Isochronous link protocol - Patent Review 5406559

Description Submit all comments and votes

The present invention relates to a data communication network, such as a local area network or wide area network, and in particular to a network for transferring isochronous data.

BACKGROUND OF THE INVENTION

Isochronous data can generally be described as data which is non-packetized and of indeterminate, potentially continuous duration. Examples of isochronous data sources include video cameras, which output a substantially continuous stream of data representing images and associated sounds, and telephones, which output a substantially continuous stream of voice data. An example of an isochronous data sink is a video monitor which receives a substantially continuous stream of video data for display.

FIG. 1A schematically depicts isochronous data transfer. The data transfer is first initiated, such as by initiating a telephone conversation or beginning a video camera transmission 12. After initiating the data transfer, transmission of the data is provided substantially continuously for an indeterminate period, until termination of the telephone conversation or video transmission 14. Every bit transferred need not necessarily represent a data bit. "Housekeeping" bits to control destination and timing may be also transferred. Furthermore, the data being transferred may comprise "Null" data such as silence during a telephone conversation or transfer of a blank video image. One type of isochronous data transfer is the Fiber Distributed Data Interface-II (FDDI-II) as described, for example, in FDDI-II Hybrid Multiplexer, Revision 2.4, dated Mar. 25, 1991.

The increasing availability of multi-media computers and work stations that include isochronous data sources and sinks in addition to non-isochronous sources and sinks has increased interest in the transfer of isochronous data in a network environment. Many existing networks use non-isochronous data communications between stations on the network. Commonly used data transfer protocols include packet transfer systems and token ring systems.

An example of packetized data transfer is the commonly-used ethernet system. One implementation known as 10BASE-T is described in the draft Nine supplement to IEEE standard 802.3, dated Nov. 15, 1989. FIG. 1B shows a packet transfer 22.

In a token ring system, a node transfers data only upon receipt of an electronic token. One commonly used token ring system is described in IEEE standard 802.5 FIG. 1C shows a data transfer 23 in a token ring system.

Many previous attempts to accommodate isochronous data on these existing data networks result in disadvantageous operating characteristics. In some systems, the bandwidth available for a given isochronous source or sink degrades in direct proportion to the total number of isochronous data sources and sinks transmitting and receiving on the network. Also, the presence of isochronous sources and sinks degrades the non-isochronous bandwidth. Furthermore, in any isochronous system employing bidirectional links, the link cable length introduces a skew between data transmitted from the node to the hub and data received from the node by the hub. These delays introduce undesirable jitter and can be disconcerting to users of video and telephonic data.

In addition, existing isochronous systems also provide little or no compatibility with previous networks. This incompatibility necessitates extensive replacement of hardware or software to accommodate both isochronous and non-isochronous traffic. Thus a multi-media PC having ethernet capabilities and a video camera cannot simultaneously utilize both the isochronous and non-isochronous source/sinks.

SUMMARY OF THE INVENTION

Copending application Ser. No. 07/969,916, Attorney Docket No. 8332-314/NS-2023, titled "Network for Data Communication with Isochronous Capability" filed the same day herewith, and incorporated by reference, describes a system that provides for communication of data to and from isochronous data sources and sinks. The bandwidth available to an isochronous source/sink is independent of changes in non-isochronous demand on the network. Furthermore, each source/sink is guaranteed an isochronous bandwidth which is independent of changes in source/sink bandwidth demands on the network. The isochronous communication system also maintains a high degree of compatibility with previous, often in-place, systems, thus requiring only minimal replacement of hardware/software.

The system is implemented as a star-topology network with data sources transmitting to a central hub which, in turn, transmits the data to data sinks. Several such star-topology systems can be connected by inter-connection of the hubs, for example, in a ring structure. Multiplexed data arriving at the hub is de-multiplexed to separate the isochronous-source data, the non-isochronous-source data and D channel and M channel information. The non-isochronous-source data can be provided to hub circuitry specialized for handling the non-isochronous data stream. Preferably, circuitry in the hub converts the separated non-isochronous data stream into a form substantially similar to the form available over previous non-isochronous networks. For example, where non-isochronous data is sourced from an ethernet MAC, the hub converts the separated non-isochronous data to a form handled by standard ethernet hub repeater circuitry.

According to one embodiment of the present invention, the hub-node system includes delay circuitry to compensate for cycle misalignments caused by system cabling. The delay system operates to minimize the amount of data buffering required. In the delay system of the present invention, the hub times the delay between the transmission of the cycle start and the arrival of the received cycle start. The hub thus senses the adjustment necessary in the node's cycle and outputs a control signal to the node. At the node, a delay circuit stores the delay value and delays the start of the cycle reference provided to the node transmitter. The node transmitter thus outputs cycles that arrive at the hub coincident with the beginning of a cycle thereby minimizing data skew.

According to another embodiment of the invention, the node delay circuitry comprises a latch controlled by the hub. The latch triggers when the cycle reference is provided to the node transmitter and thereby aligns the transmitted frames as desired.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a timing chart of an isochronous data transmission;

FIG. 1B is a timing chart for a packetized data transmission;

FIG. 1C is a timing chart for a token ring data transfer;

FIG. 2A is a diagram of a star and ring topology communication system according to an embodiment of the present invention;

FIG. 2B is a diagram of a star and ring topology communication system having multiple isochronous circuitry within a single hub according to an embodiment of the present invention;

FIG. 2C is a diagram of a tree topology communication system according to an embodiment of the present invention;

FIG. 3 is a communication system configured according to an embodiment of the present invention;

FIG. 4 is a block diagram of node circuitry according to an embodiment of the present invention;

FIG. 5 is a block diagram of hub receiver circuitry according to an embodiment of the present invention;

FIG. 6 is a block diagram of a receive interface for non-isochronous data according to an embodiment of the present invention;

FIG. 7 is a diagram of a receive interface for isochronous data and associated hub circuitry according to an embodiment of the present invention;

FIG. 8 is a block diagram of a hub transmit interface for non-isochronous data according to an embodiment of the present invention;

FIG. 9 is a block diagram of a hub transmitter interface for non-isochronous data according to an embodiment of the present invention;

FIG. 10 is a timing chart for coordinating data transfers according to an embodiment of the present invention;

FIG. 11 is a block diagram of node circuitry having a delay circuit according to an embodiment of the present invention;

Table I is a tabulation of a time division multiplexing scheme for multiplexing data streams according to an embodiment of the present invention; and

Table II lists a form of four/five encoding according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

The detailed description of this invention is given in the context of a data transfer system that supports both non-isochronous and isochronous communication. The description provided herein therefore enables a discussion of the invention in network systems employed for:

1) transfers between a non-isochronous source and non-isochronous sink connected to the same hub; and

2) transfers between an isochronous source and an isochronous sink connected to the same hub.

The description therefore illustrates several of the more common situations in which the invention might be used.

General Overview of System Operation

Pending application Ser. No. 07/969,916, Attorney Docket No. 8332-314/NS2023, titled "Network for Data Communication with Isochronous Capability", filed the same day herein and incorporated by reference, describes a data communication system for isochronous data that can be configured in a star-topology and interconnected in a ring or tree topology. Such a system is shown in FIGS. 2A, 2B or 2C. In the configuration depicted in FIG. 2A, the hubs are connected in a ring-topology with first hub 44a sending data to the second hub 44b, the second hub 44b sending data to the third hub 44c, the third hub sending data back to the first hub 44a via a cycle generator and latency adjustment circuitry. The inter-hub connections over a Time Slot Interchange (TSI) ring 58f. In one embodiment, an FDDI-II system can be used as the TSI ring 58f. FIG. 2B shows hubs 44a, 44b and 44c arranged in a star and ring topology having multiple isochronous circuitry within a single hub. FIG. 2C shows a tree topology communication system. The parent hub 44a connects to a high bandwidth backbone. Hub 44b operates as a child hub of parent hub 44a and is attached at port 2 of hub 44a. Child hub 44a cascades from child hub 44b.

The star and ring topology includes a plurality of nodes 42a, 42b, 42c attached to a single hub operating on the high bandwidth bus. The exact number of nodes varies depending on the data transmission needs and objectives of the system. Each of the nodes 42a-42c can include various types of sources and sinks such as strictly isochronous sources and sinks, strictly non-isochronous sources/sinks or both isochronous and non-isochronous sources and sinks. Data links comprising physical data transmission media, such as one-way twisted pair wires 46, couple each node to one of hubs 44a-44c.

FIG. 3 shows hub 44a and associated nodes 42a-42c in greater detail. FIG. 3 may, by itself, form a complete star topology system. Each node 42a, 42b, 42c includes circuitry 50a, 50b, 50c. Circuitry 50a-c receives data, for conversion to a form suitable for transmission onto the physical media 46a, 46c, 46e; and receives signals from the physical media 46b, 46d, 46f for conversion to a form suitable for use by the data sinks.

Hub 44a includes circuitry 54a, 54b, 54c for receiving data from the physical media 46a, 46c, 46e; separating the isochronous-sourced data from the non-isochronous-sourced data and the D channel and maintenance channel data; and converting separated data into a form suitable for handling by downstream hub circuitry 56. In the depicted embodiment the separated isochronous-sourced data is provided to isochronous switching circuitry such as a time slot interchange controller 58 for placing the data on a TSI bus so that it can be transported to and retrieved by other equivalent circuitry 54a-54c in the hub for transmission to various destination nodes 42a-42c to other hubs. The separated non-isochronous data is provided to circuitry 60 configured to convey the non-isochronous data for transmission to destination nodes 42a, 42b, 42c. In an embodiment in which non-isochronous-sourced data includes ethernet data, the hub circuitry 60 can be a standard ethernet repeater processor. In this way, the system can be at least partially backwards-compatible with previous ethernet hub systems.

The D channel and maintenance data is provided to a signaling processor 62. Signaling processor 62 performs various maintenance and control functions such as identifying and alerting users of error conditions, and setting up requested connections, i.e. source/destination paths e.g. by communicating with the isochronous and non-isochronous controllers 58, 60, e.g. over data path 64.

The operation of the components described above may be understood by describing the transfer of data from video camera, isochronous source, 48d to isochronous sink 48b and from Ethernet Mac, non-isochronous source 48c to non-isochronous sink 48g. Data sent from isochronous device 48d is a continuous stream of digitized data having, for example, a rate equal to the American "T1" standard of 1.544M6PS. Data output from the ethernet MAC 48c is provided at the standard 10BASE-T ethernet rate of 10 Mb/sec. D channel information is provided from a D channel data stream source preferably contained in a MAC or other circuitry in the system, or for example, from the virtual key pad 48f at a variable data rate, such as a rate not exceeding about 64 Kb/sec.

Lines 66a, 66b, 66c carry the data streams from sources 48d and 48c to node circuitry 50b. FIG. 4 shows circuitry 50b in greater detail. Node circuitry 50b comprises hardware that operates on the incoming data stream to enable efficient, compatible transmission between the data source and destination. A multiplexer 70 time-division multiplexes the incoming data on a four-bit basis using a repeating series of frames or templates. In this embodiment, the frames are repeated every 125 micro seconds.

Table I tabulates a scheme for time division multiplexing the various data streams, additional data and control bytes. Each symbol in Table I represents four bits of data and therefore every group of two symbols represents one 8-bit byte of data. In Table I, E represents four bits of data from the ethernet stream 66a, B designates four bits of data from the isochronous stream 66b and D represents four bits of data from the D channel stream 66c. M represents M channel data, which preferably is provided by circuitry 50b. In addition, certain byte-length patterns are provided. JK represents a frame synchronization pattern and EM (the first two bytes of block three in Table I) represents an ethernet "pad" followed by a maintenance byte.

As seen in Table I, each frame contains 256 bytes which can be considered in thirty-two groups of eight bytes each, or four blocks of sixty-four bytes each. For a data rate output from the isochronous source 48d of 1.544 Mb/sec. The frame structure described provides an isochronous bandwidth capability of 6.144 Mb/sec. Thus, the single isochronous source 48b in the present example can be entirely accommodated using only 48 of the 192 "B" symbols per frame. A basic rate ISDN Channel could be supported by using three 64 Kb/s slots within the isochronous channel. Thus, a variety of isochronous sources may be allocated among the available isochronous bandwidth. The frame structure is described more thoroughly in commonly-assigned application Ser. No. 07/969,911 (Attorney File No. 8332-315/NS-2024), titled "Network for Transmitting Isochronous-Source Data with a Frame Structure" filed on even date herewith and incorporated herein by reference. Frame structures other than that described could be used to provide an allocation of bandwidth suited for the particular purpose.

The time-multiplexed data is then encoded by an encoder 72 to maintain the AC balance of the cable which can be potentially upset by an extended string of binary zeros. In the depicted embodiment, the encoder performs four/five encoding. One particular form of four/five encoding, conforming partially to the ANSII X3T9.5 standard, is described by Table II. These patterns, when properly combined, have a maximum of three bit times with no transition. The encoding scheme depicted in Table II is described in greater detail in commonly-assigned application Ser. No. 07/970,329 (Attorney File No. 8332-316/NS-2022), titled "Frame-Based Transmission of Data" filed on even date herewith and incorporated herein by reference.

The results of the four/five encoding is then further encoded by encoder 74 of FIG. 4 using a non-return to zero, inverted (NRZI) scheme. The four/five-NRZI encoding is particularly useful in networks in which a non-isochronous source is a 10BASE-T ethernet source because the encoding provides for transmission at a signaling rate substantially compatible with the data rates provided and expected by the ethernet MAC. Other types of encoding or decoding, however, can also be used such as a scheme encoding 8 bits into 10 bits.

After encoding, the data is sent to pre-emphasis circuitry 76 and to a transmitter or driver 78b. Pre-emphasis circuitry 76 compensates the signal transmitted onto the physical medium to reduce jitter. The signal is then transmitted over the physical medium 46c to hub 44a which can include twisted pair, coaxial or fiber optic cable.

Hub 44a as seen in FIG. 3 comprises a plurality of circuit devices 54a, 54b, 54c, each one coupled to one of nodes 42a, 42b, 42c by physical media 46. As depicted in FIG. 5, the data transmitted over the physical media 46 arrives serially at a de-serializer/decoder 80. De-serializer/decoder 80 includes circuitry which is functionally an inverse of the multiplexing/encoding circuitry described above and operates to decode the four/five NRZI encoding and to separate the isochronous and non-isochronous sourced data. De-serializer/decoder 80 also outputs a synchronization signal, derived from the JK frame synchronization symbols 96 for use by a framing timing generator 98. Link detect circuitry 82 also receives the data from the physical media 46 for detection of the mode in which the node is operating (e.g. 10BASE-T, isochronous ethernet or isochronous) and outputting a mode select signal, as described more fully in commonly-assigned application Ser. No. 07/971,018 (Attorney File No. 8332-319/NS-2027, titled "Network Link Endpoint Capability Detection," filed on even date herewith and incorporated herein by reference.)

Both the non-isochronous-sourced data 94b and the isochronous-sourced data 94a are made available to the various hub circuitry components 54a, 54b, 54c, as needed for transmission to destination nodes. In one embodiment, the separated isochronous data 94a and non-isochronous data 94b are reconfigured by the respective interfaces 58, 60 to provide isochronous output 102 and non-isochronous output 104 in a form suitable for transmission to the destination nodes. In one embodiment, the non-isochronous data 94b is configured by the E interface 60 so that the output data 104 can be processed by a repeater device for provision to hub circuitry 54 and eventual transmission to destination nodes. As an alternative to using a repeater for the non-isochronous data, packet connections may be linked through media access control layer bridges.

FIG. 6 depicts one implementation of an E interface 59 of a type which will receive the non-isochronous data 94b and provide outputs 106, 108 of a type that can be processed by previously-available repeater circuitry 60. The non-isochronous data is received in a first-in-first-out (FIFO) buffer 112 to smooth out data rates. Circuitry 114 detects "no carrier" symbols, provided to emulate ethernet data packets, used by logic circuitry or state machine 116 to output carrier detect signals. The output 118 from FIFO 112 is provided to a multiplexer 120 and a de-serializer 122 to produce data output 106. Multiplexer 120 can receive a preamble stream 124 to provide the proper preamble bits in the output data 106. Output 118 from FIFO 112 is also provided to decode circuitry 128 to recognize data collision and alignment error symbols and output appropriate signals 130, 132 to state machine 116. Operation and components of the receive interface 59 are described more thoroughly in application Ser. No. 07/970,329 (Attorney File No. 8332-316/NS-2022), titled "Frame-Based Transmission of Data".

For purposes of the present example, data from isochronous-source 48d (FIG. 7) is assumed transmitted in the first 24 Isochronous bytes of each frame represented by the "B" symbols in block 0 of Table I (i.e. the first forty-eight "B" symbols in the frame structure). FIG. 5 depicts a B interface 58 according to one embodiment of the invention. In the embodiment of FIG. 7 the separated isochronous data 94a is stored in one of two buffers 132a, 132b. The timing of storage in the buffers 132a, 132b is coordinated with the 125 microsecond frame transmission timing so that data 94a from a first frame will be stored in the first buffer 132a during a first period of 125 microseconds and, during the next 125 microsecond period, the isochronous data 94a from the next frame will be stored in the second buffer 132b. In one embodiment, the data can be stored in the buffer 132 in the same order it is received, such that the eight bits represented by the first two "B" symbols in Table I is stored in the first storage location of buffer 132a, that corresponding to the second two "B" symbols in Table I is stored in the second location of buffer 132a and so on. Since the frame structure depicted in Table I contains 96 bytes of isochronous data per frame, each of the buffers 132a, 132b has the capacity to store 96 bytes of data per node supported. After isochronous data from a first frame is stored in buffer 132, during the next 125 microsecond period (while the data from the next frame is being stored in the second buffer 132b) data which was stored in the first buffer 132a may be transmitted onto a high bandwidth bus 134. The loading and ordering of the buffer 132 is dependent upon the number of nodes supported by hub 44a. Bus 134 has sufficient bandwidth to carry the isochronous data output from a plurality of nodes which are connected to the hub 44a. In an embodiment in which the hub 44a is connected to 16 nodes, the bandwidth of the bus 134 must be sufficient to receive 1.536 bytes of data (i.e. 16 nodes.times.96 bytes per node) every 125 microseconds (i.e. every frame). This corresponds to a bandwidth of about 98304 Kb/sec.

Depending upon aspects of the system configuration, such as the number of nodes attached to a hub and the bandwidth dedicated to isochronous data, other embodiments of the invention could be provided with other bandwidths in place of the TSI bus 134. However, the 98304 Kb/sec bandwidth is particularly useful because it substantially matches the bandwidth employed in FDDI-II, making it particularly easy to port the data on the TSI bus 134 to a TSI ring 58 (FIG. 3) in configurations where the TSI ring 58 is an FDDI-II system.

According to one embodiment, the data is conveyed from the buffer 132 into a time slot on the bus 134 in a time slot interchange fashion. Data carried on TSI bus 134 is transmitted in 125 microsecond time frames divided into 1.536 time slots, each of which has a duration of about 0.08138 microseconds. Each time slot has data and associated control and parity. Thus a byte could represent 10 bits of time slot information. Thus, data from buffer 132a may be placed onto TSI bus 134, by transmitting a given one of the 1.536 bytes stored in 132a on TSI bus 134 at the proper one of the 1.536 time slots of the 125 microsecond frame. Which of the time slots is "proper" depends on the use which is to be made of the data and, in particular, the destination for the data as predetermined in the connection set-up via the D channel.

The destination for data, in the depicted embodiment, has been pre-established using the D channel information. The D channel information is sent to a signaling processor 138. The D channel information, which includes source, destination, and other needed information, is used to store values in preferably a switch table 140. In one example, switch table 140 may be divided into sixteen sections 142a-142p corresponding to the sixteen nodes associated with the hub circuitry 58 of this example. Each section 142 contains 1.536 bits, corresponding to the 1.536 time slots in a TSI bus time frame. These bits can be used as a control 144 for a multiplexer 146.

In the present example, the twenty-four bytes of data from 48d per 125 microsecond frame are conveyed in the first twenty-four B slots of each 48d frame. Thus, the data from source 48d will be stored in the isochronous data buffer 132. The destination for the isochronous data of this example is monitor 48b. Thus, the 24 B slots of data will be transferred to data buffer 154a and then on the next frame transmitted to 48b in its corresponding first 24 B slots.

The 24 B slots could have been destined for the TSI bus in which case the 24 B slots in 132 would have been switched onto the TSI bus. A bit of the contents of the switch table would have controlled line 150 to control the multiplexer 146 at a rate of one bit for every TSI time slot (i.e., one bit every 0.080 microseconds). Assuming the first 10 time slots of the TSI bus do not receive the B data which is destined for a nodes attached to another hub, during the first TSI time slot, the multiplexer control 144 will be "0" and no data will be output from the buffer 132 onto bus 134. The multiplexer 146 will merely convey along the TSI bus 134 whatever data was already on the TSI bus in the first time slot. This continues until the 11th time slot of the TSI bus, at which time the B data destined for a node attached to another hub begins to be output onto the TSI bus. During each of the next 24 TSI bus time slots, the control signal for multiplexer 146 will be "1" and a byte of data stored in the appropriate data location of buffer 132 will be output through multiplexer 146 onto the bus 134. Which data location of the buffer 132 is "appropriate" can be determined by a read pointer contained in the switch table. Preferably, buffer 132 is a read access memory RAM and the read pointer will be determined according to the contents of the switch table, on representing the TSI slot time. After completion of conveying the 24 bytes onto the TSI bus, there will be no output from the buffer 132a during subsequent time slots of this TSI frame since in this example no other connections were established. In this way, time slots 11 through 35 for a frame on the TSI bus will be filled with data stored in the buffer 132a, i.e. the 24 bytes