guaranteed Quality of Service", IETF RFC 2212. September 1997]. Signaling
protocols, e.g., RSVP and UNI3.1, which carry control information to facilitate the establishment of the desired services, are specified for IP and ATM, respectively [R. Braden, "Resource ReSerVation Protocol (RSVP) - Version 1 Functional Specification, IETF Request for Comment RFC2205", September 1997] [Handel et al., "ATM Networks: Concepts, Protocols, and Applications", (2nd Ed.) Addison-Wesley, 1994].
These protocols address the transport of data to one destination known as unicast or multiple destinations multicast. In addition, SLP, a higher level protocol for facilitating the establishment of sessions that use the underlying services, is currently under definition under IETF auspices [Handley et al., "SIP-Session Initiation Protocol", <draft-draft-ietf-mmusic-sip-04.ps>, November 1997].
The methods for providing different services under packet switching fall under the general title of Quality of Service (QoS). Prior art in QoS can be divided into two parts: (1) traffic shaping with local timing without deadline scheduling, for example [M.G.H. Katevenis, "Fast Switching And Fair Control Of Congested Flow In Broadband Networks", IEEE Journal on Selected Areas in Communications, SAC-
5(8): 1315-1326, October 1987; Demers et al., "Analysis and Simulation Of A Fan- Queuing Algorithm", ACM Computer Communication Review (SIGCOMM'89), pages 3-12, 1989; S.J. Golestani, "Congestion-Free Communication In High-Speed Packet Networks", IEEE Transcripts on Communications, COM-39(12):1802-1812, December 1991; Parekh et al., "A GeneraUzed Processor Sharing Approach To Flow Control - The
Multiple Node Case", EEEFJACM T. on Networking, 2(2): 137- 150, 1994], and (2) traffic shaping with deadline scheduling, for example [Ferrari et al., "A Scheme For realtime Channel Establishment In Wide- Area Networks", IEEE Journal on Selected Areas in Communication, SAC-8(4):368-379, April 1990; Kandlur et al, "Real Time Communication In Multi-Hop Networks", EEEE Trans, on Parallel and Distributed
Systems, Vol. 5, No. 10, pp. 1044-1056, 1994]. Both of these approaches rely on manipulation of local queues by each router with little coordination with other routers. The Weighted Fair Queuing (WFQ), which typifies these approaches, is based on cyclical servicing of the output port queues where the service level of a specific class of packets is determined by the amount of time its queue is served each cycle [Demers et al., "Analysis and Simulation Of A Fair Queuing Algorithm", ACM Computer Communication Review (SIGCOMM'89), pages 3-12, 1989]. These approaches have inherent limitations when used to transport real-time streams. When traffic shaping without deadline scheduling is configured to operate at high utilization with no loss, the delay and jitter are inversely proportional to the connection bandwidth, which means that low rate connections may experience large delay and jitter inside the network. In traffic
shaping with deadline scheduling the delay and jitter are controlled at the expense of possible congestion and loss.
The recognition that the processing of packets by switches and routers constitutes a performance bottleneck resulted in the development of methods for enhancing performance by simplifying the processing of packets. Multiprotocol Label
Switching (MPLS) converts the destination address in the packet header into a short tag, which defines the routing of the packet inside the network [Callon et al., "A Proposed Architecture For MPLS" <draft-ietf-mpls-arch-OO.txt> INTERNET DRAFT, August 1997]. The real-time transport protocol (RTP) [H. Schultzrinne et. al, RTP: A Transport
Protocol for Real-Time Applications, IETF Request for Comment RFC 1889, January 1996] is a method for encapsulating time-sensitive data packets and attaching to the data time related information like time stamps and packet sequence number. RTP is currently the accepted method for transporting real time streams over IP internetworks and packet audio/video telephony based on ITU-T H.323.
Synchronous methods are found mostly in circuit switching, as compared to packet switching that uses mostly asynchronous methods. However, some packet switching synchronous methods have been proposed. IsoEthernet or IEEE 802.9a [IEEE 802.9a Editor. Integrated service (is): IEEE 802.9a "Isochronous Services With CSMA/CD MAC Service", IEEE Draft, March 1995] combines CSMA/CD (IEEE
802.3), which is an asynchronous packet switching, with N-ISDN and H.320, which is circuit switching, over existing Ethernet infrastructure (lOBase-T). This is a hybrid solution with two distinct switching methods: N-ISDN circuit switching and Ethernet packet switching. The two methods are separated in the time domain by time division multiplexing (TDM). The IsoEthernet TDM uses fixed allocation of bandwidth for the two methods - regardless of their utilization levels. This approach to resource partitioning results in undesirable side effects like under-utilization of the circuit switching part while the asynchronous packet switching is over loaded but cannot use the idle resources in the circuit switching part. One approach to an optical network that uses synchronization was introduced in the synchronous optical hypergraph [Y. Ofek, "The Topology, Algorithms And Analysis Of A Synchronous Optical Hypergraph Architecture", Ph.D. Dissertation, Electrical Engineering Department, University of Illinois at Urbana, Report No. UIUCDCS-R-87- 1343, May 1987], which also relates to how to integrate packet telephony using synchronization [Y. Ofek, "Integration Of Voice Communication On A Synchronous
Optical Hypergraph", INFOCOM'88, 1988]. In the synchronous optical hypergraph, the forwarding is performed over hyper-edges, which are passive optical stars. In [Li et
al., "Pseudo-Isochronous Cell Switching In ATM Networks", IEEE INFOCOM'94, pages 428-437, 1994; Li et al., "Time-Driven Priority: Flow Control For Real-Time Heterogeneous Internetworking", IEEE INFOCOM'96, 1996] the synchronous optical hypergraph idea was applied to networks with an arbitrary topology and with point-to- point links. The two papers [Li et al., "Pseudo-Isochronous Cell Switching In ATM
Networks", IEEE INFOCOM'94, pages 428-437, 1994; Li et al., "Time-Driven Priority: Flow Control For Real-Time Heterogeneous Internetworking", IEEE INFOCOM'96, 1996] provide an abstract (high level) description of what is called "RISC-like forwarding", in which a packet is forwarded, with little if any details, one hop every time frame in a manner similar to the execution of instructions in a Reduced Instruction Set
Computer (RISC) machine.
In U.S. Pat. No. 5,418,779 Yemini et al. discloses switched network architecture with common time reference. The time reference is used in order to determine the time in which multiplicity of nodes can transmit simultaneously over one predefined routing tree to one destination. At every time instance the multiplicity of nodes are transmitting to different single destination node.
Routing — the selection of an output port for an information segment (i.e. data packets) that arrives at an input port of a switch — is a fundamental function of communication networks. In circuit switching networks, the unit of switching is a byte, and the switching is made based on the location of the byte in a time frame.
Establishing a connection in a circuit switching network requires the network to reserve a slot for the connection in every frame. The position of the byte in the frame is different from link to link, so each switch maintains a translation table from incoming frame positions on each input port to respective output ports and frame positions therein. The sequence of frame positions on the links of the route constitute a circuit that is assigned for the exclusive use of a specific connection, which results in significant inflexibiUty: the connection is limited in traffic intensity by the capacity of the circuit and when the connection does not use the circuit no other is aUowed to use it. This feature is useful for CBR traffic, like PCM telephony, but it results in low utiUzation of the network when the traffic is bursty [C. Huitema, Routing in the Internet,
Prentice Hall, 1995, and A. Tannebaum Computer Networks (3rd Ed) Prentice HaU 1996].
SUMMARY OF THE INVENTION: In accordance with the present invention, a method is disclosed providing virtual pipes that carry real-time traffic over packet switching networks while guaranteeing end- to-end performance. The method combines the advantages of both circuit and packet
switching. It provides for aUocation for the exclusive use of predefined connections and for those connections it guarantees loss free transport with low delay and jitter. When predefined connections do not use their aUocated resources, other non-reserved data packets can use them without affecting the performance of the predefined connections. On the Internet the non-reserved data packet traffic is caUed "best effort" traffic.
This invention further describes a method for transmitting and forwarding packets over a packet switching network where the delay between two switches increases, decreases, or changes arbitrarily over time. Packets are being forwarded over each Unk inside the network in predefined periodic time intervals. The switches of the network maintain a common time reference, which is obtained either from an external source (such as GPS - Global Positioning System) or is generated and distributed internaUy. The time intervals are arranged with simple periodicity and complex periodicity (like seconds and minutes of a clock).
The invention provides methods for maintaining timely forwarding within predefined time interval over two types of Unk delay variations: (i) increasing delay and
(ti) decreasing delay. When the delay increases at some point of time a packet may be late for its predefined forwarding time interval. In such case the packet is delayed until the next time interval of its virtual pipe. When the Unk delay decreases, packets are buffered until the first time interval of its virtual pipe. Packets that are forwarded inside the network over the same route and in the same time intervals constitute a virtual pipe and share a pipe-ID. The pipe-ID can be either expUcit, such as a tag or a label that is generated inside the network, or impUcit, such as a group of IP addresses. A virtual pipe provides deterministic quality of service guarantees. The time interval in which a switch forwards a specific packet is determined by the packet's pipe-ID, the time it reaches the switch, and the current value of the common time reference.
In accordance with the present invention, the bandwidth aUocated to a connection and the delay and jitter inside the network are independent. MPLS can be used by the present invention to identify virtual pipes. The packet time-stamp that is carried in the RTP header can be used in accordance with the present invention to faciUtate time-based transport.
Under the aforementioned prior art methods for providing packet switching services, switches and routers operate asynchronously. The present invention provides real-time services by synchronous methods that utitize a time reference that is common to the switches and end stations comprising a wide area network. The common time reference can be reaUzed by using UTC (Coordinated Universal Time), which is globally available via, for example, GPS (Global Positioning System - see, for example:
http^/www.utexas.edu/depts/grg/gcraft/notes/gps/gps.html). By international agreement, UTC is the same aU over the world. UTC is the scientific name for what is commonly caUed GMT (Greenwich Mean Time), the time at the 0 (root) line of longitude at Greenwich, England. In 1967, an international agreement estabUshed the length of a second as the duration of 9,192,631,770 oscUlations of the cesium atom. The adoption of the atomic second led to the coordination of clocks around the world and the establishment of UTC in 1972. The Time and Frequency Division of the National Institute of Standards and Technologies (NIST) (see http^/www.boulder.nisLgov/timefreq) is responsible for coordinating with the International Bureau of Weights and Measures (BIPM) in Paris in maintaining UTC.
UTC timing is readily available to individual PCs through GPS cards. For example, TrueTime, Inc.'s (Santa Rosa, CA) PCI-SG provides precise time, with zero latency, to computers that have PCI extension slots. Another way by which UTC can be provided over a network is by using the Network Time Protocol (NTP) [D. MiUs, "Network Time Protocol" (version 3) IETF RFC 1305]. However, the clock accuracy of
NTP is not adequate for inter-switch coordination, on which this invention is based. Although the present invention reUes on time to control the flow of packets inside the network in a similar fashion as in circuit switching, there are major differences between the two approaches. In circuit switching, for each data unit (e.g., a byte) at the time it has been transmitted from its source, it is possible to predict deterministicaUy the future times it wiU be transmitted from any switch along its route [Ballart et al., "SONET: Now It's The Standard Optical Network", IEEE Communications Magazine, Vol. 29 No. 3, March 1989, pages 8-15]. The time resolution of this advanced knowledge is much shorter than the data unit transmission time. On the other hand, in accordance with the present invention, for each data unit (e.g., a ceU) at the time it has been transmitted from its source, it is possible to know the future time frames that this data unit wiU be forwarded along its route. However, the time frame, which constitutes the accuracy of this advance timing knowledge, is much larger than one data unit transmission time. For example, the transmission time of an ATM ceU (53 bytes) over a gigabit per second link is 424 nanoseconds, which is 294 times smaller than a typical time frame of 125 microseconds used in one embodiment of the present invention. There are several consequences that further distinguish the present invention from circuit switching.
In accordance with the present invention, the use of reserved resources is aUowed by aU packet traffic whenever the reserved resources are not in use.
In accordance with the present invention, the synchronization requirements are independent of the physical Unk transmission speed, whUe in circuit switching the synchronization becomes more and more difficult as the link speed increases.
In accordance with the present invention, timing information is not used for routing, and therefore, as in the Internet, for example, the routing is done using IP addresses or a tag/label.
In accordance with the present invention, the Internet "best effort" packet forwarding strategy can be integrated into the system.
In accordance with this invention, a method is disclosed for monitoring and poUcing the packet traffic in a packet switching network where the switches maintain a common time reference.
In accordance with this invention, a designated points inside the network is enabled to ascertain the level of packet traffic in predefine time intervals, and control the flow of packets and bring it back to predetermined levels in cases where the traffic volume exceeds predetermined levels. The information coUected by the designated points facihtates bilUng for Internet services based on network usage, and identification of faulty conditions and maUcious forwarding of packets that cause excessive delay beyond predetermined value.
In accordance with this invention, a method is described for exchanging timing messages and data packets between synchronous switches with a common time reference, and between end-stations/gateways and other synchronous switches, over an asynchronous network, i.e., a network with asynchronous switches. The method entaUs transmission of messages conveying the common time reference to end-station/gateways that have no direct access to the common time reference, and data packets that are sent responsive to the timing information and predetermined scheduled time intervals.
In accordance with this invention, methods are provided for mamtaining timely forwarding within predefined time interval over three network configurations: (i) end- station to synchronous switch over asynchronous (LAN) switches, (ii) end-station to gateway over asynchronous (LAN) switches and then to synchronous switch, (iii) synchronous switch to synchronous switch over asynchronous switches.
These and other aspects and attributes of the present invention wUl be discussed with reference to the foUowing drawings and accompanying specification.
BRIEF DESCRIPTION OF THE DRAWINGS: FIG. 1 is a schematic illustration of a virtual pipe and its timing relationship with a common time reference (CTR), wherein delay is determined by the number of time frames between the forward time out at Node A and the forward time out at Node D;
FIG. 2 is a schematic Ulustration of multiple virtual pipes sharing certain ones of the switches;
FIG. 3 is a schematic block diagram iUustration of a switch that uses a common time reference from the GPS (Global Positioning System) for the timely forwarding of packets disclosed in accordance with the present invention;
FIG. 4 iUustrates the relationship among the local common time reference (CTR) on the switches, and how the multiplicity of local times is projected on the real-time axis, wherein time is divided into time frames of a predefined duration;
FIG. 5 is a schematic illustration of how the common time reference is organized into contiguous time-cycles of k time-frames each and contiguous super-cycle of / time- cycles each;
FIG. 6 is a schematic iUustration of the relationship of the network common time reference and UTC (Coordinated Universal Time), such that, each time-cycle has 100 time-frames, of 125 microseconds each, and 80 time-cycles are grouped into one super- cycle of one second;
FIG. 7 is a schematic Ulustration of a data packet pipeline as in FIG. 1 , and correlating to data packet movement through the switches 10 versus time for forwarding over a virtual pipe with common time reference (CTR);
FIG. 8 iUustrates the mapping of the time frames into and out of a node on a virtual pipe, wherein the mapping repeats itself in every time cycle iUustrating time in versus forwarding time out;
FIG. 9 is an iUustration of a serial transmitter and a serial receiver;
FIG. 10 is a table of the 4B/5B encoding scheme for data such as is used by the AM7968 - TAXI chip set in accordance with one embodiment of the present invention; FIG. 11 is a table of the 4B/5B encoding scheme for control signals, such as, the time frame delimiter (TFD) such as is used by the AM7968, in accordance with one embodiment of the present invention;
FIG. 12 is a schematic block diagram of an input port with a routing controller;
FIG. 13 is a schematic diagram of the routing controller which determines to which output port an incoming data packet should be switched to and attaches the time of arrival (ToA) information to the data packet header;
FIG. 14 is a flow diagram of the routing controller operation;
FIGS. 15A and 15B iUustrate two generic data packet headers with virtual pipe ID (PID), and priority bit (P), wherein FIG. 15 A iUustrates a packet without time-stamp field, and wherein FIG. 15B Ulustrates a packet with time-stamp field, and also shows how the common time-reference value, time of arrival (ToA), is attached by the routing controUer;
FIG. 16 is a schematic block diagram of an output port with a scheduling controUer and a serial transmitter;
FIG. 17 is a schematic block diagram of the double-buffer scheduUng controUer;
FIG. 18 is a flow diagram of the double- buffer scheduling controller 46 operation;
FIG. 19 is a functional block diagram of the general scheduUng controUer with its transmit buffer and select buffer controUer;
FIG. 20 is a flow diagram describing the packet scheduUng controUer operation for computing the forwarding time of a packet based on the following input parameters: PID 35C, ToA 35T and the CTR 002;
FIG. 21 is a flow diagram illustrating the operation of the Select Buffer ControUer 45D;
FIG. 22 iUustrates the real-time protocol (RTP) packet header with time-stamp field of 32 bits; and FIG. 23 is a flow diagram describing the packet scheduUng controUer operation for computing the dispatching-time of a packet based on the foUowing input parameters: PID, ToA, CTR and the RTP time-stamp.
FIG. 24 is a schematic description of a switch with a common time reference partition into time-frames with predefined positions such that the input port can unambiguously identify the positions;
FIG. 25 is a description of the timing partition of the common time reference into cycle with k time frames in each, wliile each time frame is further partitioned into four predefined parts: a, b, c and d;
FIG. 26 is a schematic diagram of the time-based routing controller. This unit determines to which output port a data packet should be switched and attaches the time in and position information to the data packet header;
FIG. 27 is an example of a routing and scheduling table on one of the incoming input ports using the incoming time or time-frame of arrival (ToA) and the position counter value for determining: (i) the output port, (u) the out-going time-frame, and (Ui) the position of the out-going data packet within the out-going time-frame;
FIG. 28 is a schematic iUustration of a data packet which is sent across the fabric to the output port;
FIG. 29 is an example of a routing and scheduling table on one of the incoming input ports using the time stamp and position information for determining: (i) the output port, (u) the out-going time-frame, and (in) the position of the out-going data packet within the out-going time-frame;
FIG. 30 is a flow diagram of the routing controller operation;
FIG. 31 is a flow diagram of the data packet scheduling controUer 45 A operation;
FIG. 32 is a flow diagram of a different embodiment of the Select Buffer ControUer 45D; FIG. 33 is a schematic diagram of another alternate embodiment of the routing controller which determines to which output port an incoming data packet should be switched to and attaches the time of arrival (ToA) information to the data packet header;
FIGS. 34A and 34B are schematic Ulustrations of two generic data packet headers with virtual pipe ID (PID) and priority bits (P1/P2): (A) a packet without a time-stamp field and (B) a packet with a time-stamp field. This drawing also shows how the common time-reference value, time of arrival (ToA), is attached by the routing controller;
FIG. 35 is a table classifying the data packets;
FIG. 36 is a flow diagram of an alternate embodiment of the routing controller operation;
FIG. 37 is a schematic diagram of the scheduUng and congestion controller, where each buffer is divided into two parts, one for constant bit rate (CBR) and the other for variable bit rate (VBR);
FIG. 38 is a flow diagram describing an alternate embodiment of the packet scheduUng and rescheduUng controller operation for computing the forwarding time of a packet based on the foUowing input parameters: pipe-ID 35C, Time of arrival 35T and the common time reference 002;
FIG. 39 is a flow diagram describing an alternate embodiment of the select buffer and congestion controller 45D; FIG.40 is a schematic iUustration of a virtual pipe with a delay which varies in time between Node B and Node C;
FIG. 41 is a timing diagram of a virtual pipe with an increasing delay between Nodes B and C;
FIG. 42 is a timing diagram of a virtual pipe with a decreasing delay between Nodes B and C;
FIG. 43 is a schematic iUustration of a virtual pipe, p, with an alternate virtual pipe, /?';
FIG. 44 describes a node, Node E, in which two virtual pipes, p and p' are converging; FIG. 45 is a schematic Ulustration of a resynchronization mechanism of two virtual pipes Pipe-ID = p and Pipe- ID = p' on Node E;
FIG. 46 is a schematic iUustration of the delay analysis and scheduling controller with its transmit buffer and select buffer controUer;
FIG. 47 is a flow diagram describing an alternate embodiment of the Select Buffer ControUer; FIG. 48 is a flow diagram describing the delay analysis and scheduling controUer operation for computing the forwarding time of a data packet;
FIG. 49 specifies a program executed by the delay analysis and scheduler controUer for mobile nodes with increasing and decreasing delays in their incoming tinks; FIG. 50 specifies a program executed by the delay analysis and scheduler controUer for communication links in which their delay can change instantly, such as it is the case for SONET tinks in a self-heating SONET rings; and
FIG. 51 specifies a program executed by the delay analysis and scheduler controUer for combining two alternate paths p and p' into one path. FIG. 52 is a schematic iUustration of the delay monitoring controUer;
FIG. 53 is a flow chart of the program executed by the delay monitoring controUer;
FIG. 54 is a schematic iUustration of the policing and load controUer;
FIG. 55 is a flow chart of the program executed by the poUcing and load controUer;
FIG. 56 is a schematic iUustration of connection between an end-station and a synchronous virtual pipe switch that are separated by asynchronous LAN switches;
FIG. 57 is a schematic illustration of connections between end-stations and synchronous virtual pipe switches which are separated by some asynchronous switches and a gateway, which resynchronize the data packets before it is forwarded to the synchronous switch;
FIG. 58 is a schematic illustration of connection between two segments of virtual pipe switches which are separated by asynchronous switches and routers;
FIG. 59 is a diagram of the temporal relationship between the common time reference signals from a synchronous switch and the timely forwarding of packets from an end-station;
FIG. 60 is an Ulustration of the possible time of arrival (ToA) variations to a synchronous switch of data packets, which are forwarded over asynchronous switches;
FIG. 61 is a schematic Ulustration of a data packet resynchronization mechanism at the input port; and
FIG. 62 specifies a program executed by the delay analysis and scheduler controller for the case of finding the schedule exactly on the delay bound by using the time-stamp at the data packet header.
DETAILED DESCRIPTION OF THE ILLUSTRATED EMBODIMENTS:
Whtie this invention is susceptible of embodiment in many different forms, there is shown in the drawing, and wtil be described herein in detati, specific embodiments thereof with the understanding that the present disclosure is to be considered as an exempUfication of the principles of the invention and is not intended to limit the invention to the specific embodiments Ulustrated.
The present invention relates to a system and method for transmitting and forwarding packets over a packet switching network. The switches of the network maintain a common time reference, which is obtained either from an external source (such as GPS - Global Positioning System) or is generated and distributed internaUy. The time intervals are arranged in simple periodicity and complex periodicity (like seconds and minutes of a clock). A packet that arrives to an input port of a switch, is switched to an output port based on specific routing information in the packet's header (e.g., IPv4 destination address in the Internet, VCI/VPI labels in ATM). Each switch along a route from a source to a destination forwards packets in periodic time intervals that are predefined using the common time reference. The time interval duration can be longer than the time duration required for transmitting a packet, in which case the exact position of a packet in the time interval is not predetermined.
Packets that are forwarded inside the network over the same route and in the same periodic time intervals constitute a virtual pipe and share the same pipe-ID. Pipe- ID can be either expUcit, such as a tag or a label that is generated inside the network, or impUcit such as a group of IP addresses. A virtual pipe can be used to transport data packets from multiple sources and to multiple destinations. A virtual pipe provides deterministic quality of service guarantees. The time interval in which a switch forwards a specific packet is determined by the packet's pipe-ID, the time it reaches the switch, and the current value of the common time reference. In accordance with the present invention, congestion-free packet switching is provided for pipe-IDs in which capacity in their corresponding forwarding links and time intervals is reserved in advance. Furthermore, packets that are transferred over a virtual pipe reach their destination in predefined time intervals, which guarantees that the delay jitter is smaUer than or equal to one time interval.
Packets that are forwarded from one source to multiple destinations share the same pipe ID and the links and time intervals on which they are forwarded comprise a
virtual tree. This faciUtates congestion-free forwarding from one input port to multiple output ports, and consequently, from one source to multipUcity of destinations. Packets that are destined to multiple destinations reach aU of their destinations in predefined time intervals and with delay jitter that is no larger than one time interval. A system is provided for managing data transfer of data packets from a source to a destination. The transfer of the data packets is provided during a predefined time interval, comprised of a pluraUty of predefined time frames. The system is further comprised of a pluraUty of switches. A virtual pipe is comprised of at least two of the switches interconnected via communication links in a path. A common time reference signal is coupled to each of the switches, and a time assignment controller assigns selected predefined time frames for transfer into and out from each of the respective switches responsive to the common time reference signal. For each switch, there is a first predefined time frame within which a respective data packet is transferred into the respective switch, and a second predefined time frame within which the respective data packet is forwarded out of the respective switch. The time assignment provides consistent fixed intervals between the time between the input to and output from the virtual pipe.
In a preferred embodiment, there is a predefined subset of the predefined time frames during which the data packets are transferred in the switch, and for each of the respective switches, there are a predefined subset of the predefined time frames during which the data packets are transferred out of the switch.
Each of the switches is comprised of one or a pluraUty of addressable input and output ports. A routing controUer maps each of the data packets that arrives at each one of the input ports of the respective switch to a respective one or more of the output ports of the respective switch.
For each of the data packets, there is an associated time of arrival to a respective one of the input ports. The time of arrival is associated with a particular one of the predefined time frames. For each of the mappings by the routing controUer, there is an associated mapping by a scheduling controller, which maps of each of the data packets between the time of arrival and forwarding time out The forwarding time out is associated with a specified predefined time frame.
In the preferred embodiment, there are a pluraUty of the virtual pipes comprised of at least two of the switches interconnected via communication links in a path. The communication Unk is a connection between two adjacent switches; and each of the communications tinks can be used simultaneously by at least two of the virtual pipes.
Multiple data packets can be transferred utiUzing at least two of the virtual pipes.
In some configurations of this invention there is a fixed time difference, which is constant for all switches, between the time frames for the associated time of arrival and forwarding time out for each of the data packets. The fixed time difference is a variable time difference for some of the switches. A predefined interval is comprised of a fixed number of contiguous time frames comprising a time cycle. Data packets that are forwarded over a given virtual pipe are forwarded from an output port within a predefined subset of time frames in each time cycle. Furthermore, the number of data packets that can be forwarded in each of the predefined subset of time frames for a given virtual pipe is also predefined. The time frames associated with a particular one of the switches within the virtual pipe are associated with the same switch for aU the time cycles, and are also associated with one of input into or output from the particular respective switch.
In some configurations of this invention there is a constant fixed time between the input into and output from a respective one of the switches for each of the time frames within each of the time cycles. A fixed number of contiguous time cycles comprise a super cycle, which is periodic. Data packets that are forwarded over a given virtual pipe are forwarded from an output port within a predefined subset of time frames in each super cycle. Furthermore, the number of data packets that can be forwarded in each of the predefined subset of time frames within a super cycle for a given virtual pipe is also predefined.
In the preferred embodiment the common time reference signal is coupled from a GPS (Global Positioning System), and is in accordance with the UTC (Coordinated Universal Time) standard. The UTC time signal does not have to be received directly from GPS. Such signal can be received by using various means, as long as the delay or time uncertainty associated with that UTC time signal does not exceed half a time frame.
In one embodiment, the super cycle duration is equal to one second as measured using the UTC (Coordinated Universal Time) standard. The super cycle can also be equal to multiple UTC seconds or a fraction of a UTC second.
A select buffer controUer maps one of the time frames for output from a first switch to a second time frame for input via the communications Unk to a second switch.
The select buffer controUer uses the UTC time signal in order to identify the boundaries between two successive time frames. The select buffer controUer inserts a time frame delimiter (TFD) signal into the transmission Unk in order to the signal the second switch with the exact boundary between two time frames. Each of the data packets is encoded as a stream of data, and a time frame detimiter is inserted into the stream of data responsive to the select buffer controller. This can be implemented by using a redundant serial codewords as it is later explained.
The communication links can be of fiber optic, copper, and wireless communication tinks for example, between a ground station and a sateltite, and between two sateUites orbiting the earth. The communication link between two nodes does not have to be a serial communication link. A parallel communication Unk can be used - such Unk can simultaneously carry multiple data bits, associated clock signal, and associated control signals.
The data packets can be Internet protocol (IP) data packets, and asynchronous transfer mode (ATM) ceUs, and can be forwarded over the same virtual pipe having an associated pipe identification (PID). The PID can be an Internet protocol (IP) address, Internet protocol group multicast address, an asynchronous transfer mode (ATM), a virtual circuit identifier (VCI), and a virtual path identifier (VPI), or (used in combination as VCI/VPI).
The routing controUer determines two possible associations of an incoming data packet: (i) the output port, and (U) the time of arrival (ToA). The ToA is then used by the scheduUng controUer for determining when a data packet should be forwarded by the select buffer controller to the next switch in the virtual pipe. The routing controller utUizes at least one of Internet protocol version 4 (IPv4), Internet protocol version 6 (IPv6) addresses, Internet protocol group multicast address, Internet MPLS (multi protocol label swapping or tag switching) labels, ATM virtual circuit identifier and virtual path identifier (VCI VPI), and IEEE 802 MAC (media access control) addresses, for mapping from an input port to an output port.
Each of the data packets is comprised of a header, which includes an associated time stamp. For each of the mappings by the routing controUer, there is an associated mapping by the scheduUng controller, of each of the data packets between the respective associated time-stamp and an associated forwarding time out, which is associated with one of the predefined time frames. The time stamp can record the time in which a packet was created by its application.
In one embodiment the time-stamp is generated by an Internet real-time protocol (RTP), and by a predefined one of the switches. The time-stamp can be used by a scheduling controUer in order to determine the forwarding time of a data packet from an output port.
Each of the data packets originates from an end station, and the time-stamp is generated at the respective end station for inclusion in the respective originated data packet. Such generation of a time-stamp can be derived from UTC either by receiving it directly from GPS or by using the Internet's Network Time Protocol (NTP).
1 Synchronous virtual pipe
In accordance with the present invention, a system is provided for transferring data packets across a data network whUe maintaining for reserved data traffic constant bounded jitter (or delay uncertainty) and no congestion-induced loss of data packets. Such properties are essential for many multimedia applications, such as, telephony and video teleconferencing.
In accordance with the design, method, and iUustrated implementation of the present invention, one or a plurality of virtual pipes 25 are provided, as shown in FIGS. 1-2, over a data network with general topology. Such data network can span the globe. Each virtual pipe 25 is constructed over one or more switches 10, shown in FIG. 1, which are interconnected via communication links 41 in a path.
FIG. 1 iUustrates a virtual pipe 25 from the output port 40 of switch A, through switches B and C. This virtual pipe ends at the output port 40 of node D. The virtual pipe 25 transfers data packets from at least one source to at least one destination. FIG. 2 iUustrates three virtual pipes: virtual pipe 1 from the output of switch A to the output of switch D, virtual pipe 2 from the output of switch B to the output of switch D, and virtual pipe 3 from the output of switch A to the output of switch C.
The data packet transfers over the virtual pipe 25 via switches 10 are designed to occur during a pluraUty of predefined time intervals, wherein each of the predefined time intervals is comprised of a plurality of predefined time frames. The timely transfers of data packets are achieved by coupling a common time reference 002 (CTR) signal to each of the switches 10.
FIG. 3 illustrates the structure of a pipetine switch 10. The switch 10 is comprised of one or a plurality of input ports 30, one or a plurality of output ports 40, switching fabric 50, and global positioning system (GPS) time receiver 20 with a GPS antenna 001. The GPS time receiver provides a common time reference signal (CTR)
002 to aU input and output ports.
1.1 The common time reference (CTR) 002
As shown in FIG. 4, the common time reference 002 that is coupled to the switches 10 provides the following property: the local clock ticks 004, shown in FIG. 4, at aU the pipetine switches (e.g., switches A, B, C, and D in FIGS. 1 and 2) when projected on the real-time axis 005 wiU aU occur within predefined synchronization envelopes 003. In other words, the local clock ticks 004 occur within the synchronization envelopes 003, and therefore, outside relative to the synchronization envelopes aU local clocks have the same clock value.
The common time reference is divided in a predefined manner into time frames, Tf, of equal duration, as shown in FIG. 4, typically Tf= 125 microseconds. The time
frames are grouped into time cycles. Each time cycle has predefined number of time frames.
Referring to FIG. 5, there are k time frames in each time cycle. Contiguous time cycles are grouped together into contiguous super cycles, and as shown in FIG. 5, there are / time cycles in each super cycle.
FIG. 6 Ulustrates how the common time reference can be aligned with the UTC (Coordinated Universal Time) standard. In this Ulustrated example, the duration of every super cycle is exactly one second as measured by the UTC standard. Moreover, the beginning of each super cycle coincides with the beginning of a UTC second, as shown in FIG. 6. Consequently, when leap seconds are inserted or deleted for UTC corrections (due to changes in the earth rotation period) the cycle and super cycle periodic scheduling wtil not be affected.
The time frames, time cycles, and super cycles are associated in the same manner with aU respective switches within the virtual pipe at aU times. 1.2 Pipeline forwarding
Pipeline forwarding relates to data packets being forwarded across a virtual pipe 25 with a predefined delay in every stage (either across a communication Unk 41 or across a switch 10 from input port 30 to output port 40). Data packets enter a virtual pipe 25 from one or more sources and are forwarded to one or more destinations. This sort of pipeline forwarding used in accordance with the present invention is iUustrated in FIG. 7. Data packet 41A is forwarded out of switch A during time frame t-1. This data packet 41 A wiU reach switch B after a delay of T-ab. This data packet 41A wUl be forwarded out of switch B as data packet 41B during time frame t+7 and will reach switch C after a delay of T-bc. This data packet 41B will be forwarded out of switch C as data packet 41C during time frame t+5. Data packet 41C wiU reach switch
D after a delay of T-cd. Consequently, the delay from the output of switch A to the output of switch C is 6=t+5-(t-7) time frames. As iUustrated in FIG. 7, all data packets that are forwarded over that virtual pipe wtil have a delay of six time frames from the output of switch A to the output of switch C. Referring again to FIG. 1 , the timely pipeline forwarding of data packets over the virtual pipe 25 is Ulustrated. A data packet is received by one of the input ports 30 of switch A at time frame 1, and is forwarded along this virtual pipe 25 in the following manner: (i) the data packet 41A is forwarded from the output port 40 of switch A at time frame 2 of time cycle 1, (ii) the data packet 41B is forwarded from the output port 40 of switch B, after 18 time frames, at time frame 10 of time cycle 2, (iii) the data packet 41C is forwarded from the output port 40 of switch C, after 42 time frames, at time frame 2
of time cycle 7, and (iv) the data packet 41D is forwarded from the output port 40 of switch D, after 19 time frames, at time frame 1 of time cycle 9. As iUustrated in FIG. 1,
• All data packets enter the virtual pipe 25 (i.e., forwarded out of the output port 40 of switch A) periodically at the second time frame of a time cycle, are output from this virtual pipe 25 (i.e., are forwarded out of the output port 40 of switch D) after 79 time frames.
• The data packets that enter the virtual pipe 25 (i.e., forwarded out of the output port 40 of switch A) can come from one or more sources and can reach switch A over one or more input links 41.
• The data packets that exit the virtual pipe 25 (i.e., forwarded out of the output port 40 of switch D) can be forwarded over plurality of output links 41 to one of pluraUty of destinations.
• The data packets that exit the virtual pipe 25 (i.e., forwarded out of the output port 40 of switch D) can be forwarded simultaneously to multiple destinations, (i.e., multicast (one-to-many) data packet forwarding).
• The communication link 41 between two adjacent ones of the switches 10 can be used simultaneously by at least two of the virtual pipes.
In FIG. 2, where there are three virtual pipes: • The three virtual pipes can multiplex (i.e., mix their traffic) over the same communication links.
• The three virtual pipes can multiplex (i.e., mix their traffic) during the same time frames and in an arbitrary manner.
• The same time frame can be used by multiple data packets from one or more virtual pipes.
1.3 Virtual pipe capacity assignment
For each virtual pipe there are predefined time frames within which respective data packets are transferred into its respective switches, and separate predefined time frames within which the respective data packets are transferred out of its respective switches. Though the time frames of each virtual pipe on each of its switches can be assigned in an arbitrary manner along the common time reference, it is convenient and practical to assign time frames in a periodic manner in time cycles and super cycles. FIG. 8 iUustrates the timing of a switch of a virtual pipe wherein there are a predefined subset of time frames (i, 75, and 80) of every time cycle, during which data packets are transferred into that switch, and wherein for that virtual pipe there are a predefined subset time frames (i+3, 1, and 3) of every time cycle, during which the data packets are transferred out of that switch. If each of the three data packets has 125 bytes
or 1000 bits, and there are 80 time frames of 125 microseconds in each time cycle (i.e., time cycle duration of 10msec), then the bandwidth aUocated to this virtual pipe is 300,000 bits per second.
In general, the bandwidth or capacity allocated for a virtual pipe is computed by dividing the number of bits transferred during each of the time cycles by the time cycle duration. In the case of a super cycle, the bandwidth allocated to a virtual pipe is computed by dividing the number of bits transferred during each of the super cycles by the super cycle duration.
The switch 10 structure, as shown in FIG. 3, can also be referred to as a pipeline switch, since it enables a network comprised of such switches to operate as a large distributed pipetine architecture, as it is commonly found inside digital systems and computer architectures.
Each pipeline switch 10 is comprised of a pluraUty of addressable input ports 30 and output ports 40. As illustrated in FIG. 12, the input port 30 is further comprised of a routing controUer 35 for mapping each of the data packets that arrives at each one of the input ports to a respective one of the output ports. As Ulustrated in FIG. 16, the output port 40 is further comprised of a scheduUng controUer and transmit buffer 45. An output port 40 is connected to an input port 30 via a communication Unk 41, as shown in FIG. 9. The communication Unk can be realized using various technologies compatible with the present invention.
As shown in FIG. 3, the common time reference 002 is provided to the input ports 30 and output ports 40 from the GPS time receiver 20, which receives its timing signal from the GPS antenna 001. GPS time receivers are avaUable from variety of manufacturers, such as, TrueTime, Inc. (Santa Rosa, CA). With such equipment, it is possible to maintain a local clock with accuracy of ±1 microsecond from the UTC
(Coordinated Universal Time) standard everywhere around the globe. 1.4 The communication link and time frame delimiter encoding
The communication links 41 used for the system disclosed is in this invention can be of various types: fiber optic, wireless, etc. The wireless Unks can be between at least one of a ground station and a satellite, between two sateUites orbiting the earth, or between two ground stations, as examples.
Referring to FIG. 9, a serial transmitter 49 and serial receiver 31 are Ulustrated as coupled to each link 41. A variety of encoding schemes can be used for a serial line link 41 in the context of this invention, such as, SONET/SDH, 8B/10B Fiber Channel, 4B/5B FDDI (fiber distributed data interface). In addition to the encoding and decoding of the data transmitted over the serial Unk, the serial transmitter/receiver (49 in FIG. 12 and 31 in FIG. 16) sends/receives control words for a variety of control purposes,
mostly unrelated to the present invention description. However, one control word, time frame delimiter (TFD), is used in accordance with the present invention. The TFD marks the boundary between two successive time frames and is sent by a serial transmitter 49 when a CTR 002 clock tick occurs in a way that is described hereafter as part of the output port operation. It is necessary to distinguish in an unambiguous manner between the data words, which carry the information, and the control signal or words (e.g., the TFD is a control signal) over the serial Unk 41. There are many ways to do this. One way is to use the known 4B/5B encoding scheme (used in FDDI). In this scheme, every 8-bit character is divided into two 4-bit parts and then each part is encoded into a 5-bit codeword that is transmitted over the serial Unk 41.
FIG. 10 iUustrates an encoding table from 4-bit data to 5-bit serial codewords. The 4B/5B is a redundant encoding scheme, which means that there are more codewords than data words. Consequently, some of the unused or redundant serial codewords can be used to convey control information. FIG. 11 is a table with 15 possible encoded control codewords, which can be used for transferring the time frame delimiter (TFD) over the serial link. The TFD transfer is completely transparent to the data transfer, and therefore, it can be sent in the middle of the data packet transmission in a non-destructive manner.
When the communication Unks 41 are SONET/SDH, the time frame delimiter cannot be embedded as redundant serial codewords, since SONET/SDH serial encoding is based on scrambting with no redundancy. Consequently, the TFD is implemented using the SONET/SDH frame control fields: transport overhead (TOH) and path overhead (POH). Note that although SONET/SDH uses a 125 microseconds frame, it cannot be used directly in accordance with the present invention, at the moment, since SONET/SDH frames are not globally aUgned and are also not aligned to UTC.
However, if SONET/SDH frames are globaUy aligned, SONET/SDH can be used compatibly with the present invention. 1.5 The input port
As shown in FIG. 12, the input port 30 has three parts: serial receiver 31, a routing controller 35 and separate queues to the output ports 36. The serial receiver 31 transfers the data packets and the time frame detimiters to the routing controUer 35.
The routing controller 35 is constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data packet, read only memory (ROM) for storing the routing controUer processing program and the routing table that is used for determining the output port that the incoming data packet should be switched to.
As Ulustrated in FIG. 13, the incoming data packet header includes a virtual pipe identification, PID 35C, that is used to lookup in the routing table 35D the address 35E
of the queue 36 that the incoming data packet should be transferred into. Before the packet is transferred into its queue 36, the time of arrival (ToA) 35T is attached to the packet header as Ulustrated in FIGS 15A and 15B. The ToA 35T is used by the scheduUng controller 45 of the output port 40 in the computation of the forwarding time out of the output port and shown in FIG. 16.
The data packet can have various formats, such as, Internet protocol version 4 (IPv4), Internet protocol version 6 (IPv6), asynchronous transfer mode (ATM) cells, etc. The data packets PID can be determined by one of the following: an Internet protocol (IP) address, an asynchronous transfer mode (ATM) a virtual circuit identifier, a virtual path identifier (VCI/VPI), Internet protocol version 6 (IPv6) addresses, Internet MPLS
(multi protocol label swapping or tag switching) labels, and IEEE 802 MAC (media access control) address, etc..
FIG. 14 iUustrates the flow chart for the router controUer 35 processing program executed by the routing controUer 35B. The program is responsive to two basic events from the serial receiver 31 of FIG. 12: the received time frame deUmiter TFD at step 35-
01, and the receive data packet at step 35-02. After receiving a TFD, the routing controUer 35 computes the time of arrival (ToA) 35T value at step 35-03 that is attached to the incoming data packets. For this computation it uses a constant, Dcorist, which is the time difference between the common time reference (CTR) 002 tick and the reception of the TFD at time t2 (generated on an adjacent switch by the CTR 002 on that node). This time difference is caused by the fact that the delay from the serial transmitter 49 to the serial receiver 31 is not an integer number of time frames. When the data packet is received at step 35-02, the routing controUer 35B executes three operations as set forth in step 35-04: attach the ToA, lookup the address of the queue 36 using the PID, and storing the data packet in that queue 36.
1.6 The switching fabric
There are various ways to implement a switching fabric. However, the switching fabric is peripheral to the present invention, and so it wiU be described only briefly. The main property that the switching fabric should ensure is that packets for which the priority bit P (35P in FIGS. 15A and 15B) is set to high, then priority (i.e., reserved traffic) wUl be switched into the output port in a constant bounded delay - measured in time frames.
This is possible in accordance with the present invention, where the packets in the input ports are already separated into queues to their respective output ports. Then, by using the Clos theorem in the time domain (see J.Y. Hui, "Switching and Traffic
Theory for Integrated Broadband Networks", page 65), the delay can be bounded by two time frames, one time frame at the input port and one time frame to get across the
switching fabric. Other implementations can be used, such as based on shared bus with round robin service of the high priority data packets, or on a crossbar switch.
Another possible switch design is shared memory, which ensures a deterministic delay bound from an input port to an output port. Shared memory packet switches are commerciaUy avaUable from various vendors, for example, MMC Networks Inc. (Santa
Clara, CA).
FIGS. 15A and 15B illustrate data packets without and with a time stamp attached, respectively. 1.7 The output port The output port 40 is iUustrated in FIG. 16, comprised of a scheduUng controller with a transmit buffer 45, and serial transmitter 49 (as previously described herein). The scheduUng controller 45 performs a mapping of each of the data packets between the associated respective time of arrival (ToA) and an associated forwarding time out of the output port via the serial transmitter 49. The forwarding time is determined relative to the common time reference (CTR) 002.
Three output port configurations are Ulustrated herein: a double-buffer scheduUng controUer, as shown in FIGS. 17 and 18, a general scheduling controller, as shown in FIGS. 19, 20, and 21, and a general scheduUng controller with time-stamp, as shown in FIGS. 22 and 23. The double-buffer scheduUng controUer 46, as iUustrated in the block diagram of
FIG. 17 and flow chart of FIG. 18, is constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data packet, and read only memory (ROM) for storing the controUer processing program. Each time frame as specified by the common time reference 002 is considered to be one of an even tick or an odd tick. The determination of even tick vs. odd tick is made relative to the beginning of a time cycle. In the preferred embodiment, the first time frame of a time cycle is determined to be an odd tick, the second time frame of the time cycle is determined to be an even tick, the third time frame of the time cycle is determined to be an odd tick, and so forth, where the determination of even tick vs. odd tick alternates as shown for the duration of the time cycle. In an alternate embodiment the first time frame of a time cycle is determined to be an even tick, the second time frame of the time cycle is determined to be an odd tick, the third time frame of the time cycle is determined to be an even tick, and so forth, where the determination of even tick vs. odd tick alternates as shown for the duration of the time cycle. The actual sequence of even ticks vs. odd ticks of time frames within a time cycle may be arbitrarily started with no loss in generality.
The double-buffer scheduUng controUer 46 operates in the following manner. Data packets arrive from the switching fabric 50 via link 51. When the priority bit 35P
is asserted (i.e., reserved traffic), the packet is switched through the packet DMUX (demultiplexer) 51S (during odd ticks of the common time reference) to buffer Ba via Unk 51-1, and during even ticks of the common time reference to buffer Bb, via link 51- 2. Data packets in which the priority bit 35P is not asserted (i.e., non-reserved traffic) are switched through the packet DMUX (demultiplexer) 51S to the "best effort" buffer
Be via Unk 51-3. The transmit buffer selection operation is controlled by the select signal 46A, which connects the double-buffer scheduUng controller with the packet DMUX (demultiplexer) 51S.
Data packets are forwarded to the serial transmitter 49 through the packet MUX (multiplexer) 47S, and link 47C in FIG. 17, during odd ticks of the common time reference from buffer Bb via link 46-2, and during even ticks of the common time reference from buffer Ba via link 46-1. If during odd ticks of the common time reference buffer Bb is empty, data packets from the "best effort" buffer Be are forwarded to the serial transmitter. If during even ticks of the common time reference buffer Ba is empty, data packets from the "best effort" buffer Be are forwarded to the serial transmitter. The transmit buffer selection operation is controUed by the select signal 46B, which connects the double-buffer scheduling controller 46 with the packet MUX (multiplexer) 47S.
A more general scheduling controller 45 operation is described in FIGS. 19, 20, and 21, which includes a transmit buffer 45C and a select buffer controller 45D. The data packet scheduling controUer 45A, together with the select buffer controUer 45D, perform the mapping, using the PID 35C and the data packet time of arrival (ToA) 35T in order to determine the respective time frame a respective packet should be forwarded out of the output port. Both controllers 45A and 45D are constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data, and read only memory (ROM) for storing the controUer processing program.
Data packets arrive from the switching fabric 50 via Unk 51. Data packets which have the priority bit 35P asserted (i.e., reserved traffic) are switched by the scheduling controller 45A to one of the k transmit buffers 45C (B-1, B-2, ...., E-k). Each of the it buffers is designated to store packets that wUl be forwarded in each of the k time frames in every time cycle, as shown in FIG. 5.
The flow chart for the program executed by the scheduling controUer is Ulustrated in FIG. 20. When the data packet is received from the fabric at step 45-03, the PID 35C in the data packet header is used to look-up the forward parameter 45F in the forwarding table (45B of FIG. 19), as specified in step 45-04. Next, the index of the transmit buffer, between B-1 and B-k is computed in step 45-05 by subtracting the time of arrival ToA 35T from the common time reference CTR 002 and by adding the
forward parameter 45F, and then switching the incoming data packet to transmit buffer B- , as specified in step 45-06.
Incoming data packets in which the priority bit 35P is not asserted (i.e., non- reserved traffic) are switched by the scheduUng controUer to the transmit "best effort" buffer B-E via Unk 45-be.
FIG. 21 Ulustrates the flow chart for the select buffer controller 45D operation. The controUer 45D is responsive to the common time reference (CTR) tick 002, and at step 45-11, increments the transmit buffer index i (i.e., i:=i+l modk', where k' is the number of buffers for scheduled traffic) and sends a time frame delimiter TFD to the serial transmitter at step 45-12. Then, if the transmit buffer B-i is not empty, at step 45-
13, it wUl send a data packet from transmit buffer B-z, as specified in at step 45-14, else it wiU send a "best effort" data packet from the "best effort" buffer B-be, as specified at step 45-15.
FIGS. 22 and 23 tilustrate a system with a scheduling controUer, wherein each of the data packets is comprised of a header, including an associated time stamp. The time- stamp is generated by an Internet real-time protocol (RTP) in which its data packet format is Ulustrated in FIG. 22. Alternatively, the time-stamp can be generated by a predefined one of the switches 10 in the system, or the time stamp can be generated at a respective end station for inclusion in the respective originated data packet FIG. 23 iUustrates the operation of the scheduling controUer for the case where the packet header contains a time-stamp 35TS. Data packets arrive from the switching fabric 50 via Unk 51. Data packets in which the priority bit 35P is set (i.e., reserved traffic) are switched by the scheduling controUer to one of the k transmit buffers 45C (B-1, B-2,..., B-k). Each of the k buffers is designated to store packets that wiU be forwarded in each of the k time frames in every time cycle, as shown in FIG. 5. The flow chart for the program executed by the scheduUng controUer is Ulustrated in FIG. 23. When a data packet is received from the fabric at step 45-21, the PID 35C in the data packet header is used to look-up the forward parameter 45F in the forwarding table 45B, as specified in step 45-22. Next the index i of the transmit buffer, between B-1 and B-k, is computed in step 45-23 by subtracting the time of arrival ToA 35T from the common time reference CTR 002 and by adding the forward parameter 45F, and then switching the incoming data packet to transmit buffer B-i, as specified in step 45-24. 2 Time-based routing
In this variant of this invention, a packet that arrives to an input port of a switch is switched to an output port based on (i) its position within the predefined time interval and (u) the unique address of the incoming input port. Each switch along a route from a source to a destination forwards packets in periodic time intervals that are predefined
using the common time reference. The time interval duration can be longer than the time duration required for transmitting a packet.
FIG. 24 depicts a schematic description of a switch 10. The switch 10 is constructed of four components: a plurality of uniquely addressable input ports 30 (in FIG. 24 there are N such ports), a pluraUty of uniquely addressable output ports 40 (in
FIG. 24 there are N such ports), a switching fabric 50, and a global positioning system
(GPS) time receiver 20 with a GPS antenna 001. The GPS time receiver provides a common time reference (CTR) 002 to aU input and output ports. The common time reference is partitioned into time frames. Each of the time frames is further comprised of predefined positions such that the input port can unambiguously identify the positions. The time and position that a data packet arrives into the input port are used by the routing controller 35 in FIG. 12 for determining the output port that incoming data packet should be switched to.
In FIG. 24, each of the time frames, t=i and t=i+l, has four predefined positions: a, b, c and d. In each of the positions, one data packet can be stored. The positions can be marked explicitly with position delimiters (PDs) between the variable size data packets, as it wtil be explained below, or implicitly. Implicit position within a time frame can be achieved by either measuring time delays - this is suitable for sending a fixed size ATM (asynchronous transfer mode) ceUs, or by placing data packets of variable size in the predefined positions within each of the time frames - if the output port 40 does not have a data packet to transmit in a predefine position an empty or nuU data packet should be sent.
FIG. 25 depicts a common time reference (CTR) 002 axis that is divided into time cycles. Each time cycle is divided into predefined frames. Each of the time frame has predefined positions: a, b, c, and d of either fixed size (in time duration) or variable size (in time duration), consequently, the predefined position can have ether fixed size data packets or variable size data packets, respectively.
2.1 The time-based input port
The input port 30, shown in FIG. 12, has three parts: serial receiver 31, time- based routing controUer 35 and separate queues 36 to the plurality of output ports 40.
The serial receiver 31 transfers to the time-based routing controUer 35 data packets, time frame delimiters (TFD) and position delimiters (PD).
The routing controUer is constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data packets, read only memory (ROM) for storing the time-based routing controller processing program, and a time-based routing table is used for determining the foUowing parameters (see 35D in FIGS. 26, 27, and 29):
1. Parameter 35- 1 in table 35D (FIGS. 27 and 29) - the output port 40 that the incoming data packet should be switched to - this parameter is used for switching the data packet to the queue 36 that is leading to the corresponding output port;
2. Parameter 35-2 in table 35D (FIGS. 27 and 29) - the out-going time frame in which the data packet wtil be forwarded out of the output port - this parameter is attached to the data packet header in FIG. 27, and
3. Parameter 35-3 in table 35D (FIGS. 27 and 29) - the position within the out-going time frame in which the data packet wtil be forwarded out of the output port - this parameter is attached to the data packet header in FIG. 28. The time-based routing controUer 35B determines the entry to the time-based routing table 35D, in FIGS. 27 and 29, in various ways, such as:
1. Local time and position by using (1) the time frame of arrival (ToA) 35T - the time frame using the common time reference 002, and (2) the position value 35P within that time frame as measured by the position counter 35PC. This is depicted in FIG. 12.
2. Time stamp 35TS and position 35PC by using (1) the time stamp 35TS in the data packet header in FIG. 28B, and (2) the position value 35P within that time frame as measured by the position counter 35PC.
3. Time stamp, PID (shown in the packet headers in FIG. 28) and position 35PC by (1) the time stamp 35TS in the data packet header in FIG. 28B, (2) the virtual pipe ID (PID) 35C in the data packet header in FIG. 28B (the virtual pipe is discussed in details at the end of this description), and (3) the position value 35P within that time frame as measured by the position counter 35PC. This is depicted in FIG. 29.
The data packets, see for example FIG. 28, can have various formats, such as, Internet protocol version 4 (IPv4), Internet protocol version 6 (IPv6), asynchronous transfer mode (ATM) ceUs. The data packets PID 35C can be determined by one of the foUowing: an Internet protocol (IP) address, an asynchronous transfer mode (ATM), a virtual circuit identifier, and a virtual path identifier (VCI VPI), Internet protocol version 6 (IPv6) addresses, Internet MPLS (multi protocol label swapping or tag switching) labels, and IEEE 802 MAC (media access control) address.
The time stamp 35TS in the packet header in FIG. 28B can be generated by an application using Internet real-time protocol (RTP) and is used also in the TTU-T H.323 standard. Such data packets use the format depicted in FIG. 22. Alternatively the time- stamp can be generated by a predefined one of the switches in the system, or alternatively the time stamp is generated at the respective end node for inclusion in the respective originated data packet.
FIG. 30 is a detatied description of the program executed by the time-based routing controUer 35B. The program is responsive to three events from the serial receiver 31 and the position value 35P within that time frame as measured by the position counter 35PC. The time-based routing controller program FIG. 30 using the three parameters in table 35D in FIGS. 27 and 29 that is associated with this incoming packet operates as foUows:
1. Receive time frame delimiter TFD 35-1 - responsive to this event the routing controller resets the position counter (35P:=0 in 135-04 of FIG. 30) and computes the time-frame of arrival (ToA) 35T value as specified in 135-04 of FIG. 30. For this computation it uses a constant, Dconst, which is the time difference between the common time reference (CTR) 002 tick and the reception of the TFD at time t2 (note that the TFD was generated on an adjacent switch by the CTR 002 on that node). This time difference is caused by the fact that the delay from the serial transmitter 49 to the serial receiver 31 is not an integer number of time frames. 2. Receive position delimiter PD 135-02 - responsive to this event it increments the position counter, 35P:=35P+1, 135-05 of FIG. 30.
3. Receive data packet 135-03 -responsive to this event three operations are performed as shown in 135-06 of FIG. 30: (1) the out-going time frame parameter 35-2 is attached to the packet header, (2) the position within the out-going time frame parameter 35-3 is attached to the packet header, and (3) the data packet is stored in the queue 36 using the output port parameter 35-1 in table 35D in FIGS. 27 and 29. 2.2 The time-based output port
The output is depicted in FIG. 16, it has two parts a scheduUng controller with a transmit buffer 45, and serial transmitter 49, which was described before. The data packet scheduUng controller 45A, in FIG. 19, transfers the data packet the transmit buffer which is a random access memory (RAM) 45C, as described below.
The data packet scheduling controUer 45 operation is described in FIGS. 19, 31, and 32, which includes a transmit buffer 45C and a select buffer controller 45D. The scheduling controUer 45A together with the select buffer controller 45D perform the mapping, using the two parameters, 35-2 and 35-3, that were attached to the data packet by the routing controller 35B. Both controUers are constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data, and read only memory (ROM) for storing the controUer processing program.
Data packets that arrive from the switching fabric 50 via link 51 in which their priority bit 35P is asserted (i.e., reserved traffic) wtil be switched by the data packet scheduling controUer 45A to one of the k' transmit buffers 45C: B-1, B-2, ..., B-k' (one special case is when k'=k, where k is the time cycle size measured in time frames). Each
of the k' buffers is designated to store packet that wtil be forwarded in a cycUcaUy recurring order in each of the it time frames in every time cycle, as shown in FIGS. 5 and 6. The actual program executed by the data packet scheduling controUer is described in FIG. 31. When data packet is received from the fabric 145-01 (in FIG. 31) the two parameters, 35-2 and 35-3, in the data packet header are used to determine in which of the transmit buffer, between B-1 and B-k', to store that data packet and in what position, as specified in 145-02 in FIG. 31.
Incoming data packets in which their priority bit 35P, see FIG. 28, is not asserted (i.e., non-reserved traffic) wtil be switched by the data packet scheduUng controUer to the transmit "best effort" buffer B-E via link 45-be.
FIG. 32 depicts the select buffer controller 45D operation, which is responsive to the common time reference (CTR) tick 002, as specified in 145-11 (FIG. 32). Consequently, the select buffer controUer increments the transmit buffer index / 145-12 (i.e., i:=i+l modk', where k' is the number of buffers for scheduled traffic), sends a time frame delimiter TFD 47A to the serial transmitter 145-12, and reset the position pointer to one, p:=l 145-12. Then while the transmit buffer B- i is not empty 145-13, it wtil send a data packets from transmit buffer B- i, as specified in 145-14, 145-15 and 145-16, else if the transmit buffer B- is empty, it wtil send "best effort" data packets from the "best effort" buffer B-be, as specified in 145-17, until the end of the time frame (the next CTR 002 tick) or until buffer B-E becomes empty.
When the transmit buffer B- i is not empty 145-13, the select buffer controUer sends data packets from aU of the non-empty predefined positions in that buffer, as specified in 145-14. After sending a data packet or if position p in buffer B- is empty 145-15, the select buffer controller sends a position delimiter (PD) 47B to the serial transmitter and increments the position pointer p:=p+l, as specified in 145-16.
3 Traffic integration 3.1 The integrated innut port
The input port 30, shown in FIG. 12, has three parts: serial receiver 31, routing controller 35 and separate queues to the output ports 36. The serial receiver 31 transfers to the routing controller 35 the data packets and the time frame delimiters.
The routing controUer is constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data packet, read only memory (ROM) for storing the routing controller processing program, and routing table is used for determining the output port that the incoming data packet should be switched to. The incoming data packet header includes a virtual pipe identification - PID 35C in FIG. 34, that is used to lookup in the routing table 35D the address 35E of the queue the incoming data packet should be transferred into its queue 36. Before the packet is
transferred into its queue 36 the time of arrival (ToA) 35T in FIG. 34 is attached to the packet header. The ToA 35T wiU be used by the scheduUng controUer 45 in FIG. 16 in the computation of the forwarding time out of the output port.
The data packet can have various formats, such as, internet protocol version 4 (IPv4), Internet protocol version 6 (IPv6), asynchronous transfer mode (ATM) cells.
The data packets PID can be determined by one of the following: an Internet protocol (IP) address, an asynchronous transfer mode (ATM) a virtual circuit identifier, and a virtual path identifier (VCI/VPI), Internet protocol version 6 (IPv6) addresses, Internet MPLS (multi protocol label swapping or tag switching) labels, and IEEE 802 MAC (media access control) address.
FIG. 35 is a table for defining two bits, PI and P2, in the packet headers in FIG. 34. The two bits classify three types of data packets: P1/P2 are "00" constant bit rate (CBR) data traffic; PI, P2 are 01 variable bit rate (VBR) data traffic; and PI, P2 are "10" "best effort" data traffic. The above classification is used by the program executed by the routing controUer 35B, as shown in FIG. 36, in order to determine into which of the three parts of the queue to the output port 36, shown in FIG. 33, the data packet should be switched into.
FIG. 36 is a detailed description of the program executed by the routing controller 35B. The program is responsive to two basic events from the serial receiver 31: receive time frame delimiter TFD 235-01, and receive data packet 235-02. After receiving a TFD the routing controUer computes the time of arrival (ToA) 35T value 235-03 in FIG. 34, that is attached to the incoming data packets. For this computation it uses a constant, Dconst which is the time difference between the common time reference (CTR) 002 tick and the reception of the TFD at time t2 (note that the TFD was generated on an adjacent switch by the CTR 002 on that node). This time difference is caused by the fact that the delay from the serial transmitter 49 to the serial receiver 31 is not an integer number of time frames. When data packet is received 235-02 the routing controller 35B executes three operations 235-04 in FIG. 36: attach the ToA, lookup the address of the queue 36 using the PID, and storing the data packet in the queue 36 to the output port 37, while using P1/P2 in the header, in FIG. 34, in order to determine in what part, CBR/VBR/Best effort, of that queue to store the incoming data packet. 3.2 Integrated scheduling controUer in the output port
A more general scheduling controUer 45 operation is described in FIGS. 37-39, which includes a scheduling and rescheduling controUer 145A, a transmit buffer 145C, and a select buffer and congestion controller 145D, as shown in FIG. 37. The scheduling and rescheduUng controller 145A together with the select buffer controller 145D perform the mapping of the data packet into the time frame. The mapping is done
on the scheduUng and rescheduUng controUer using the PID 35C and the data packet time of arrival (ToA) 35T in order to determine the respective time frame in which the respective packet should be forwarded out of the output port. The details are presented below. Both controUers, 145A and 145D, are constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data, and read only memory
(ROM) for storing the controUer processing program.
In the more general configuration, depicted in FIG. 37, data packets that arrive from the switching fabric 50 via link 51 in which their priority bits 35P (P1/P2) are either "00" or "01" (i.e., reserved CBR traffic, or VBR traffic, respectively) wiU be switched by controUer 145A to one of the k transmit buffers in 145C: B-1, B-2, ..., B-k.
Each of the k buffers is designated to store packet that wtil be forwarded in each of the k time frames in every time cycle, that where defined in FIGS. 5 and 6. Another possible operation is to map the incoming packets separately to each of the time frames of a super-cycle. When a super-cycle mapping is implemented there are k*l transmit buffers in 145C: B-1, B-2, ..., B-it*/, i.e., k buffers to each of the / cycles of a super-cycle.
The actual program executed by the scheduling and rescheduling controller is described in FIG. 38. When a data packet is received from either the fabric via link 51 or from the select buffer and congestion controUer 145D via link 45R, as specified in 145-03, the 35C, 35T and 35P in the data packet header are used to look-up the forward parameter 45F in the forwarding table 145B, as specified in 145-04. Next the index i of the transmit buffer, between B-1 and B-k, is computed in 145-05 by subtracting the time of arrival ToA 35T from the common time reference CTR 002 and by adding the forward parameter 45F, and then switching the incoming data packet to transmit buffer B-i, as specified in 145-05. Incoming data packets in which their priority bits 35P, P1/P2, are either "10"
(i.e., non-reserved traffic or "best effort") or "11" (i.e., rescheduled packet) are switched by the scheduling and rescheduUng controUer to the transmit "best effort" buffer B-E via Unk 45-be.
FIG. 39 depicts the operation of the select buffer and congestion controUer 145D operation, which is responsive to the common time reference (CTR) tick 002.
When CTR signal is received 245-11 the foUowing operations are executed by controller 45D in 245-15:
1. Send time frame delimiter (TFD) control signal 47A to the serial transmitter 49; 2. Forward back to controller 145A aU unsent packets in transmit buffer B-f in which the 35P field in their header is not "11", i.e., it is not a rescheduled packet and
set the 35P field to "11", i.e., a rescheduled packet (note that a packet can be rescheduled only once);
3. If the number of data packets need to be rescheduled exceeds some predefined number, say n, then select at random n data packets and sends them back to the packet scheduling and rescheduling controUer 145A via link 45R and discard the remainder of the packets.
4. Increment the transmit buffer index f (i.e.,f:=f+l mod k', where k' is the number of buffers for scheduled traffic). If the CBR part of buffer B-f is not empty 245-12, then it will send a data packet from transmit buffer B-f first CBR packets and then VBR packets, as specified in 245-16 and 245-13, else it wiU send a "best effort" data packet from the "best effort" buffer B-E as specified in 245-14. 4 Operation with links with variable delay
The present invention further relates to a system and method for transmitting and forwarding packets over a packet switching network in which some of its communication tinks have dynamically varying delays. Such variations in the link delay can be the consequence of having mobile switching node (e.g., satellites).
FIG. 40 Ulustrates a virtual pipe 25 from the output port 40 of switch A, through switches B and C. This virtual pipe ends at the output port 40 of node D. The virtual pipe 25 transfers data packets from at least one source to at least one destination. In FIG. 40 the communication link that connects switch B to switch C may have a delay that varies in time with a defined delay bound. Such communication links are found in various network architectures, such as, mobile wireless networks, satellite networks, and self-healing SONET rings. In satellite networks the delay between a sateUite in space and a base-station on earth changes in the foUowing manner. First the delay decreases, as the sateUite appears above the horizon and is moving towards the base-station, and then the delay increases as the satellite is moving away from the base-station until it disappears below the horizon.
FIGS. 41 and 42 describe the delay changes on the link between Nodes B and C as it is projected on a common time reference (CTR), which is discussed in details in FIGS. 4, 5, and 6 above. FIG. 41 describes a communication link between Nodes B and
C, where the time of arrival to Node C decreases, i.e., the delay between Nodes B and C gets shorter. In FIG. 41, the delay of data packet 0a is longer than data packet lb, the delay of data packet lb is longer than data packet 2c, and so on. FIG. 42 describes a communication link between Nodes B and C, where the time of arrival to Node C increases, i.e., the delay between Nodes B and C gets longer. In FIG. 42, the delay of data packet 0a is shorter than data packet lb, the delay of data packet lb is shorter than data packet 2c, and so on. A complete description of the above ^synchronization
operation is part of the output operation described below in FIGS. 46, 47, 48, 50, and 51.
FIGS. 43, 44, and 45 describe a delay variations that are due to the forwarding of successive data packets on alternate paths or routes in the network. FIG. 43 shows two virtual pipes (defined below), p (from Node A to B to C to D and to E) and p' (from
Node A to B' to D' and to E), with the requirement that data packets wtil be forwarded out of Node E at the same predefined time regardless which virtual pipe,/? or/?', they were forwarded on. FIG. 44 shows the scenario in which the data packets on virtual pipe/?' arrive to Node E before the time they would have arrived to Node E on virtual pipe p. Consequently, the data packets on path /?' should be delayed, as shown in FIG.
44 and forwarded, in time, as if they have arrived on virtual pipe/?. FIG. 45 shows how such resynchronization can be achieved by using a resynchronization buffer on node E. In FIG. 45, a data packet from virtual pipe/?', 1 /?', enters a resynchronization buffer 10R, such that, when this data packet exits this buffer, 2/?', it wtil be forwarded from the output of Node E as if this packet was forwarded on virtual pipe /?. A complete description of the above resynchronization operation is part of the output port operation is described below in FIGS. 16, 46, 47, 48, 49, 50, and 51.
The output port 40 is iUustrated in FIG. 16, comprised of a scheduling controller with a transmit buffer 45, and serial transmitter 49 (as previously described herein). The scheduUng controller 45 performs a mapping of each of the data packets between the associated respective time of arrival (ToA) and an associated forwarding time out of the output port via the serial transmitter 49. The forwarding time is determined relative to the common time reference (CTR) 002.
The scheduling controller and transmit buffer 45 has various modes of operation which are described in FIGS. 46, 47, 48, 49, 50, and 51. The different operation modes correspond to some of the possible variations in the communications Unk delay as was discussed in FIGS. 40-45. The scheduUng controller and transmit buffer 45 in FIG. 46 includes three parts:
1. a delay analysis and scheduUng controller 245A which further comprises a forwarding table 245B,
2. a transmit buffer 245C which is typicaUy realized as a random access memory (RAM), and
3. a select buffer controUer 245D which forward data packets to the serial transmitter. The delay analysis and scheduling controUer 245A, together with the select buffer controller 245D, perform the mapping, using the PID 35C, the time-stamp 35TS and the data packet time of arrival (ToA) 35T in order to determine the respective time
frame a respective packet should be forwarded out of the output port Both controUers 245A and 245D are constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data, and read only memory (ROM) for storing the controller processing program. Data packets arrive from the switching fabric 50 via link 51. Data packets which have the priority bit 35P asserted (i.e., reserved traffic) are switched by the delay analysis and scheduling controller 245A to one of the l*k transmit buffers 45C (B-1, B- 2..., B-l*k). Each of the l*k buffers is designated to store packets that wtil be forwarded in each of the l*k time frames in every super cycle, as shown in FIG. 5 and FIG. 6. Having /*it transmit buffers enables the delay analysis and scheduling controller 245A to schedule data packets in a wide range of delay variations. When the super cycle is one second the scheduUng capability of the delay analysis and scheduling controller 245A is up to one second. However, this is an extreme case and in most practical scenarios the scheduUng requirements, even with delay varying links, is only a small number of time frames.
The transmit buffer 245C includes an additional buffer B-E for "best effort" data packets. The priority bit 35P in the "best effort" data packets is not asserted and this how the delay analysis and scheduling controller determines that such data packets should be stored in the "best effort" buffer. The "best effort" data packets are forwarded to the serial transmitter 49 whenever there are no more scheduled data packets
(priority bit 35P asserted).
FIG. 47 illustrates the flow chart for the select buffer controller 45D operation. The controller 45D is responsive to the common time reference (CTR) tick 002 at step 345-11, and then, at step 345-12, it increments the transmit buffer index i (i.e., i:=i+l mod l'*k', where T*k' is the number of buffers for scheduled traffic), and sends a time frame delimiter TFD to the serial transmitter 49. Then, if the transmit buffer B-i is not empty, at step 345-13, it wtil send a data packet from transmit buffer B-i, as specified in at step 345-14, else it wiU send a "best effort" data packet from the "best effort" buffer B-E, as specified at step 345-15. The flow chart for the program executed by the delay analysis and scheduUng controller is iUustrated in FIG. 48. The main task of the program is to compute the index, i, of the transmit buffer, B-i, between B-1 and B- T*k is computed in step 245- 05. There are several possible methods to perform the computation in step 245-05, which depends on the type of delay variations that can occur on the communication links. In FIGS. 49 and 50 three possible computation methods are described:
I. FIG. 49 - the case of continuous delay variations as described in FIGS. 40- 42, as specified in 45-051:
1. Let <sl, s2, s3, ... , s > be the set of time frames of a PED=/?, which repeats in every super cycle, as it is specified in the forwarding table 245B at the p entry,
2. Controller 245A searches the set <sl, s2, s3, ... , s/> in order to determine the first feasible time frame, si, that occur after (ToA 35T)+CONST (where
CONST is a constant bound on the delay across the switching fabric, and
3. si is the time frame the data packet is scheduled for transmission via the serial transmitter - where i is the index transmit buffer B-i.
The set <sl, s2, s3, ... , sj> constitute plurality of time frame in which a data packet can be scheduled for transmission out of the output port of a switch.
II. FIG. 50 - the case of multiple path with resynchronization as described in FIGS. 43, 44, and 45: When the data packet is received from the fabric at step 245-03, the PID 35C in the data packet header is used to look-up the resynchronization parameter 45R in the forwarding table (245B of FIG. 46, as specified in step 45-151, and then in step 45-152: Compute the index, i, of the transmit buffer 245C: i = [(ToA
35T)+45R] mod T*k' (where T*k' is the number of buffers for scheduled traffic). In the case of two virtual pipes: /? and /?', where one is an alternative to the other, as shown in FIGS. 43, 44, and 45, the above resynchronization is needed on both ends. More specifically resynchronization is needed on both Node A and Node E, as shown in FIGS. 43, 44, and 45. in. FIG. 51 - the case when using a time-stamp in the packet header, FIGS. 15 and 22: When the data packet is received from the fabric at step 45-03, the PID 35C in the data packet header is used to look-up the forward parameter 45F in the forwarding table (245B of FIG. 46), as specified in step 45-251, and then in step 45-252: Compute the index, i, of the transmit buffer 245C: i = [(Time-stamp 35TS)+45F] mod / ' *k '
(where l'*k' is number of buffers for scheduled traffic). 5 Monitoring, policing and billing
The present invention further relates to a system and method for monitoring, poticing and btiling of the transmission and forwarding of data packets over a packet switching network. The switches of the network 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 output port 40 is iUustrated in FIG. 16, comprised of a scheduUng controller with a transmit buffer 45, serial transmitter 49 (as previously described herein), and the monitoring and policing controllers. The scheduling controUer 45 performs a mapping of each of the data packets between the associated respective time of arrival (ToA) and
an associated forwarding time out of the output port via the serial transmitter 49. The forwarding time is determined relative to the common time reference (CTR) 002.
A general scheduUng controller 45 operation was previously described in FIGS. 19-21, which includes a transmit buffer 45C and a select buffer controller 45D. The data packet scheduling controUer 45A, together with the select buffer controUer 45D, perform the mapping, using the PID 35C and the data packet time of arrival (ToA) 35T in order to determine the respective time frame a respective packet should be forwarded out of the output port. Both controllers 45A and 45D are constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data, and read only memory (ROM) for storing the controUer processing program.
5.1 The monitoring and policing controllers
The monitoring and poticing controllers 65 (FIGS. 52-55) are part of both the input port in FIG. 12 and the output port in FIG. 16. Monitoring and policing controllers 65 are of two basic types: 1. The delay monitoring controUer 65D - for ensuring the correct timing behavior by PJD=p (FIGS. 52, 53). 2. The poticing and load controUer 65P - for ensuring the correct capacity usage by PID=/? (FIGS. 54, 55). Both controllers 65D and 65P are constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data, and read only memory
(ROM) for storing the controller processing program.
5.2 The delay monitoring controUer 65D
FIGS. 52 and 53 describe the operation of a delay monitoring controller 65D. This controUer checks data packets in which their reserved priority bit, 35P in their headers, is asserted for three cases:
1. Data packet is within two predefined delay parameters range (see box 65D- 05): between the two delay parameters: 65-par-L and 65-par-H, which were found PID=/? 35C entry in the parameters table 65-Tab (see box 65D-02). More specificaUy, the delay monitoring controUer 65D computes the actual delay the data packet already experienced: 65-Del = Time of arrival 35T -
Time- stamp 35TS (see box 65D-03), then comparing that it is in the predefined delay range: (65-Del > 65 - Par-L and 65-Del < 65 - Par-H) (see box 65D-04).
2. Data packet is late (see box 65D-07): its delay is greater than 65-par-H, i.e., 65-Del >65 - Par-H (see box 65D-06), and
3. Data packet is early (see box 65D-08): its delay is smaller than 65-par-L, i.e., 65-Del < 65 - Par-L.
The three cases have importance on ensuring proper network operations and the adherence to the user quality of service (QoS) requirements. Furthermore, the information coUected by the delay monitoring controller is reported to upper layer protocols, which are outside the scope of this invention. 5.3 The policing and load controUer 65P
FIGS. 54 and 55 describe the operation of an alternate embodiment of policing and load controUer 65P that checks and ensures that a data packets in which its reserved priority bit 35P (in its header) is asserted will not exceed the predefined load of its virtual pipe PID=p. This controller operation can be used for both: 1. Poticing - detecting PID=p that exceeds its reserved capacity, and
2. Billing - recording the actual capacity usage of PID=p.
The two cases have importance on ensuring proper network operations and the adherence to the user quality of service (QoS) requirements.
When a data packet is received (see box 65P-01) the poticing an load controller 65P first computes the current load, L(p) for PID=p (see box 65P-02) by L(p): =
L(p)+1 (see box 65P-02) using the load table 65L that stores previous values of L(p). The load can be computed in various ways: (i) per time frame of PID=p, (ii) per time cycle of PID=p, or (iii) per super cycle of PID=p.
Next the policing and load controller 65P using the PID=p 35C looks-up the parameter 65-Par in the table 65-Tab. Then if 65-Par>L(p) (see box 65P-03) the data packet is dropped (see box 65P-05), otherwise the data packet is forwarded (see box 65P-04).
In the above two cases, the load L(p) information on PID=p is recorded and reported to upper layer protocols for billing the usage for the usage of PTD=p. Furthermore, the policing and load information is used also for ensuring proper network operations and the adherence to the user quality of service (QoS) requirements. The information collected by the policing and load controller is reported to upper layer protocols, which are outside the scope of this invention. 6 Interconnecting a synchronous with an asynchronous switching networks The present invention further relates to a system and method for transmitting and forwarding packets over a heterogeneous packet switching network in which some of its some of its switches are synchronous and some switches are asynchronous. The invention specificaUy ensures that the synchronous switches wiU forward data packets in predefined time intervals although are arriving to the synchronous switches with relatively large, but bounded, delay uncertainty. Such delay uncertainty is the consequence of having asynchronous switches on the route of a data packet before it reaches the synchronous switch.
A system is provided for managing in a timely manner transfer of data packets across asynchronous switches with three configurations:
1. From an end-station 100 across asynchronous LAN switches 20 to a synchronous virtual pipe switch 10, as shown in FIG. 56. In this embodiment, an end-station 100 sends data packets from its network interface 102. The data packet is switched through one or more asynchronous switches 20 (e.g., Nodes B and C in FIG. 56) until it reaches the synchronous virtual pipe switch 10. In this configuration, the synchronous switch 10 wtil resynchronized the incoming data packet as it is explained below.
2. From an end-station 100 network interface 102 across asynchronous LAN switches 20 to an asynchronous to synchronous gateway 15, which converts the asynchronous data packets stream to a synchronous stream and then forward the data packets to a synchronous virtual pipe switch, as shown in FIG. 57.
3. From a first synchronous virtual pipe switch 10 across asynchronous switches or routers 20 to a second synchronous virtual pipe switch 10, as shown in FIG. 58. In this configuration, the incoming data packet to the second synchronous virtual pipe switch 10 wtil be resynchronized as it is explained below.
6.1 Resynchronization bv the router controller at the input port
A data packet that has been forwarded by one or asynchronous switches 20, i.e., switches without common time references can be resynchronized to its original schedule at either the input port or the output port. FIGS. 59-61 describe the resynchronization operation at the input port of either a synchronous virtual pipe switch 10 or a gateway
15.
In order to facilitates the resynchronization of data packets sent by the end- station 100, the synchronous virtual pipe switch 10 periodically sends timing messages M002 to the end-station 100, as shown in FIG. 59. The timing messages M002 represent the value of the common time reference (CTR) 002 as it is received by the synchronous switch 10.
The network interface 102 at the end-station 100 incorporates the timing information obtained from the timing message M002 into the time-stamp field 35TS (FIG. 15) in the header of the data packets, 1DP, 2DP, 3DP, ... , 9DP in FIG. 59, it sends across the asynchronous switches 20 to the synchronous virtual pipe switch 10.
FIG. 60 show the possible time of arrival of data packets, 1DP, 2DP, 3DP, ... , 9DP, to the input port 35. A data packet can experience delay that is ranging between minimum
delay and maximum delay, which constitute the delay bound uncertainty to be Maximum delay - Minimum delay, as shown in FIG. 60.
FIGS. 61A and 61B show the resynchronization operation performed by the routing controUer 35 at the input port 30 using a resynchronization buffer 30R, which is executed in the foUowing two steps:
1. Using the PID 35C look-up the delay bound uncertainty parameter 35DB (which is [Maximum delay - Minimum delay]) in the routing table 35B.
2. Delay the packet at the input port by: 35DB - [(CTR 002) - (Time-stamp 35TS)]. 6.2 The router controller operation with a gateway
When there is an asynchronous to synchronous gateway 15 before the input to virtual pipe switch the router controUer operates as was previously specified herein in FIG. 14. In this case, the gateway receives the CTR 002 and also forward time frame delimiter (TFD) 47A to the serial receiver, as it wtil be specified below. More specifically, its operation can be identical to the output operation which will be described below.
FIG. 14 iUustrates the flow chart for the router controller 35 processing program executed by the routing controller 35B. The program is responsive to two basic events from the serial receiver 31 of FIG. 14: the receive time frame delimiter TFD at step 35- 01, and the receive data packet at step 35-02. After receiving a TFD, the routing controUer 35 computes the time of arrival (ToA) 35T value at step 35-03 that is attached to the incoming data packets. For this computation it uses a constant, Dconst, which is the time difference between the common time reference (CTR) 002 tick and the reception of the TFD at time t2 (generated on an adjacent switch by the CTR 002 on that node). This time difference is caused by the fact that the delay from the serial transmitter 49 to the serial receiver 31 is not an integer number of time frames. When the data packet is received at step 35-02, the routing controUer 35B executes three operations as set forth in step 35-04: attach the ToA, lookup the address of the queue 36 using the PID, and storing the data packet in that queue 36. 6.3 The router controller operation without a gateway
When the virtual pipe switch 10 is connected directly to the asynchronous network the virtual pipe switch 10 cannot receive TFD, and therefore, the operation of the routing controller, as specified was previously specified herein in FIG. 14, should be changed in the foUowing way. The operations in 35-01 and 35-03 are taken out and in part one of the operation specified in 35-04, in FIG. 14, should clearly state that the time of arrival value attached to the packet header should be derived directly from the common time reference (CTR) 002.
6.4 The output port
The output operation described herein applies to both the synchronous virtual pipe switch 10, in FIGS. 56 and 58, and the asynchronous to synchronous gateway 15, in FIG. 57. The output port 40 was previously specified herein in FIG. 16, comprised of a scheduling controller with a transmit buffer 45, and serial transmitter 49 (as previously described herein). The scheduUng controller 45 performs a mapping of each of the data packets between the associated respective time of arrival (ToA) and an associated forwarding time out of the output port via the serial transmitter 49. The forwarding time is determined relative to the common time reference (CTR) 002. As it wtil be described, this mapping resynchronized the stream of data packets forwarded by the virtual pipe switch.
The scheduling controller and transmit buffer 45 has various modes of operation which are described in FIGS. 19, 48-49, and 62. The different operation modes corresponds to some of the possible configurations with the asynchronous switches, as was discussed in FIGS. 56-58. The scheduling controller and transmit buffer 45 in FIG. 19 includes three parts:
1. a delay analysis and scheduUng controller which further comprises a forwarding table 45B; 2. a transmit buffer 45C which is typically realized as a random access memory (RAM); and 3. a select buffer controUer 45D which forward data packets to the serial transmitter. The delay analysis and scheduling controller 45A, together with the select buffer controller 45D, perform the mapping, using the PID 35C, the time-stamp 35TS and the data packet time of arrival (ToA) 35T in order to determine the respective time frame a respective packet should be forwarded out of the output port. Both controllers 45A and 45D are constructed of a central processing unit (CPU), a random access memory (RAM) for storing the data, and read only memory (ROM) for storing the controUer processing program.
Data packets arrive from the switching fabric 50 via Unk 51. Data packets which have the priority bit 35P asserted (i.e., reserved traffic) are switched by the delay analysis and scheduling controUer 45A to one of the k transmit buffers 45C (B-1, B-2, ...., B-k). Each of the k buffers is designated to store packets that wtil be forwarded in each of the it time frames in every time cycle, as shown in FIGS. 5 and 6. Having it transmit buffers enables the delay analysis and scheduling controller 45A to schedule data packets in a wide range of delay variations.
The transmit buffer 45C includes an additional buffer B-E for "best effort" data packets. The priority bit 35P in the "best effort" data packets is not asserted and this is how the delay analysis and scheduling controller determine that such data packets should be stored in the "best effort" buffer. The "best effort" data packets are forwarded to the serial transmitter 49 whenever there are no more scheduled data packets
(priority bit 35P asserted).
6.5 The select buffer controUer operation:
FIG. 21 Ulustrates the flow chart for the select buffer controller 45D operation. The controller 45D is responsive to the common time reference (CTR) tick 002 at step 45-11, and then, at step 45-12, it increments the transmit buffer index i (i.e., i:=i+l mod k', where k' is the number of buffers for scheduled traffic) and sends a time frame delimiter TFD to the serial transmitter 49. Then, if the transmit buffer B-i is not empty, at step 45-13, it wtil send a data packet from transmit buffer B-i, as specified in at step 45-14, else it will send a "best effort" data packet from the "best effort" buffer B-E, as specified at step 45-15.
6.6 The delay analysis and scheduling controller
The flow chart for the program executed by the delay analysis and scheduUng controller is iUustrated in FIG. 48. The main task of the program is to compute the index, i, of the transmit buffer, B-i, between B-1 and B-k, is computed in step 245-05. There are several possible methods to perform the computation in step 245-05, which depends on the type of delay variations that can occur on the communication tinks. In FIGS. 49 and 62, two possible computation methods are described:
I. FIG. 49 - the case of arbitrary, but bounded, delay variations as described in FIGS. 56-58, as specified in 45-051: 1. Let <s 1 , s2, s3, ... , s > be the set of time frames of a PID=p, which repeats in every time cycle, as it is specified in the forwarding table 45B at the p entry,
2. controller 45A searches the set <sl, s2, s3, ... , s > in order to determine the first feasible time frame, si, that occur after (ToA 35T)+CONST (where CONST is a constant bound on the delay across the switching fabric); and
3. si is the time frame the data packet is scheduled for transmission via the serial transmitter - where i is the index transmit buffer B-i.
The set <s 1, s2, s3, ... , s > constitute plurality of time frame in which a data packet can be scheduled for transmission out of the output port of a switch.
II. FIG. 62 - the case when using a time-stamp in the packet header (FIGS. 15 and 22): When the data packet is received from the fabric at step 45-03, the PID 35C in
the data packet header is used to look-up the forward parameter 45F in the forwarding table (45B of FIG. 19), as specified in step 45-351, and then in step 45-352: Compute the index, i, of the transmit buffer 45C: i = [(Time-stamp 35TS)+45F] mod k', where it' is the number of buffers for scheduled traffic.
From the foregoing, it wtil be observed that numerous variations and modifications may be effected without departing from the spirit and scope of the invention. It is to be understood that no limitation with respect to the specific apparatus illustrated herein is intended or should be inferred. It is, of course, intended to cover by the appended claims all such modifications as fall within the scope of the claims.