Joint Flow Control Mechanism in a Telecommunications Network

This application relates to communications methods and apparatus in a distributed switching architecture, and in particular to a joint flow control mechanism in a distributed switching architecture.

A Flow Controlled Virtual Connection (FCVC) protocol for use in a distributed switching architecture is presently known in the art, and is briefly discussed below with reference to FIG. 1. This protocol involves communication of status (buffer allocation and current state) on a per virtual connection, such as a virtual channel connection or virtual path connection, basis between individual upstream and downstream network elements to provide a "no cell loss" guarantee. A cell is the unit of data to be transmitted. Each cell requires a buffer to store it.

One example of this protocol involves a credit-based flow control system, where a number of connections exist within the same link with the necessary buffers established and flow control monitored on a per-connection basis. Buffer usage over a known time interval, the link round-trip time, is determined in order to calculate the per-connection bandwidth. A trade-off is established between maximum bandwidth and buffer allocation per connection. Such per-connection feedback and subsequent flow control at the transmitter avoids data loss from an inability of the downstream element to store data cells sent from the upstream element. The flow control protocol isolates each connection, ensuring lossless cell transmission for that connection.

Connection-level flow control results in a trade-off between update frequency and the realized bandwidth for the connection. High update frequency has the effect of minimizing situations in which a large number of receiver cell buffers are available, though the transmitter incorrectly believes the buffers to be unavailable. Thus it reduces the number of buffers that must be set aside for a connection.

Under certain circumstances, it is desirable to model a single transmitter queue as a Virtual Path Connection (VPC) providing data cells to plural queues associated with a single receiver processor, each receiver queue regarded as Virtual Channel Connection (VCC). Prior art protocols fail, however, to enable connection-level flow control elements associated with the single transmitter queue to jointly flow control the plural receiver queues. Rather, each connection from transmitter queue to receiver queue must be individually flow controlled, resulting in a redundant and bandwidth-inefficient control mechanism.

The presently claimed invention provides a mechanism for grouping plural queue control elements in a downstream receiver into a single, joint set of flow control elements. This is particularly desirable in the situation where a single transmitter queue is the source of data cells communicated to plural queues all associated with the same receiver processor. While the prior art requires that each such connection be flow controlled distinctly, thus resulting in the maintenance of duplicative flow control elements and the consumption of bandwidth through the requirement of transmitting parallel update and check information for each connection, the presently disclosed mechanism enables such plural downstream queues to be flow controlled as a unit.

Each queue has associated therewith a data set known as a queue descriptor. Among the elements of a queue descriptor in a receiver are counters and register which are used in connection-level flow control. The queue descriptor further includes fields useful for indicating that certain flow control information is tracked in a queue descriptor of another queue and for pointing to that queue descriptor. Thus, plural queues are flow controlled through the use of counters in a common queue descriptor.

The use of such a mechanism results in a reduced frequency of check and update cell transfer, thus reducing overhead bandwidth, while maintaining a desired characteristic of connection-level flow control, zero cell loss.

In alternative embodiments, the presently disclosed mechanism is also used in conjunction with link-level flow control of a virtual link, with a prioritized buffer mechanism providing prioritized access to a shared buffer resource, and with a mechanism capable of guaranteeing a minimum, allocated bandwidth to a connection.

The above and further advantages may be more fully understood by referring to the following description and accompanying drawings of which:

FIG. 1 is a block diagram of a connection-level flow control apparatus as known in the prior art;

FIG. 2 is a block diagram of a link-level flow control apparatus according to the present invention;

FIGS. 3A and 3B are flow diagram representations of counter initialization and preparation for cell transmission within a flow control method according to the present invention;

FIG. 4 is a flow diagram representation of cell transmission within the flow control method according to the present invention;

FIGS. 5A and 5B are flow diagram representations of update cell preparation and transmission within the flow control method according to the present invention;

FIGS. 6A and 6B are flow diagram representations of an alternative embodiment of the update cell preparation and transmission of FIGS. 5A and 5B;

FIGS. 7A and 7B are flow diagram representations of update cell reception within the flow control method according to the present invention;

FIGS. 8A, 8B and 8C are flow diagram representations of check cell preparation, transmission and reception within the flow control method according to the present invention;

FIGS. 9A, 9B and 9C are flow diagram representations of an alternative embodiment of the check cell preparation, transmission and reception of FIGS. 8A, 8B and 8C;

FIG. 10 illustrates a cell buffer pool according to the present invention as viewed from an upstream element;

FIG. 11 is a block diagram of a link-level flow control apparatus in an upstream element providing prioritized access to a shared buffer resource in a downstream element according to the present invention;

FIGS. 12A and 12B are flow diagram representations of counter initialization and preparation for cell transmission within a prioritized access method according to the present invention;

FIGS. 13A and 13B illustrate alternative embodiments of cell buffer pools according to the present invention as viewed from an upstream element;

FIG. 14 is a block diagram of a flow control apparatus in an upstream element providing guaranteed minimum bandwidth and prioritized access to a shared buffer resource in a downstream element according to the present invention;

FIGS. 15A and 15B are flow diagram representations of counter initialization and preparation for cell transmission within a guaranteed minimum bandwidth mechanism employing prioritized access according to the present invention;

FIG. 16 is a block diagram representation of a transmitter, a data link, and a receiver in which the presently disclosed joint flow control mechanism is implemented; and

FIG. 17 illustrates data structures associated with queues in the receiver of FIG. 16.

In FIG. 1, the resources required for connection-level flow control are presented. As previously stated, the illustrated configuration of FIG. 1 is presently known in the art. However, a brief discussion of a connection-level flow control arrangement will facilitate an explanation of the presently disclosed link-level flow control method and apparatus.

One link 10 is shown providing an interface between an upstream transmitter element 12, also known as an UP subsystem, and a downstream receiver element 14, also known as a DP subsystem. Each element 12, 14 can act as a switch between other network elements. For instance, the upstream element 12 in FIG. 1 can receive data from a PC (not shown). This data is communicated through the link 10 to the downstream element 14, which in turn can forward the data to a device such as a printer (not shown). Alternatively, the illustrated network elements 12, 14 can themselves be network end-nodes.

The essential function of the presently described arrangement is the transfer of data cells from the upstream element 12 via a connection 20 in the link 10 to the downstream element 14, where the data cells are temporarily held in cell buffers 28. Cell format is known, and is further described in "Quantum Flow Control", Version 1.5.1, dated Jun. 27, 1995 and subsequently published in a later version by the Flow Control Consortium. In FIG. 1, the block labelled Cell Buffers 28 represents a set of cell buffers dedicated to the respective connection 20. Data cells are released from the buffers 28, either through forwarding to another link beyond the downstream element 14, or through cell utilization within the downstream element 14. The latter event can include the construction of data frames from the individual data cells if the downstream element 14 is an end-node such as a work station.

Each of the upstream and downstream elements 12, 14 are controlled by respective processors, labelled UP (Upstream Processor) 16 and DP (Downstream Processor) 18. Associated with each of the processors 16, 18 are sets of buffer counters for implementing the connection-level flow control. These buffer counters are each implemented as an increasing counter/limit register set to facilitate resource usage changes. The counters of FIG. 1, described in further detail below, are implemented in a first embodiment in UP internal RAM. The counter names discussed and illustrated for the prior art utilize some of the same counter names as used with respect to the presently disclosed flow control method and apparatus. This is merely to indicate the presence of a similar function or element in the prior art with respect to counters, registers, or like elements now disclosed.

Within the link 10, which in a first embodiment is a copper conductor, multiple virtual connections 20 are provided. In an alternative embodiment, the link 10 is a logical grouping of plural virtual connections 20. The number of connections 20 implemented within the link 10 depends upon the needs of the respective network elements 12, 14, as well as the required bandwidth per connection. In FIG. 1, only one connection 20 and associated counters are illustrated for simplicity.

First, with respect to the upstream element 12 of FIG. 1, two buffer state controls are provided, BS.sub.-- Counter 22 and BS.sub.-- Limit 24. In a first embodiment, each are implemented as fourteen bit counters/registers, allowing a connection to have 16,383 buffers. This number would support, for example, 139 Mbps, 10,000 kilometer round-trip service. The buffer state counters 22, 24 are employed only if the connection 20 in question is flow-control enabled. That is, a bit in a respective connection descriptor, or queue descriptor, of the UP 16 is set indicating the connection 20 is flow-control enabled.

BS.sub.-- Counter 22 is incremented by the UP 16 each time a data cell is transferred out of the upstream element 12 and through the associated connection 20. Periodically, as described below, this counter 22 is adjusted during an update event based upon information received from the downstream element 14. BS.sub.-- Counter 22 thus presents an indication of the number of data cells either currently being transmitted in the connection 20 between the upstream and downstream elements 12, 14, or yet unreleased from buffers 28 in the downstream element 14.

BS.sub.-- Limit 24 is set at connection configuration time to reflect the number of buffers 28 available within the receiver 14 for this connection 20. For instance, if BS.sub.-- Counter 22 for this connection 20 indicates that twenty data cells have been transmitted and BS.sub.-- Limit 24 indicates that this connection 20 is limited to twenty receiver buffers 28, the UP 16 will inhibit further transmission from the upstream element 12 until an indication is received from the downstream element 14 that further buffer space 28 is available for that connection 20.

Tx.sub.-- Counter 26 is used to count the total number of data cells transmitted by the UP 16 through this connection 20. In the first embodiment, this is a twenty-eight bit counter which rolls over at OxFFFFFFF. As described later, Tx.sub.-- Counter 16 is used during a check event to account for errored cells for this connection 20.

In the downstream element 14, the DP 18 also manages a set of counters for each connection 20. Buffer.sub.-- Limit 30 performs a policing function in the downstream element 14 to protect against misbehaving transmitters. Specifically, the buffer.sub.-- limit register 30 indicates the maximum number of cell buffers 28 in the receiver 14 which this connection 20 can use. In most cases, BS.sub.-- Limit 24 is equal to Buffer.sub.-- Limit 30. At some point, though, it may be necessary to adjust the maximum number of cell buffers 28 for this connection 20 up or down. This function is coordinated by network management software. To avoid the "dropping" of data cells in transmission, an increase in buffers per connection is reflected first in Buffer.sub.-- Limit 30 prior to BS.sub.-- Limit 24. Conversely, a reduction in the number of receiver buffers per connection is reflected first in BS.sub.-- Limit 24 and thereafter in Buffer.sub.-- Limit 30.

Buffer.sub.-- Counter 32 provides an indication of the number of buffers 28 in the downstream element 14 which are currently being used for the storage of data cells. As described subsequently, this value is used in providing the upstream element 12 with a more accurate picture of buffer availability in the downstream element 14. Both the Buffer.sub.-- Limit 30 and Buffer.sub.-- Counter 32 are fourteen bits wide in the first embodiment.

N2.sub.-- Limit 34 determines the frequency of connection flow-rate communication to the upstream transmitter 12. A cell containing such flow-rate information is sent upstream every time the receiver element 14 forwards a number of cells equal to N2.sub.-- Limit 34 out of the receiver element 14. This updating activity is further described subsequently. In the first embodiment, N2.sub.-- Limit 34 is six bits wide.

The DP 18 uses N2.sub.-- Counter 36 to keep track of the number of cells which have been forwarded out of the receiver element 14 since the last time the N2.sub.-- Limit 34 was reached. In the first embodiment, N2.sub.-- Counter 36 is six bits wide.

In a first embodiment, the DP 18 maintains Fwd.sub.-- Counter 38 to maintain a running count of the total number of cells forwarded through the receiver element 14. This includes buffers released when data cells are utilized for data frame construction in an end-node. When the maximum count for this counter 38 is reached, the counter rolls over to zero and continues. The total number of cells received by the receiver element 14 can be derived by adding Buffer.sub.-- Counter 32 to Fwd.sub.-- Counter 38. The latter is employed in correcting the transmitter element 12 for errored cells during the check event, as described below. Fwd.sub.-- Counter 38 is twenty-eight bits wide in the first embodiment.

In a second embodiment, the DP 18 maintains Rx.sub.-- Counter 40, a counter which is incremented each time the downstream element 14 receives a data cell through the respective connection 20. The value of this counter 40 is then usable directly in response to check cells and in the generation of an update cell, both of which will be described further below. Similar to the Fwd Counter 38, Rx.sub.-- Counter 40 is twenty-eight bits wide in this second embodiment.

There are two events in addition to a steady state condition in the connection-level flow controlled protocol: update; and check. In steady state, data cells are transmitted from the transmitter element 12 to the receiver element 14. In update, buffer occupancy information is returned upstream by the receiver element 14 to correct counter values in the transmitter element 12. Check mode is used to check for cells lost or injected due to transmission errors between the upstream transmitter and downstream receiver elements 12, 14.

In the accompanying figures, connection level counters are augmented with "[i]" to indicate association with one connection [i] of plural possible connections.

Prior to any activity, counters in the upstream and downstream elements 12, 14 are initialized, as illustrated in FIG. 3A. Initialization includes zeroing counters, and providing initial values to limit registers such as Link.sub.-- BS.sub.-- Limit and Link.sub.-- Buffer.sub.-- Limit. In FIG. 3A, Buffer.sub.-- Limit[i] is shown being initialized to (RTT*BW)+N2, which represents the round-trip time times the virtual connection bandwidth, plus accommodation for delays in processing the update cell. As for Link.sub.-- N2.sub.-- Limit, "X" represents the buffer state update frequency for the link, and for N2.sub.-- Limit[i], "Y" represents the buffer state update frequency for each connection.

In steady state operation, the UP 16 of the transmitter element 12 determines which virtual connection 20 (VC) has a non-zero cell count (i.e. has a cell ready to transmit), a BS.sub.-- Counter value less than the BS.sub.-- Limit, and an indication that the VC is next to send (also in FIGS. 3A and 3B).

The UP 16 increments BS.sub.-- Counter 22 and Tx.sub.-- Counter 26 whenever the UP 16 transmits a data cell over the respective connection 20, assuming flow control is enabled (FIG. 4). Upon receipt of the data cell, the DP 18 checks whether Buffer.sub.-- Counter 32 equals or exceeds Buffer.sub.-- Limit 30, which would be an indication that there are no buffers available for receipt of the data cell. If Buffer.sub.-- Counter>=Buffer.sub.-- Limit, the data cell is discarded (FIG. 3B). Otherwise, the DP 18 increments Buffer.sub.-- Counter 32 and Rx.sub.-- Counter 40 and the data cell is deposited in a buffer cell 28, as in FIG. 4. The Tx.sub.-- Counter 26 and the Rx.sub.-- Counter 40 roll over when they reach their maximum.

If flow control is not enabled, none of the presently described functionality is implemented. Connections that do not utilize flow control on the link can coexist with connections using link flow control. The flow control accounting is not employed when cells from non-flow controlled connections are transmitted and received. This includes both connection level accounting and link level accounting. Thereby, flow control and non-flow control connections can be active simultaneously.

When a data cell is forwarded out of the receiver element 14, Buffer.sub.-- Counter 32 is decremented. Buffer.sub.-- Counter 32 should never exceed Buffer.sub.-- Limit 30 when the connection-level flow control protocol is enabled, with the exception of when BS.sub.-- Limit 24 has been decreased and the receiver element 14 has yet to forward sufficient cells to bring Buffer.sub.-- Counter 32 below Buffer.sub.-- Limit 30.

A buffer state update occurs when the receiver element 14 has forwarded a number of data cells equal to N2.sub.-- Limit 34 out of the receiver element 14. In the first embodiment in which the DP 18 maintains Fwd.sub.-- Counter 38, update involves the transfer of the value of Fwd.sub.-- Counter 38 from the receiver element 14 back to the transmitter element 12 in an update cell, as in FIG. 6A. In the embodiment employing Rx.sub.-- Counter 40 in the downstream element 14, the value of Rx.sub.-- Counter 40 minus Buffer.sub.-- Counter 32 is conveyed in the update cell, as in FIG. 5A. At the transmitter 12, the update cell is used to update the value in BS.sub.-- Counter 22, as shown for the two embodiments in FIG. 7A. Since BS.sub.-- Counter 22 is independent of buffer allocation information, buffer allocation can be changed without impacting the performance of this aspect of connection-level flow control.

Update cells require an allocated bandwidth to ensure a bounded delay. This delay needs to be accounted for, as a component of round-trip time, to determine the buffer allocation for the respective connection.

The amount of bandwidth allocated to the update cells is a function of a counter, Max.sub.-- Update.sub.-- Counter (not illustrated) at an associated downstream transmitter element (not illustrated). This counter forces the scheduling of update and check cells, the latter to be discussed subsequently. There is a corresponding Min.sub.-- Update.sub.-- Interval counter (not shown) in the downstream transmitter element, which controls the space between update cells. Normal cell packing is seven records per cell, and Min.sub.-- Update.sub.-- Interval is similarly set to seven. Since the UP 16 can only process one update record per cell time, back-to-back, fully packed update cells received at the UP 16 would cause some records to be dropped.

An update event occurs as follows, with regard to FIGS. 1, 5A and 6A. When the downstream element 14 forwards (releases) a cell, Buffer.sub.-- Counter 32 is decremented and N2.sub.-- Counter 36 and Fwd.sub.-- Counter 38 are incremented. When the N2.sub.-- Counter 36 is equal to N2.sub.-- Limit 34, the DP 18 prepares an update cell for transmission back to the upstream element 12 and N2.sub.-- Counter 36 is set to zero. The upstream element 12 receives a connection indicator from the downstream element 14 forwarded cell to identify which connection 20 is to be updated. In the first embodiment, the DP 18 causes the Fwd.sub.-- Counter 38 value to be inserted into an update record payload (FIG. 6A). In the second embodiment, the DP 18 causes the Rx.sub.-- Counter 40 value minus the Buffer.sub.-- Counter 32 value to be inserted into the update record payload (FIG. 5A). When an update cell is fully packed with records, or as the minimum bandwidth pacing interval is reached, the update cell is transmitted to the upstream element 12.

Once received upstream, the UP 16 receives the connection indicator from the update record to identify the transmitter connection, and extracts the Fwd.sub.-- Counter 38 value or the Rx.sub.-- Counter 40 minus Buffer.sub.-- Counter 32 value from the update record. BS.sub.-- Counter 22 is reset to the value of Tx.sub.-- Counter 26 minus the update record value (FIG. 7A). If this connection was disabled from transmitting due to BS.sub.-- Counter 22 being equal to or greater than BS.sub.-- Limit 24, this condition should now be reversed, and if so the connection should again be enabled for transmitting.

In summary, the update event provides the transmitting element 12 with an indication of how many cells originally transmitted by it have now been released from buffers within the receiving element 14, and thus provides the transmitting element 12 with a more accurate indication of receiver element 14 buffer 28 availability for that connection 20.

The buffer state check event serves two purposes: 1) it provides a mechanism to calculate and compensate for cell loss or cell insertion due to transmission errors; and 2) it provides a mechanism to start (or restart) a flow if update cells were lost or if enough data cells were lost that N2.sub.-- Limit 34 is never reached.

One timer (not shown) in the UP subsystem 16 serves all connections. The connections are enabled or disabled on a per connection basis as to whether to send check cells from the upstream transmitter element 12 to the downstream receiver element 14. The check process in the transmitter element 12 involves searching all of the connection descriptors to find one which is check enabled (see FIGS. 8A, 9A). Once a minimum pacing interval has elapsed (the check interval), the check cell is forwarded to the receiver element 14 and the next check enabled connection is identified. The spacing between check cells for the same connection is a function of the number of active flow-controlled connections times the mandated spacing between check cells for all connections. Check cells have priority over update cells.

The check event occurs as follows, with regard to FIGS. 8A through 8C and 9A through 9C. Each transmit element 12 connection 20 is checked after a timed check interval is reached. If the connection is flow-control enabled and the connection is valid, then a check event is scheduled for transmission to the receiver element 14. A buffer state check cell is generated using the Tx.sub.-- Counter 26 value for that connection 20 in the check cell payload, and is transmitted using the connection indicator from the respective connection descriptor (FIGS. 8A and 9A).

In the first embodiment, a calculation of errored cells is made at the receiver element 14 by summing Fwd.sub.-- Counter 38 with Buffer.sub.-- Counter 32, and subtracting this value from the contents of the transmitted check cell record, the value of Tx.sub.-- Counter 26 (FIG. 9B). The value of Fwd.sub.-- Counter 38 is increased by the errored cell count. An update record with the new value for Fwd.sub.-- Counter 38 is then generated. This updated Fwd.sub.-- Counter 38 value subsequently updates the BS.sub.-- Counter 22 value in the transmitter element 12.

In the second embodiment, illustrated in FIG. 8B, the same is accomplished by resetting the Rx.sub.-- Counter 40 value equal to the check cell payload value (Tx.sub.-- Counter 26). A subsequent update record is established using the difference between the values of Rx.sub.-- Counter 40 and Buffer.sub.-- Counter 32.

Thus, the check event enables accounting for cells transmitted by the transmitter element 12, through the connection 20, but either dropped or not received by the receiver element 14.

A "no cell loss" guarantee is enabled using buffer state accounting at the connection level since the transmitter element 12 has an up-to-date account of the number of buffers 28 in the receiver element 14 available for receipt of data cells, and has an indication of when data cell transmission should be ceased due to the absence of available buffers 28 downstream.

In order to augment the foregoing protocol with a receiver element buffer sharing mechanism, link-level flow control, also known as link-level buffer state accounting, is added to connection-level flow control. It is possible for such link-level flow control to be implemented without connection-level flow control. However, a combination of the two is preferable since without connection-level flow control there would be no restriction on the number of buffers a single connection might consume.

It is desirable to perform buffer state accounting at the link level, in addition to the connection level, for the following reasons. Link-level flow control enables cell buffer sharing at a receiver element while maintaining the "no cell loss" guarantee afforded by connection-level flow control. Buffer sharing results in the most efficient use of a limited number of buffers. Rather than provide a number of buffers equal to bandwidth times RTT for each connection, a smaller number of buffers is employable in the receiver element 14 since not all connections require a full compliment of buffers at any one time.

A further benefit of link-level buffer state accounting is that each connection is provided with an accurate representation of downstream buffer availability without necessitating increased reverse bandwidth for each connection. A high-frequency link-level update does not significantly effect overall per-connection bandwidth.

Link-level flow control is described now with regard to FIG. 2. Like elements found in FIG. 1 are given the same reference numbers in FIG. 2, with the addition of a prime. Once again, only one virtual connection 20' is illustrated in the link 10', though the link 10' would normally host multiple virtual connections 20'. Once again, the link 10' is a physical link in a first embodiment, and a logical grouping of plural virtual connections in a second embodiment.

The upstream transmitter element 12' (FSPP subsystem) partially includes a processor labelled From Switch Port Processor (FSPP) 16'. The FSPP processor 16' is provided with two buffer state counters, BS.sub.-- Counter 22' and BS.sub.-- Limit 24', and a Tx.sub.-- Counter 26' each having the same function on a per-connection basis as those described with respect to FIG. 1.

The embodiment of FIG. 2 further includes a set of resources added to the upstream and downstream elements 12', 14' which enable link-level buffer accounting. These resources provide similar functions as those utilized on a per-connection basis, yet they operate on the link level.

For instance, Link.sub.-- BS.sub.-- Counter 50 tracks all cells in flight between the FSPP 16' and elements downstream of the receiver element 14', including cells in transit between the transmitter 12' and the receiver 14' and cells stored within receiver 14' buffers 28'. As with the update event described above with respect to connection-level buffer accounting, Link.sub.-- BS.sub.-- Counter 50 is modified during a link update event by subtracting either the Link.sub.-- Fwd.sub.-- Counter 68 value or the difference between Link.sub.-- Rx.sub.-- Counter 70 and Link.sub.-- Buffer.sub.-- Counter 62 from the Link.sub.-- TX.sub.-- Counter 54 value. In a first embodiment, the link-level counters are implemented in external RAM associated with the FSPP processor 16'.

Link.sub.-- BS.sub.-- Limit 52 limits the number of shared downstream cell buffers 28' in the receiver element 14' to be shared among all of the flow-control enabled connections 20'. In a first embodiment, Link.sub.-- BS.sub.-- Counter 50 and Link.sub.-- BS.sub.-- Limit 52 are both twenty bits wide.

Link.sub.-- TX.sub.-- Counter 54 tracks all cells transmitted onto the link 10'. It is used during the link-level update event to calculate a new value for Link.sub.-- BS.sub.-- Counter 50. Link.sub.-- TX.sub.-- Counter 54 is twenty-eight bits wide in the first embodiment.

In the downstream element 14', To Switch Port Processor (TSPP) 18' also manages a set of counters for each link 10' in the same fashion with respect to the commonly illustrated counters in FIGS. 1 and 2. The TSPP 18' further includes a Link.sub.-- Buffer.sub.-- Limit 60 which performs a function in the downstream element 14' similar to Link.sub.-- BS.sub.-- Limit 52 in the upstream element 12' by indicating the maximum number of cell buffers 28' in the receiver 14' available for use by all connections 10'. In most cases, Link.sub.-- BS.sub.-- Limit 52 is equal to Link.sub.-- Buffer.sub.-- Limit 60. The effect of adjusting the number of buffers 28' available up or down on a link-wide basis is the same as that described above with respect to adjusting the number of buffers 28 available for a particular connection 20. Link.sub.-- Buffer.sub.-- Limit 60 is twenty bits wide in the first embodiment.

Link.sub.-- Buffer.sub.-- Counter 62 provides an indication of the number of buffers in the downstream element 14' which are currently being used by all connections for the storage of data cells. This value is used in a check event to correct the Link.sub.-- Fwd.sub.-- Counter 68 (described subsequently). The Link.sub.-- Buffer.sub.-- Counter 62 is twenty bits wide in the first embodiment.

Link.sub.-- N2.sub.-- Limit 64 and Link.sub.-- N2.sub.-- Counter 66, each eight bits wide in the first embodiment, are used to generate link update records, which are intermixed with connection-level update records. Link.sub.-- N2.sub.-- Limit 64 establishes a threshold number for triggering the generation of a link-level update record (FIGS. 5B and 6B), and Link.sub.-- N2.sub.-- Counter 66 and Link.sub.-- Fwd.sub.-- Counter 68 are incremented each time a cell is released out of a buffer cell in the receiver element 14'. In a first embodiment, N2.sub.-- Limit 34' and Link.sub.-- N2.sub.-- Limit 64 are both static once initially configured.

However, in a further embodiment of the present invention, each is dynamically adjustable based upon measured bandwidth. For instance, if forward link bandwidth is relatively high, Link.sub.-- N2.sub.-- Limit 64 could be adjusted down to cause more frequent link-level update record transmission. Any forward bandwidth impact would be considered minimal. Lower forward bandwidth would enable the raising of Link.sub.-- N2.sub.-- Limit 64 since the unknown availability of buffers 28' in the downstream element 14' is less critical.

Link.sub.-- Fwd-Counter 68 tracks all cells released from buffer cells 28' in the receiver element 14' that came from the link 10' in question. It is twenty-eight bits wide in a first embodiment, and is used in the update event to recalculate Link.sub.-- BS.sub.-- Counter 50.

Link.sub.-- Rx.sub.-- Counter 70 is employed in an alternative embodiment in which Link.sub.-- Fwd.sub.-- Counter 68 is not employed. It is also twenty-eight bits wide in an illustrative embodiment and tracks the number of cells received across all connections 20' in the link 10'.

With regard to FIG. 2 et seq., a receiver element buffer sharing method is described. Normal data transfer by the FSPP 16' in the upstream element 12' to the TSPP 18' in the downstream element 14' is enabled across all connections 20' in the link 10' as long as the Link BS Counter 50 is less than or equal to Link.sub.-- BS.sub.-- Limit 52, as in FIG. 3B. This test prevents the FSPP 16' from transmitting more data cells than it believes are available in the downstream element 14'. The accuracy of this belief is maintained through the update and check events, described next.

A data cell is received at the downstream element 14' if neither connection-level or link-level buffer limit are exceeded (FIG. 3B). If a limit is exceeded, the cell is discarded.

The update event at the link level involves the generation of a link update record when the value in Link.sub.-- N2.sub.-- Counter 66 reaches (equals or exceeds) the value in Link.sub.-- N2.sub.-- Limit 64, as shown in FIGS. 5B and 6B. In a first embodiment, Link.sub.-- N2.sub.-- Limit 64 is set to forty.

The link update record, the value taken from Link.sub.-- Fwd.sub.-- Counter 68 in the embodiment of FIG. 6B, is mixed with the per-connection update records (the value of Fwd.sub.-- Counter 38') in update cells transferred to the FSPP 16'. In the embodiment of FIG. 5B, the value of Link.sub.-- Rx.sub.-- Counter 70 minus Link.sub.-- Buffer.sub.-- Counter 62 is mixed with the per-connection update records. When the upstream element 12' receives the update cell having the link update record, it sets the Link.sub.-- BS.sub.-- Counter 50 equal to the value of Link.sub.-- Tx.sub.-- Counter 54 minus the value in the update record (FIG. 7B). Thus, Link.sub.-- BS.sub.-- Counter 50 in the upstream element 12' is reset to reflect the number of data cells transmitted by the upstream element 12', but not yet released in the downstream element 14'.

The actual implementation of the transfer of an update record, in a first embodiment, recognizes that for each TSPP subsystem 14', there is an associated FSPP processor (not illustrated), and for each FSPP subsystem 12', there is also an associated TSPP processor (not illustrated). Thus, when an update record is ready to be transmitted by the TSPP subsystem 14' back to the upstream FSPP subsystem 12', the TSPP 18' conveys the update record to the associated FSPP (not illustrated), which constructs an update cell. The cell is conveyed from the associated FSPP to the TSPP (not illustrated) associated with the upstream FSPP subsystem 12'. The associated TSPP strips out the update record from the received update cell, and conveys the record to the upstream FSPP subsystem 12'.

The check event at the link level involves the transmission of a check cell having the Link.sub.-- Tx.sub.-- Counter 54 value by the FSPP 16' every "W" check cells (FIGS. 8A and 9A). In a first embodiment, W is equal to four. At the receiver element 14', the TSPP 18' performs the previously described check functions at the connection-level, as well as increasing the Link.sub.-- Fwd.sub.-- Counter 68 value by an amount equal to the check record contents, Link.sub.-- Tx.sub.-- Counter 54, minus the sum of Link.sub.-- Buffer.sub.-- Counter 62 plus Link.sub.-- Fwd.sub.-- Counter 68 in the embodiment of FIG. 9C. In the embodiment of FIG. 8C, Link.sub.-- Rx.sub.-- Counter 70 is modified to equal the contents of the check record (Link.sub.-- Tx.sub.-- Counter 54). This is an accounting for errored cells on a link-wide basis. An update record is then generated having a value taken from the updated Link.sub.-- Fwd.sub.-- Counter 68 or Link.sub.-- Rx.sub.-- Counter 70 values (FIGS. 8C and 9C).

It is necessary to perform the check event at the link level in addition to the connection level in order to readjust the Link.sub.-- Fwd.sub.-- Counter 68 value (FIG. 9C) or Link.sub.-- Rx.sub.-- Counter 70 value (FIG. 8C) quickly in the case of large transient link failures.

Again with regard to FIG. 2, the following are exemplary initial values for the illustrated counters in an embodiment having 100 connections in one link.

The BS.sub.-- Limit value equals the Buffer.sub.-- Limit value for both the connections and the link. Though BS.sub.-- Limit 24' and Buffer.sub.-- Limit 30' are both equal to twenty, and there are 100 connections in this link, there are only 1000 buffers 28' in the downstream element, as reflected by Link.sub.-- BS-Limit 52 and Link.sub.-- Buffer.sub.-- Limit 60. This is because of the buffer pool sharing enabled by link-level feedback.

Link-level flow control can be disabled, should the need arise, by not incrementing: Link.sub.-- BS.sub.-- Counter; Link.sub.-- N2.sub.-- Counter; and Link.sub.-- Buffer.sub.-- Counter, and by disabling link-level check cell transfer. No updates will occur under these conditions.

The presently described invention can be further augmented with a dynamic buffer allocation scheme, such as previously described with respect to N2.sub.-- Limit 34 and Link.sub.-- N2.sub.-- Limit 64. This scheme includes the ability to dynamically adjust limiting parameters such as BS.sub.-- Limit 24, Link.sub.-- BS.sub.-- Limit 52, Buffer.sub.-- Limit 30, and Link.sub.-- Buffer.sub.-- Limit 60, in addition to N2.sub.-- Limit 34 and Link.sub.-- N2.sub.-- Limit 64. Such adjustment is in response to measured characteristics of the individual connections or the entire link in one embodiment, and is established according to a determined priority scheme in another embodiment. Dynamic buffer allocation thus provides the ability to prioritize one or more connections or links given a limited buffer resource.

The Link.sub.-- N2.sub.-- Limit is set according to the desired accuracy of buffer accounting. On a link-wide basis, as the number of connections within the link increases, it may be desirable to decrease Link.sub.-- N2.sub.-- Limit in light of an increased number of connections in the link, since accurate buffer accounting allows greater buffer sharing among many connections. Conversely, if the number of connections within the link decreases, Link.sub.-- N2.sub.-- Limit may be increased, since the criticality of sharing limited resources among a relatively small number of connections is decreased.

In addition to adjusting the limits on a per-link basis, it may also be desirable to adjust limits on a per-connection basis in order to change the maximum sustained bandwidth for the connection.

The presently disclosed dynamic allocation schemes are implemented during link operation, based upon previously prescribed performance goals.

In a first embodiment of the present invention, incrementing logic for all counters is disposed within the FSPP processor 16'. Related thereto, the counters previously described as being reset to zero and counting up to a limit can be implemented in a further embodiment as starting at the limit and counting down to zero. The transmitter and receiver processors interpret the limits as starting points for the respective counters, and decrement upon detection of the appropriate event. For instance, if Buffer.sub.-- Counter (or Link.sub.-- Buffer.sub.-- Counter) is implemented as a decrementing counter, each time a data cell is allocated to a buffer within the receiver, the counter would decrement. When a data cell is released from the respective buffer, the counter would increment. In this manner, the counter reaching zero would serve as an-indication that all available buffers have been allocated. Such implementation is less easily employed in a dynamic bandwidth allocation scheme since dynamic adjustment of the limits must be accounted for in the non-zero counts.

A further enhancement of the foregoing zero cell loss, link-level flow control technique includes providing a plurality of shared cell buffers 28" in a downstream element 14" wherein the cell buffers 28" are divided into N prioritized cell buffer subsets, Priority 0 108a, Priority 1 108b, Priority 2 108c, and Priority 3 108d, by N-1 threshold level(s), Threshold(1) 102, Threshold(2) 104, and Threshold(3) 106. Such a cell buffer pool 28" is illustrated in FIG. 10, in which four priorities labelled Priority 0 through Priority 3 are illustrated as being defined by three thresholds labelled Threshold(1) through Threshold(3).

This prioritized buffer pool enables the transmission of high priority connections while lower priority connections are "starved" or prevented from transmitting cells downstream during periods of link congestion. Cell priorities are identified on a per-connection basis. The policy by which the thresholds are established is defined according to a predicted model of cell traffic in a first embodiment, or, in an alternative embodiment, is dynamically adjusted. Such dynamic adjustment may be in response to observed cell traffic at an upstream transmitting element, or according to empirical cell traffic data as observed at the prioritized buffer pool in the downstream element. For example, in an embodiment employing dynamic threshold adjustment, it may be advantageous to lower the number of buffers available to data cells having a priority less than Priority 0, or conversely to increase the number of buffers above Threshold(3), if a significantly larger quantity of Priority 0 traffic is detected.

The cell buffer pool 28" depicted in FIG. 10 is taken from the vantage point of a modified version 12" of the foregoing link-level flow control upstream element 12', the pool 28" being resident within a corresponding downstream element 14". This modified upstream element 12", viewed in FIG. 11, has at least one Link.sub.-- BS Threshold(n) 100, 102, 104 established in association with a Link BS Counter 50" and Link BS Limit 52", as described above, for characterizing a cell buffer pool 28" in a downstream element 14". These Link.sub.-- BS.sub.-- Thresholds 102, 104, 106 define a number of cell buffers in the pool 28" which are allocatable to cells of a given priority, wherein the priority is identified by a register 108 associated with the BS.sub.-- Counter 22" counter and BS-Limit 24" register for each connection 20". The Priorities 108a, 108b, 108c, 108d illustrated in FIG. 11 are identified as Priority 0 through Priority 3, Priority 0 being the highest. When there is no congestion, as reflected by Link.sub.-- BS.sub.-- Counter 50" being less than Link.sub.-- BS.sub.-- Threshold(1) 102 in FIGS. 10 and 11, flow-controlled connections of any priority can transmit. As congestion occurs, as indicated by an increasing value in the Link.sub.-- BS.sub.-- Counter 50", lower priority connections are denied access to downstream buffers, in effect disabling their transmission of cells. In the case of severe congestion, only cells of the highest priority are allowed to transmit. For instance, with respect again to FIG. 10, only cells of Priority 0 108a are enabled for transmission from the upstream element 12" to the downstream element 14" if the link-level Link.sub.-- BS.sub.-- Threshold(3) 106 has been reached downstream. Thus, higher priority connections are less effected by the state of the network because they have first access to the shared downstream buffer pool. Note, however, that connection-level flow control can still prevent a high-priority connection from transmitting, if the path that connection is intended for is severely congested. As above, Link.sub.-- BS.sub.-- Counter 50" is periodically updated based upon a value contained within a link-level update record transmitted from the downstream element 14" to the upstream element 12". This periodic updating is required in order to ensure accurate function of the prioritized buffer access of the present invention. In an embodiment of the present invention in which the Threshold levels 102, 104, 106 are modified dynamically, either as a result of tracking the priority associated with cells received at the upstream transmitter element or based upon observed buffer usage in the downstream receiver element, it is necessary for the FSPP 16" to have an accurate record of the state of the cell buffers 28", as afforded by the update function.

The multiple priority levels enable different categories of service, in terms of delay bounds, to be offered within a single quality of service. Within each quality of service, highest priority to shared buffers is typically given to connection/network management traffic, as identified by the cell header. Second highest priority is given to low bandwidth, small burst connections, and third highest for bursty traffic. With prioritization allocated as described, congestion within any one of the service categories will not prevent connection/management traffic from having the lowest cell delay.

Initialization of the upstream element 12" as depicted in FIG. 11 is illustrated in FIG. 12A. Essentially, the same counters and registers are set as viewed in FIG. 3A for an upstream element 12' not enabling prioritized access to a shared buffer resource, with the exception that Link.sub.-- BS.sub.-- Threshold 102, 104, 106 values are initialized to a respective buffer value T. As discussed, these threshold buffer values can be pre-established and static, or can be adjusted dynamically based upon empirical buffer usage data.

FIG. 12B represents many of the same tests employed prior to forwarding a cell from the upstream element 12" to the downstream element 14" as shown in FIG. 3B, with the exception that an additional test is added for the provision of prioritized access to a shared buffer resource. Specifically, the FSPP 16" uses the priority value 108 associated with a cell to be transferred to determine a threshold value 102, 104, 106 above which the cell cannot be transferred to the downstream element 14". Then, a test is made to determine whether the Link.sub.-- BS.sub.-- Counter 50" value is greater than or equal to the appropriate threshold value 102, 104, 106. If so, the data cell is not transmitted. Otherwise, the cell is transmitted and connection-level congestion tests are executed, as previously described.

In alternative embodiments, more or less than four priorities can be implemented with the appropriate number of thresholds, wherein the fewest number of priorities is two, and the corresponding fewest number of thresholds is one. For every N priorities, there are N-1 thresholds.

In yet a further embodiment, flow-control is provided solely at the link level, and not at the connection level, though it is still necessary for each connection to provide some form of priority indication akin to the priority field 108 illustrated in FIG. 11.

The link level flow controlled protocol as previously described can be further augmented in yet another embodiment to enable a guaranteed minimum cell rate on a per-connection basis with zero cell loss. This minimum cell rate is also referred to as guaranteed bandwidth. The connection can be flow-controlled below this minimum, allocated rate, but only by the receiver elements associated with this connection. Therefore, the minimum rate of one connection is not affected by congestion within other connections.

It is a requirement of the presently disclosed mechanism that cells present at the upstream element associated with the FSPP 116 be identified by whether they are to be transmitted from the upstream element using allocated bandwidth, or whether they are to be transmitted using dynamic bandwidth. For instance, the cells may be provided in queues associated with a list labelled "preferred," indicative of cells requiring allocated bandwidth. Similarly, the cells may be provided in queues associated with a list labelled "dynamic," indicative of cells requiring dynamic bandwidth.

In a frame relay setting, the present mechanism is used to monitor and limit both dynamic and allocated bandwidth. In a setting involving purely internet traffic, only the dynamic portions of the mechanism may be of significance. In a setting involving purely CBR flow, only the allocated portions of the mechanism would be employed. Thus, the presently disclosed method and apparatus enables the maximized use of mixed scheduling connections--those requiring all allocated bandwidth to those requiring all dynamic bandwidth, and connections therebetween.

In the present mechanism, a downstream cell buffer pool 128, akin to the pool 28' of FIG. 2, is logically divided between an allocated portion 300 and a dynamic portion 301, whereby cells identified as to receive allocated bandwidth are buffered within this allocated portion 300, and cells identified as to receive dynamic bandwidth are buffered in the dynamic portion 301. FIG. 13A shows the two portions 300, 301 as distinct entities; the allocated portion is not a physically distinct block of memory, but represents a number of individual cell buffers, located anywhere in the pool 128.

In a further embodiment, the presently disclosed mechanism for guaranteeing minimum bandwidth is applicable to a mechanism providing prioritized access to downstream buffers, as previously described in conjunction with FIGS. 10 and 11. With regard to FIG. 13B, a downstream buffer pool 228 is logically divided among an allocated portion 302 and a dynamic portion 208, the latter logically subdivided by threshold levels 202, 204, 206 into prioritized cell buffer subsets 208a-d. As with FIG. 13A, the division of the buffer pool 228 is a logical, not physical, division.

Elements required to implement this guaranteed minimum bandwidth mechanism are illustrated in FIG. 14, where like elements from FIGS. 2 and 11 are provided with like reference numbers, added to 100 or 200. Note that no new elements have been added to the downstream element; the presently described guaranteed minimum bandwidth mechanism is transparent to the downstream element.

New aspects of flow control are found at both the connection and link levels. With respect first to the connection level additions and modifications, D.sub.-- BS.sub.-- Counter 122 highlights resource consumption by tracking the number of cells scheduled using dynamic bandwidth transmitted downstream to the receiver 114. This counter has essentially the same function as BS.sub.-- Counter 22' found in FIG. 2, where there was no differentiation between allocated and dynamically scheduled cell traffic. Similarly, D.sub.-- BS.sub.-- Limit 124, used to provide a ceiling on the number of downstream buffers available to store cells from the transmitter 112, finds a corresponding function in BS-Limit 24' of FIG. 2. As discussed previously with respect to link level flow control, the dynamic bandwidth can be statistically shared; the actual number of buffers available for dynamic cell traffic can be over-allocated. The amount of "D" buffers provided to a connection is equal to the RTT times the dynamic bandwidth plus N2. RTT includes delays incurred in processing the update cell.

A.sub.-- BS.sub.-- Counter 222 and A.sub.-- BS.sub.-- Limit 224 also track and limit, respectively, the number of cells a connection can transmit by comparing a transmitted number with a limit on buffers available. However, these values apply strictly to allocated cells; allocated cells are those identified as requiring allocated bandwidth (the guaranteed minimum bandwidth) for transmission. Limit information is set up at connection initialization time and can be raised and lowered as the guaranteed minimum bandwidth is changed. If a connection does not have an allocated component, the A.sub.-- BS.sub.-- Limit 224 will be zero. The A.sub.-- BS.sub.-- Counter 222 and A.sub.-- BS.sub.-- Limit 224 are in addition to the D.sub.-- BS.sub.-- Counter 122 and D.sub.-- BS.sub.-- Limit 124 described above. The amount of "A" buffers dedicated to a connection is equal to the RTT times the allocated bandwidth plus N2. The actual number of buffers dedicated to allocated traffic cannot be over-allocated. This ensures that congestion on other connections does not impact the guaranteed minimum bandwidth.

A connection loses, or runs out of, its allocated bandwidth through the associated upstream switch once it has enqueued a cell but has no more "A" buffers as reflected by A.sub.-- BS.sub.-- Counter 222 and A.sub.-- BS.sub.-- Limit 224. If a connection is flow controlled below its allocated rate, it loses a portion of its allocated bandwidth in the switch until the congestion condition is alleviated. Such may be the case in multipoint-to-point (M2P) switching, where plural sources on the same connection, all having a minimum guaranteed rate, converge on a single egress point which is less than the sum of the source rates. In an embodiment of the presently disclosed mechanism in which the transmitter element is a portion of a switch having complimentary switch flow control, the condition of not having further "A" buffer states inhibits the intra-switch transmission of further allocated cell traffic for that connection.

The per-connection buffer return policy is to return buffers to the allocated pool first, until the A.sub.-- BS.sub.-- Counter 222 equals zero. Then buffers are returned to the dynamic pool, decreasing D.sub.-- BS.sub.-- Counter 122.

Tx.sub.-- Counter 126 and Priority 208 are provided as described above with respect to connection-level flow control and prioritized access.

On the link level, the following elements are added to enable guaranteed minimum cell rate on a per-connection basis. Link.sub.-- A.sub.-- BS.sub.-- Counter 250 is added to the FSPP 116. It tracks all cells identified as requiring allocated bandwidth that are "in-flight" between the FSPP 116 and the downstream switch fabric, including cells in the TSPP 118 cell buffers 128, 228. The counter 250 is decreased by the same amount as the A.sub.-- BS.sub.-- Counter 222 for each connection when a connection level update function occurs (discussed subsequently).

Link.sub.-- BS.sub.-- Limit 152 reflects the total number of buffers available to dynamic cells only, and does not include allocated buffers. Link.sub.-- BS.sub.-- Counter 150, however, reflects a total number of allocated and dynamic cells transmitted. Thus, connections are not able to use their dynamic bandwidth when Link.sub.-- BS.sub.-- Counter 150 (all cells in-flight, buffered, or in downstream switch fabric) minus Link.sub.-- A.sub.-- BS.sub.-- Counter 250 (all allocated cells transmitted) is greater than Link.sub.-- BS.sub.-- Limit 152 (the maximum number of dynamic buffers available). This is necessary to ensure that congestion does not impact the allocated bandwidth. The sum of all individual A.sub.-- BS.sub.-- Limit 224 values, or the total per-connection allocated cell buffer space 300, 302, is in one embodiment less than the actually dedicated allocated cell buffer space in order to account for the potential effect of stale (i.e., low frequency) connection-level updates.

Update and check events are also implemented in the presently disclosed allocated/dynamic flow control mechanism. The downstream element 114 transmits connection level update cells when either a preferred list and a VBR-priority 0 list are empty and an update queue is fully packed, or when a "max.sub.-- update.sub.-- interval" (not illustrated) has been reached.

At the upstream end 112, the update cell is analyzed to identify the appropriate queue, the FSPP 116 adjusts the A.sub.-- BS.sub.-- Counter 222 and D.sub.-- BS.sub.-- Counter 122 for that queue, returning cell buffers to "A" first then "D", as described above, since the FSPP 116 cannot distinguish between allocated and dynamic buffers. The number of "A" buffers returned to individual connections is subtracted from Link.sub.-- A.sub.-- BS Counter 250.

Other link level elements used in association with the presently disclosed minimum guaranteed bandwidth mechanism, such as Link.sub.-- Tx.sub.-- Counter 154, function as described in the foregoing discussion of link level flow control. Also, as previously noted, a further embodiment of the presently described mechanism functions with a link level flow control scenario incorporating prioritized access to the downstream buffer resource 228 through the use of thresholds 202, 204, 206. The function of these elements are as described in the foregoing.

The following is an example of a typical initialization in a flow controlled link according to the present disclosure:

Downstream element has 3000 buffers;

Link is short haul, so RTT*bandwidth equals one cell;

100 allocated connections requiring 7 "A" buffers each, consuming 700 buffers total;

3000-700=2300 "D" buffers to be shared among 512 connections having zero allocated bandwidth;

Link.sub.-- BS.sub.-- Limit=2300.

If D.sub.-- BS.sub.-- Counter>=D.sub.-- BS.sub.-- Limit, then the queue is prevented from indicating that it has a cell ready to transmit. In the embodiment referred to above in which the upstream element is a switch having composite bandwidth, this occurs by the queue being removed from the dynamic list, preventing the queue from being scheduled for transmit using dynamic bandwidth.

For allocated cells, a check is made when each cell is enqueued to determine whether the cell, plus other enqueued cells, plus A.sub.-- BS.sub.-- Counter, is a number greater than A.sub.-- BS.sub.-- Limit. If not, the cell is enqueued and the queue is placed on the preferred list. Else, the connection is prevented from transmitting further cells through the upstream element 112 switch fabric.

Initialization of the upstream element 112 as depicted in FIG. 14 is illustrated in FIG. 15A. Essentially, the same counters and registers set in FIG. 3A for an upstream element 12' (when prioritized access to a shared buffer resource is not enabled), and in FIG. 12A for an upstream element 12" (when prioritized access is enabled). Exceptions include: Link.sub.-- A.sub.-- BS.sub.-- Counter 250 initialized to zero; connection-level allocated and dynamic BS.sub.-- Counters 122, 222 set to zero; and connection-level allocated and dynamic BS.sub.-- Limits 124, 224 set to respective values of NA and NB. Similarly, on the downstream end at the connection level, the allocated and dynamic Buffer.sub.-- Limits and Buffer.sub.-- counters are set, with the Buffer.sub.-- Limits employing a bandwidth value for the respective traffic type (i.e., BW.sub.A =allocated cell bandwidth and BW.sub.D =dynamic cell bandwidth). Further, each cell to be transmitted is identified as either requiring allocated or dynamic bandwidth as the cell is received from the switch fabric.

FIG. 15B represents many of the same tests employed prior to forwarding a cell from the upstream element 112 to the downstream element 114 as shown in FIGS. 3B and 12B, with the following exceptions. Over-allocation of buffer states per connection is checked for dynamic traffic only and is calculated by subtracting Link.sub.-- A.sub.-- BS.sub.-- Counter from Link BS Counter and comparing the result to Link.sub.-- BS.sub.-- Limit. Over-allocation on a link-wide basis is calculated from a summation of Link.sub.-- BS.sub.-- Counter (which tracks both allocated and dynamic cell traffic) and Link.sub.-- A.sub.-- BS.sub.-- Counter against the Link.sub.-- BS.sub.-- Limit. Similarly, over-allocation at the downstream element is tested for both allocated and dynamic traffic at the connection level. As previously indicated, the presently disclosed mechanism for providing guaranteed minimum bandwidth can be utilized with or without the prioritized access mechanism, though aspects of the latter are illustrated in FIGS. 15A and 15B for completeness.

As discussed, connection-level flow control as known in the art relies upon discrete control of each individual connection. In particular, between network elements such as a transmitting element and a receiving element, the control is from transmitter queue to receiver queue. Thus, even in the situation illustrated in FIG. 16 in which a single queue Q.sub.A in a transmitter element is the source of data cells for four queues Q.sub.W, Q.sub.X, Q.sub.Y, and Q.sub.Z associated with a single receiver processor, the prior art does not define any mechanism to handle this situation.

In FIG. 16, the transmitter element 10 is an FSPP element having a FSPP 11 associated therewith, and the receiver element 12 is a TSPP element having a TSPP 13 associated therewith. The FSPP 11 and TSPP 13 as employed in FIG. 16 selectively provide the same programmable capabilities as described above, such as link-level flow control, prioritized access to a shared, downstream buffer resource, and guaranteed minimum cell rate on a connection level, in addition to a connection-level flow control mechanism. Whether one or more of these enhanced capabilities are employed in conjunction with the connection-level flow control is at the option of the system configurator.

Yet another capability provided by the FSPP and TSPP according to the present disclosure is the ability to treat a group of receiver queues jointly for purposes of connection-level flow control. In FIG. 16, instead of utilizing four parallel connections, the presently disclosed mechanism utilizes one connection 16 in a link 14, terminating in four separate queues Q.sub.W, Q.sub.X, Q.sub.Y, and Q.sub.Z, though the four queues are treated essentially as a single, joint entity for purposes of connection-level flow control. This is needed because some network elements need to use a flow controlled service but cannot handle the bandwidth of processing update cells when N2 is set to a low value, 10 or less (see above for a discussion of the update event in connection-level flow control). Setting N2 to a large value, such as 30, for a large number of connections requires large amounts of downstream buffering because of buffer orphaning, where buffers are not in-use but are accounted for up-stream as in-use because of the lower frequency of update events. This mechanism is also useful to terminate Virtual Channel Connections (VCC) within a Virtual Path Connection (VPC), where flow control is applied to the VPC.

This ability to group receiver queues is a result of manipulations of the queue descriptor associated with each of the receiver queues Q.sub.W, Q.sub.X, Q.sub.Y, and Q.sub.Z. With reference to FIG. 17, queue descriptors for the queues in the receiver are illustrated. Specifically, the descriptors for queues Q.sub.W, Q.sub.X, and Q.sub.Y are provided on the left, and in general have the same characteristics. One of the first fields pertinent to the present disclosure is a bit labelled "J." When set, this bit indicates that the associated queue is being treated as part of a joint connection in a receiver. Instead of maintaining all connection-level flow control information in each queue descriptor for each queue in the group, certain flow control elements are maintained only in one of the queue descriptors for the group. In the illustrated case, that one queue is queue Q.sub.Z.

In each of the descriptors for queues Q.sub.W, Q.sub.X, and Q.sub.Y, a "Joint Number" field provides an offset or pointer to a set of flow control elements in the descriptor for queue Q.sub.Z. This pointer field may provide another function when the "J" bit is not set. While Buffer.sub.-- Limit (labelled "Buff-Limit" in FIG. 17) and N2.sub.-- Limit are maintained locally within each respective descriptor, Joint.sub.-- Buffer.sub.-- Counter (labelled "Jt.sub.-- Buff.sub.-- Cntr"), Joint.sub.-- N2.sub.-- Counter (labelled "Jt.sub.-- N2.sub.-- Cntr"), and Joint.sub.-- Forward.sub.-- Counter (labelled "Jt.sub.-- Fwd.sub.-- Cntr") are maintained in the descriptor for queue Q.sub.Z for all of the queues in the group. The same counters in the descriptors for queues Q.sub.W, Q.sub.X, and Q.sub.Y go unused. The joint counters perform the same function as the individual counters, such as those illustrated in FIG. 2 at the connection level, but are advanced or decremented as appropriate by actions taken in association with the individual queues. Thus, for example, Joint.sub.-- Buffer.sub.-- Counter is updated whenever a buffer cell receives a data cell or releases a data cell in association with any of the group queues. The same applies to Joint.sub.-- N2.sub.-- Counter and Joint.sub.-- Forward.sub.-- Counter. In an alternate embodiment of the previously described flow control mechanism, each Forward.sub.-- Counter is replaced with Receive.sub.-- Counter. Similarly, in an alternative embodiment of the presently disclosed mechanism, Joint.sub.-- Forward.sub.-- Counter is replaced with Joint.sub.-- Receive.sub.-- Counter, depending upon which is maintained in each of the group queues. Only the embodiment including Forward.sub.-- Counter and Joint.sub.-- Forward.sub.-- Counter are illustrated.

Not all of the per-queue descriptor elements are superseded by functions in a common descriptor. Buffer.sub.-- Limit (labelled "Buff Limit" in FIG. 17) is set and referred to on a per-queue basis. Thus, Joint.sub.-- Buffer.sub.-- Counter is compared against the Buffer.sub.-- Limit of a respective queue. Optionally, the Buffer.sub.-- Limit could be Joint.sub.-- Buffer.sub.-- Limit, instead of maintaining individual, common limits. The policy is to set the same Buffer.sub.-- Limit in all the TSPP queues associated with a single Joint.sub.-- Buffer.sub.-- Counter.

An update event is triggered, as previously described, when the Joint.sub.-- N2.sub.-- Counter reaches the queue-level N2.sub.-- Limit. The policy is to set all of the N2.sub.-- Limits equal to the same value for all the queues associated with a single joint flow control connection.

When a check cell is received for a connection, an effort to modify the Receive.sub.-- Counter associated with the receiving queue results in a modification of the Joint.sub.-- Receive.sub.-- Counter. Thus, the level of indirection provided by the Joint.sub.-- Number is applicable to both data cells and check cells.

At the transmitter element 10, only one set of upstream flow control elements are maintained. At connection set-up time, the joint connection is set-up as a single, point-to-point connection, as far as the upstream elements are concerned. Therefore, instead of maintaining four sets of upstream elements for the embodiment of FIG. 16, the presently disclosed mechanism only requires one set of elements (Tx.sub.-- Counter, BS.sub.-- Counter, BS.sub.-- Limit, all having the functionality as previously described).

Once a joint flow control entity has been established, other TSPP queues for additional connections may be added. To do so, each new queue must have the same N2.sub.-- Limit and Buffer.sub.-- Limit values. The queues for the additional connections will reference the common Joint.sub.-- N2.sub.-- Counter and either Joint Forward Counter or Joint.sub.-- Receive.sub.-- Counter.

As previously noted, when J=1, the Joint.sub.-- Number field is used as an offset to the group descriptor. The Joint.sub.-- Number for the group descriptor is set to itself, as shown in FIG. 17 with regard to the descriptor for queue Q.sub.Z. This is also the case in point-to-point connections (VCC to VCC rather than the VPC to VCC, as illustrated in FIG. 16), where each Joint.sub.-- Number points to its own descriptor. Implementation for each of point-to-point and the presently described point-to-multipoint connections is thus simplified.

Having described preferred embodiments of the invention, it will be apparent to those skilled in the art that other embodiments incorporating the concepts may be used. These and other examples of the invention illustrated above are intended by way of example and the actual scope of the invention is to be determined from the following claims.