EP1027789A1 - Method and system for dynamic allocation of bandwidth in asynchronous transfer mode (atm) switching systems - Google Patents
Method and system for dynamic allocation of bandwidth in asynchronous transfer mode (atm) switching systemsInfo
- Publication number
- EP1027789A1 EP1027789A1 EP98944557A EP98944557A EP1027789A1 EP 1027789 A1 EP1027789 A1 EP 1027789A1 EP 98944557 A EP98944557 A EP 98944557A EP 98944557 A EP98944557 A EP 98944557A EP 1027789 A1 EP1027789 A1 EP 1027789A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- bottlenecked
- switching system
- virtual circuits
- bandwidth
- output port
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/54—Store-and-forward switching systems
- H04L12/56—Packet switching systems
- H04L12/5601—Transfer mode dependent, e.g. ATM
- H04L12/5602—Bandwidth control in ATM Networks, e.g. leaky bucket
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/04—Selecting arrangements for multiplex systems for time-division multiplexing
- H04Q11/0428—Integrated services digital network, i.e. systems for transmission of different types of digitised signals, e.g. speech, data, telecentral, television signals
- H04Q11/0478—Provisions for broadband connections
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/54—Store-and-forward switching systems
- H04L12/56—Packet switching systems
- H04L12/5601—Transfer mode dependent, e.g. ATM
- H04L2012/5629—Admission control
- H04L2012/5631—Resource management and allocation
- H04L2012/5632—Bandwidth allocation
- H04L2012/5635—Backpressure, e.g. for ABR
Definitions
- the present invention relates generally to asynchronous transfer mode (ATM) switching systems, and more particularly, to a method and system for dynamically allocating bandwidth to Available Bit Rate (ABR) virtual circuits in ATM switching systems.
- ATM asynchronous transfer mode
- ABR Available Bit Rate
- a virtual source transmits data in the form of fixed sized cells to a virtual destination (VD) through a connection (referred to as virtual circuit) established between the virtual source and the virtual destination.
- the virtual source and virtual destination may be a telephone, video equipment, facsimile, computer, edge-router, edge-switch, etc.
- the cells may include any type of digitized information, including audio, computer data, video, multimedia, Internet data, etc.
- TCP/IP Transmission Control Protocol/Internet Protocol
- a virtual source may be an edge-router at the entry to an ATM network.
- An entry edge-router segments the incoming TCP/IP data packets into one or more ATM cells before transmitting each cell to the ATM network.
- a virtual destination may be an edge router at the exit of the ATM network.
- An exit edge-router reassembles incoming ATM cells into TCP/IP data packets before transmitting each packet to its destination.
- a virtual source When establishing a virtual circuit through an ATM network, a virtual source can select one of five different categories of service: Constant Bit rate (CBR), Variable Bit Rate - Real Time (VBR-RT), Variable Bit Rate - Non Real Time (VBR- NRT), Available Bit Rate (ABR), and Unspecified Bit Rate (UBR).
- CBR Constant Bit rate
- VBR-RT Variable Bit Rate - Real Time
- VBR- NRT Variable Bit Rate - Non Real Time
- ABR Available Bit Rate
- UBR Unspecified Bit Rate
- ATM Forum Traffic Management Standard af-tm-0056.00 describes each of these services.
- the ABR service determines excess bandwidth in the network and uses network management methods to reallocate the excess bandwidth among the virtual circuits in the network to reduce network congestion and cell loss.
- PCR peak cell rate
- MCR minimum cell rate
- the ABR service uses the negotiated PCR and MCR parameters to provide a guaranteed quality of service concerning bandwidth availability and cell loss in a virtual circuit.
- a virtual source When a virtual source selects the ABR service, the virtual source periodically generates a resource management (RM) cell to get feedback from the network on the rate at which the virtual source can transmit cells on a virtual circuit without causing loss of cells due to network congestion.
- a virtual source generates an RM cell for every thirty-one cells it transmits or at the expiration of a fixed time interval, whichever occurs first.
- the network processes the RM cell, updates virtual circuit bandwidth information in the RM cell, and returns the RM cell to the virtual source.
- the virtual source then dynamically adjusts its rate of cell transmission based on the bandwidth information contained in the RM cell.
- An RM cell generated by a virtual source is referred to as a Forward RM cell.
- the Forward RM cell passes through one or more switching systems in the network before reaching a virtual destination.
- the virtual destination processes the Forward RM cell and returns a Backward RM cell to the virtual source.
- the Backward RM cell passes through one or more switching systems in the network before reaching the virtual source.
- a virtual source maintains the MCR, current Allowed Cell Rate (ACR), and the PCR associated with the virtual circuit.
- ACR is the rate at which the network allows the virtual source to transmit cells on a virtual circuit.
- a Forward RM cell includes an MCR field, current cell rate (CCR) field, and an explicit rate (ER) field.
- CCR is the rate at which a virtual source is transmitting cells on a virtual circuit at the time the virtual source generates a Forward RM cell.
- ER is the rate at which the virtual source wishes to transmit cells on a virtual circuit.
- a virtual source cannot set the ER field in a Forward RM cell to be greater than PCR. After generating a Forward RM cell and setting the MCR, CCR, and ER fields in the Forward RM cell, the virtual source transmits the Forward RM cell to the network.
- ABR-ER ABR Explicit Rate
- the switching system determines an upper threshold (referred to as "cutoff") for the bandwidth that can be made available to the virtual circuit in the switching system. If the switching system determines that the computed cutoff for the virtual circuit sets the ACR in the virtual source (i.e, the switching system computes the smallest cutoff among all of the switching systems on the path of the virtual circuit), then the switching system considers the virtual circuit to be "bottlenecked here" or bottlenecked in the switching system.
- the switching system determines that the computed cutoff does not set the ACR in the virtual source (i.e, the switching system does not compute the smallest cutoff among all of the switching systems on the path of the virtual circuit), then the switching system identifies the virtual circuit as "bottlenecked elsewhere.”
- the switching system determines a new bandwidth that it can allocate to the virtual circuit by determining a new ER for the virtual circuit and setting the new ER in the Forward RM cell.
- the switching system determines an estimated rate (Exp_Rate) at which the switching system "expects" the virtual source to transmit data cells after the virtual source adjusts its ACR based on the newly set ER.
- the switching system sends the Forward RM cell to the next switching system on the path of the virtual circuit.
- the virtual destination When the Forward RM cell reaches the virtual destination, the virtual destination returns the Forward RM cell as a Backward RM cell.
- the Backward RM cell passes through one or more switching systems on the path of the virtual circuit without any further modification to the bandwidth information set in the Forward RM cell.
- the virtual source uses the new ER in the RM cell to determine a new ACR. Based on the new ACR, the virtual source adjusts the rate at which it transmits cells.
- the switching system Every time a switching system determines a cutoff and ER for a virtual circuit, the switching system also recomputes certain global bandwidth parameters for the virtual circuits that the switching system handles. These global bandwidth parameters include the total bandwidth available to all ABR-ER virtual circuits, the total Exp_Rate for ABR-ER virtual circuits that are bottlenecked elsewhere, the total number of ABR-ER virtual circuits that are bottlenecked elsewhere, and the total number of ABR-ER virtual circuits that are bottlenecked at the switching system.
- a switching system allocates an optimum bandwidth to a virtual circuit when the virtual source uses the entire or nearly the entire bandwidth allocated to the virtual circuit. If the global bandwidth parameters are inaccurate, the switching system may allocate an insufficient amount of bandwidth to a virtual circuit, even though the switching system may in fact have sufficient bandwidth available. Similarly, the switching system may allocate an excessive amount of bandwidth to a virtual circuit, even though the switching system may in fact have insufficient bandwidth available. In either case, the switching system would fail to allocate an optimum bandwidth to the virtual circuit.
- the methods known prior to the present invention for determining and updating the global bandwidth parameters in an ABR-ER switching system have the additional disadvantage that, when a virtual circuit does not use the entire bandwidth a switching system allocates to it, the switching system cannot dynamically identify and reallocate the unused bandwidth to other virtual circuits in the switching system. For example, when a virtual source stops transmitting cells on a virtual circuit or transmits cells at a much lower rate than the bandwidth allocated to the virtual circuit permits, the virtual source does not use the bandwidth allocated to the virtual circuit in an optimum fashion.
- the methods known prior to the present invention for allocating bandwidth to an ABR virtual circuit have the disadvantage that a switching system may allocate more bandwidth to the virtual circuit than a virtual source could use. For example, when a virtual source requests an ER that is below the MCR for a virtual circuit, the switching system cannot decrease the ER below the MCR.
- the present invention comprises a method and system for dynamically adjusting the total bandwidth that an ABR-ER switching system can make available to ABR virtual circuits in the switching system by recomputing, at fixed time intervals, the total bandwidth that the switching system can make available to active ABR virtual circuits that are bottlenecked in the switching system.
- the method and system classifies the ABR virtual circuits that are established through an output port of the switching system into three categories of "recently active,” “active,” and "inactive" virtual circuits.
- An ABR virtual circuit is "recently active” if an output port receives at least one RM cell for the virtual circuit since the most recent recomputation of the total bandwidth that the switching system can make available to active ABR virtual circuits that are bottlenecked in the switching system (hereinafter referred to as "bandwidth allocation update").
- An ABR virtual circuit is "active” if an output port receives at least one RM cell within the two most recent bandwidth allocation updates.
- An ABR virtual circuit is "inactive” if an output port does not receive at least one RM cell within the two most recent bandwidth allocation updates.
- the present invention further comprises a method and system for dynamically adjusting the total bandwidth that an ABR-ER switching system can make available to ABR virtual circuits in the switching system by determining a set of weighted "cutoffs" for the ABR virtual circuits, respectively, determining a first total expected rate (Exp_Rate) for recently active virtual circuits that are bottlenecked elsewhere based on the weighted cutoffs, setting a second total expected rate for active virtual circuits that are bottlenecked elsewhere to the first total expected rate, determining a total bandwidth available in the switching system for active virtual circuits that are bottlenecked in the switching system based on the first total expected rate, and setting to zero the first total expected rate.
- Exp_Rate first total expected rate
- a switching system consistent with the present invention determines the total bandwidth allocated to inactive ABR virtual circuits in the switching system without explicitly identifying the inactive ABR virtual circuits, and makes available the bandwidth allocated to inactive ABR virtual circuits to active ABR virtual circuits that are established through the switching system. When transmission of cells resumes in an inactive ABR virtual circuit, the switching system reallocates new bandwidth to the virtual circuit.
- Figure la is a block diagram of an ATM communications network.
- Figure lb is a block diagram of an end-to-end path of a virtual circuit that passes through N switching systems in an ATM communications network.
- FIG. 2 is a block diagram of a switching system in accordance with an embodiment of the invention.
- FIG. 3 is a block diagram of an output port in a switching system in accordance with an embodiment of the invention.
- Figure 4 is a flow chart of the steps a switching system performs to process a resource management cell in accordance with an embodiment of the invention.
- Figure 5 is a flow chart of the steps a switching system performs to process a resource management cell in accordance with an embodiment of the invention.
- Figure 6 is a flow chart of the steps a switching system performs to determine bandwidth parameters for an output port and a virtual circuit in accordance with an embodiment of the invention.
- Figure 7 is a flow chart of the steps a switching system performs to determine a new explicit rate for a virtual circuit in accordance with an embodiment of the invention.
- Figure 8 is a flow chart of the steps a switching system performs to determine the total expected rate for active Available Bit Rate virtual circuits that are bottlenecked elsewhere in accordance with an embodiment of the invention.
- Figure 9 is a flow chart of the steps a switching system performs to determine the total expected rate for recently active Available Bit Rate virtual circuits that are bottlenecked elsewhere in accordance with an embodiment of the invention.
- Figure 10 is a flow chart of the steps a switching system performs to determine the sum of weight factors for active Available Bit Rate virtual circuits that are bottlenecked at an output port of the switching system in accordance with an embodiment of the invention.
- Figure 11 is a flow chart of the steps a switching system performs to determine the sum of weight factors for recently active Available Bit Rate virtual circuits that are bottlenecked at an output port of the switching system in accordance with an embodiment of the invention.
- Figure 12 is a flow chart of the steps a switching system performs to determine the total bandwidth available to active Available Bit Rate virtual circuits that are bottlenecked at an output port of the switching system in accordance with an embodiment of the invention.
- Figure 13 is a flow chart of the steps a switching system performs to make available a new total bandwidth to active Available Bit Rate virtual circuits that are bottlenecked at an output port of the switching system in accordance with an embodiment of the invention.
- FIG. la illustrates a block diagram of an ATM communications network in accordance with an embodiment of the invention.
- Virtual source (VS) 100 interfaces with ATM network 160 via transmission link 120.
- Virtual destination (VD) 110 interfaces with ATM network 160 via transmission link 125.
- ATM network 160 sets up a virtual circuit (VC) (not shown) between VS 100 and VD 110.
- VC virtual circuit
- VS 100 negotiates with ATM network 160 for an Available Bit Rate- Explicit Rate (ABR-ER) service.
- ABR-ER Available Bit Rate- Explicit Rate
- W is also assigned to the VC.
- the weight factor may depend on, for example, the tariff payed by the owner of the VC, the service priority of the VC, and/or the amount of traffic the VC is expected to carry.
- VS 100 After completion of the set-up phase of the connection, VS 100 generates Forward resource management (RM) cell 130 and transmits it on transmission link 120.
- Forward RM cell 130 passes through ATM network 160 and transmission link 125, and arrives at VD 110.
- VD 110 processes Forward RM cell 130, and generates Backward RM cell 140 in response to Forward RM cell 130.
- VD 110 then transmits Backward RM cell 140 on transmission link 125.
- Backward RM cell 140 passes through ATM network 160 and transmission link 120, and arrives at VS 100.
- RM Forward resource management
- VS 100 generates data cell 150 and transmits it on transmission link 120.
- Data cell 150 passes through ATM network 160 and transmission link 125, and arrives at VD 110.
- VD 110 generates data cell 150 and transmits it on transmission link
- Data cell 150 passes through ATM network 160 and transmission link 120, and arrives at VS 100.
- Figure lb illustrates a block diagram of the end-to-end path of a VC in ATM network 160 in accordance with an embodiment of the invention.
- the path of the VC established between VS 100 and VD 110 passes through N switching systems 105, through 105 N .
- Each switching system 105, through 105 N connects to its adjacent switching system via a corresponding transmission link 121, 122, and 123.
- SWITCHING SYSTEM Figure 2 illustrates a block diagram of a switching system, for example switching system 105 2 , in accordance with an embodiment of the invention.
- switching system 105 2 comprises M input ports 240, through 240 M .
- Each input port interfaces with an input line, which can carry K virtual circuits (not shown), where K is an integer.
- input port 240 M interfaces with input line 121 M , which carries K virtual circuits.
- Input ports 240, through 240 M preferably interface with a cross-point switch fabric 250 via input links 220, through 220 M respectively.
- switch fabric 250 may be a Batcher-Banyan switch network, Sunshine switch, or any other switch fabric capable of switching ATM cells.
- Switch fabric 250 interfaces with N output ports 260, through 260 N via output links 230, through 230 N respectively.
- Each output port 260 interfaces with an output line 122, which can carry K virtual circuits.
- OUTPUT PORTS Figure 3 illustrates a preferred embodiment for each output port 260, through 260 N in accordance with an embodiment of the invention.
- each output port for example output port 260 N , preferably has access to central processing unit 310 (CPU), memory unit 360, controller 390, secondary storage 380, and high speed buses 370, 371, and 372.
- CPU 310 interfaces with controller 390 via high speed bus 372.
- Controller 390 interfaces with memory unit 360 and secondary storage 380 via high speed buses 370 and 371, respectively.
- Memory unit 360 preferably includes VC_Handler 320, Port_Handler 330, Bandwidth_Allocation_Update 340, and Buffer 350.
- VC_Handler 320 VC_Handler 320,
- Port_Handler 330, Bandwidth_Allocation_Update 340, and Buffer 350 each preferably include a set of instructions in the form of software, which CPU 310 executes.
- VCJHandler 320 receives and processes RM cells for virtual circuits VC, through VC K in output port 260 N .
- Port_Handler 330 determines a new ER and Exp_Rate for the VC counter and recomputes the global bandwidth parameters for all VCs (i.e.,VC, through VC K ) in output port 260 N .
- Bandwidth_Allocation_Update 340 recomputes the total bandwidth that output port 260 N can make available to active VCs that are bottlenecked at output port 260 N .
- Buffer 350 stores cells (i.e, data cells and RM cells) that arrive at output port 260 N in queues (not shown), which are in memory unit 360.
- each VC_Handler 320, PortJHandler 330, Bandwidth_Allocation_Update 340, and Buffers 350 may be implemented in hardware using hardware technology known to one of ordinary skill in the art to provide the functionality described herein.
- Secondary storage 380 includes disk drive unit 382 and tape cartridge 381.
- disk drive unit 382 Stored in disk drive unit 382 are software and data for switching system 105 2 .
- disk drive unit 382 contains software for VCJHandler 320, PortJHandler 330, Bandwidth_Allocation_Update 340, and Buffer 350.
- Secondary storage 380 can copy software and data for switching system 105 2 from tape cartridge 381 into disk drive unit 382. Controller 390 can then upload the software and data from disk drive unit 382 into memory unit 360. Similarly, controller 390 can download software and data from memory unit 360 into disk drive unit 382. Secondary storage 380 can then copy the downloaded software and data from disk drive unit 382 into tape cartridge 381.
- a VC_Handler processes and stores data specific to each virtual circuit VC, through VC K in an output port. Specifically, a VCJHandler preferably stores for each VC forbidden where 1 ⁇ i ⁇ K, a MCR, the time that an RM cell last arrived for the VC cramp an ExpJRate, and a Bottlenecked_Else where indicator.
- the VC_Handler sets Bottlenecked Elsewhere indicator to a logical value of 1, when at the time the VC_Handler receives an RM cell for a VC, the VC, is bottlenecked elsewhere. If the VC, is instead bottlenecked at an output port of the switching system, then the VCJHandler sets Bottlenecked_Elsewhere indicator to a logical value of 0.
- FIG. 4 illustrates a flow chart of the steps a VC_Handler preferably performs to process an RM cell for a VC,.
- VCJHandler 320 receives RM cell 130 from switch fabric 250 (step 400).
- VCJHandler 320 then generates an RM_Cell_Request (step 405).
- An RM_Cell_Request includes a MCR, BottleneckedJElsewhere indicator, current Exp_Rate, CCR, and a current ER for the VC,.
- VCJHandler 320 then sends the RM_Cell_Request to PortJHandler 330 (step 410).
- Port_Handler 330 computes a new ER, Exp_Rate, and Bottlenecked_Elsewhere indicator, and incorporates them in the RM_Cell_Request.
- VC_Handler 320 waits until it receives the RM_Cell_Request back from
- PortJHandler 330 (step 415). If VCJHandler 320 does not receive the RM_Cell_Request back from PortJHandler 330 (step 415), then VCJHandler continues to wait until it receives the RM_Cell_Request from PortJHandler 330 (step 420). If VCJHandler 320 receives the RM_Cell_Request back from PortJHandler 330, then VCJHandler 320 resumes the processing of the RM_Cell_Request (step 425).
- VCJHandler 320 stores the new ER, Exp_Rate, and the Bottlenecked_Elsewhere indicator in memory unit 360 (step 430). Alternatively, VCJHandler 320 may store each of these values in a register. Finally, VCJHandler 320 replaces the current ER in RM cell 130 with the new ER in RM_Cell_Request (step 435). VC_Handler 320 then sends RM cell 130 to Buffer 350 in output port 260 N (step 440).
- a PortJHandler determines a new ER and Exp_Rate for a V , and recomputes the global bandwidth parameters for virtual circuits VC, through VC K in the corresponding output port.
- a PortJHandler preferably stores a weight factor, W j , for each VC,, where 0 ⁇ W, and 1 ⁇ i ⁇ K, the time of the most recent bandwidth allocation update, the time of the second most recent bandwidth allocation update, the total bandwidth available to all VCs (i.e., CBR, VBR, ABR, and UBR VCs) in the output port, the total bandwidth available to all ABR VCs (i.e., active and inactive ABR VCs), the total ExpJRate for active ABR VCs bottlenecked elsewhere, total Exp_Rate for recently active ABR VCs bottlenecked elsewhere, the total bandwidth available to ABR VCs bottlenecked at the output port, the sum of weight factors for active ABR VCs bottle
- FIG. 5 illustrates a flow chart of the steps a PortJHandler preferably performs to process an RM_Cell_Request for a VC,.
- PortJHandler 330 receives RM_Cell_Request from VC_Handler 320 (step 500).
- PortJHandler 330 recomputes global bandwidth parameters for output port 260 N , and new ER and Exp_Rate for the VC, (step 505).
- Port_Handler 330 stores the new global bandwidth parameters for output port 260 N in memory unit 360 (step 510).
- Port_Handler 330 sets the global bandwidth parameters for the VC, in the RM_Cell_Request (step 515).
- PortJHandler 330 then returns the updated RM_Cell_Request to VCJHandler 320 (step 520).
- FIG. 6 illustrates a flow chart of the steps a Port_Handler preferably performs to determine new global bandwidth parameters for an output port and a VC,.
- Port_Handler 330 determines whether the last RM cell for the VC, arrived at output port 260 N within the two most recent bandwidth allocation updates in output port 260 N (step 600).
- Port_Handler 330 determines that the last RM cell for the VC- arrived at output port 260 N before the two most recent bandwidth allocation updates in output port 260 N (step 610).
- PortJHandler 330 resets the current Exp_Rate for the VC, to zero (step 615), and determines that the VC, is bottlenecked elsewhere and sets the Bottlenecked J ⁇ lse where indicator in the RM_Cell_Request equal to a logical 1 (step 620).
- Port_Handler determines whether the VC, is bottlenecked elsewhere (step 625).
- PortJHandler 330 determines that the last RM cell for the VC, arrived at output port 260 N within the two most recent bandwidth allocation updates in output port 260 N (step 605), then PortJHandler 330 determines whether the VC, is bottlenecked elsewhere (step 625).
- PortJHandler 330 determines an upper threshold, namely "cutoff,," for the amount of bandwidth that output port 260 N can make available to the VC,.
- PortJHandler 330 preferably determines cutoff, for the VC, by dividing the total bandwidth available to active ABR VCs bottlenecked at output port 260 N plus the current Exp_Rate for the VC, by the sum of weight factors for active ABR VCs bottlenecked at output port 260 N plus weight factor, W,, for the VC, and multiplying the result by W, (step 645).
- PortJHandler 330 determines that the VC, is bottlenecked at output port
- Port_Handler 330 sets cutoff, for the VC, equal to the total bandwidth available to active ABR VCs bottlenecked at output port 260 N divided by the sum of weight factors for active ABR VCs bottlenecked at output port 260 N multiplied by weight factor, W,, for the VC, (step 650).
- Step 650 may alternatively be expressed as follows:
- PortJHandler 330 determines a new ER for the VC, (step 655), a new ExpJRate for the VC, (step 660), a new total Exp_Rate for recently active ABR VCs bottlenecked elsewhere (step 665), a new total Exp_Rate for active ABR VCs bottlenecked elsewhere (step 670), a new sum of weight factors for active ABR VCs bottlenecked at output port 260 N (step 675), a new total bandwidth that output port 260 N can make available to active ABR VCs bottlenecked at output port 260 N (step 680), and a new sum of weight factors for active ABR VCs bottlenecked at output port 260 N (step 685).
- FIG. 7 illustrates a flow chart of the steps a Port Handler preferably performs to determine a new ER for a VC,.
- Port_Handler 330 determines whether cutoff, for the VC, is less than or equal to MCR for that VC, (step 700). If
- Port_Handler 330 determines that cutoff, for the VC, is greater than MCR (step 710), Port_Handler 330 sets the new ER equal to the minimum of the current ER and MCR (step 715). Port_Handler 330 then determines that the VC, is bottlenecked elsewhere, and sets the Bottlenecked_Elsewhere indicator in the RM_Cell_Request to a logical 1 (step 755).
- PortJHandler 330 determines that cutoff, for the VC, is less than or equal to MCR for that VC, (step 705), PortJHandler 330 determines a most likely new Exp_Rate for the VC, (step 720). Port Handler 330 selects the maximum of CCR and MCR for the VC font and sets the most likely new Exp_Rate to the minimum of that maximum value and the current ER for the VC,.
- PortJHandler 330 determines whether cutoff, for the VC, is greater than the most likely new Exp_Rate for VC (step 725). If PortJHandler 330 determines that cutoff, is not greater than the most likely new Exp_Rate (step 735), then Port_Handler 330 sets the new ER for the VC, to the minimum of current ER and cutoff, (step 740). PortJHandler 330 then determines that the VC, is bottlenecked elsewhere and sets the BottleneckedJElsewhere indicator in the RM_Cell_Request to a logical 1 (step 755).
- PortJHandler 330 determines that cutoff, for the VC, is greater than the most likely new Exp_Rate (step 730), PortJHandler 330 sets the new ER equal to cutoff, (step 745). PortJHandler 330 then determines that the VC, is bottlenecked at output port 260 N , and sets the BottleneckedJBlsewhere indicator in the RM_Cell_Request to a logical 0 (step 750).
- FIG. 8 illustrates a flow chart of the steps a PortJHandler preferably performs to determine a new total Exp_Rate for active ABR VCs that are bottlenecked elsewhere.
- PortJHandler 330 determines whether a VC, is currently bottlenecked elsewhere (step 800). If Port-Handler 330 determines that the VC, is bottlenecked at output port 260 N (step 810), then Port_Handler 330 determines whether, when last time output port 260 N received an RM cell for that VC travers the VC, was bottlenecked elsewhere (step 815).
- Port_Handler 33 determines that last time the VC, was bottlenecked at output port 260 N (step 820), then PortJHandler 330 determines that the total Exp_Rate for active ABR VCs that are bottlenecked elsewhere remains unchanged (step 825). If Port_Handler 330 determines that, when last time output port 260 N received an RM cell for the VC literal the VC, was bottlenecked elsewhere (step 830), then Port_Handler 330 decreases the total Exp_Rate for active ABR VCs that are bottlenecked elsewhere by the current Exp_Rate for the VC, (step 835).
- PortJHandler 330 determines whether, when last time output port 260 N received an RM cell for that VC folk the VC, was bottlenecked elsewhere (step 840). If Port_Handler 330 determines that last time the VC, was bottlenecked at output port 260 N (step 845), PortJHandler 330 increases the total Exp_Rate for active ABR VCs that are bottlenecked elsewhere by the new Exp_Rate for the VC, (step 850).
- PortJHandler 330 determines that last time the VC, was bottlenecked elsewhere (step 855), the PortJHandler 330 decreases the total Exp_Rate for active ABR VCs by the current ExpJRate for the VC, and increases the total ExpJRate for active ABR VCs by the new Exp_Rate for the VC, (step 860).
- Figure 9 illustrates a flow chart of the steps a Port_Handler preferably performs to determine a new total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere.
- Port Handler 330 determines whether the last RM cell for a VC, arrived at output port 260 N after the most recent bandwidth allocation update in output port 260 N (step 900).
- Port Handler 330 determines that the last RM cell for the VC, arrived at output port 260 N before the most recent bandwidth allocation update in output port 260 N (step 905), then PortJHandler 330 determines whether the VC, is currently bottlenecked elsewhere (step 920). If PortJHandler determines that the VC, is currently bottlenecked elsewhere (step 920), the Port_Handler 330 increases the total ExpJRate for recently active ABR VCs that are bottlenecked elsewhere by the new Exp_Rate for the VC, (step 980).
- Port Handler 330 determines that the VC, is currently bottlenecked at output port 260 N (step 915), then PortJHandler 330 determines that the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere remains unchanged (step 925).
- PortJHandler 330 determines whether the VC, is currently bottlenecked elsewhere (step 935). If PortJHandler 330 determines that the VC, is bottlenecked at output port 260 N (step 940), then PortJHandler 330 determines whether, when last time output port 260 N received an RM cell for that VC travers the VC, was bottlenecked elsewhere (step 950).
- PortJHandler 330 determines that last time the VC, was bottlenecked at output port 260 N (step 955), then PortJHandler 330 determines that the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere remains unchanged (step 925).
- PortJHandler 330 determines that the VC, is bottlenecked at output port 260 N (step 940) and determines that, when last time output port 260 N received an RM cell for that VC,, the VC, was bottlenecked elsewhere (step 960), then PortJHandler 330 decreases the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere by the current Exp_Rate for the VC, (step 965).
- PortJHandler 330 determines whether, when last time output port 260 N received an RM cell for that VC literal the VC, was bottlenecked elsewhere (step 970). If PortJHandler determines that last time the VC, was bottlenecked at output port 260 N (step 975), then PortJHandler 330 increases the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere by the new Exp_Rate for the VC, (step 980).
- Port JHandler 330 determines that last RM cell for the VC, arrived after the most recent bandwidth allocation update in output port 260 N (step 930), determines that the VC, is currently bottlenecked elsewhere (step 945), and determines that, when last time output port 260 N received an RM cell for that VC,, the VC, was bottlenecked elsewhere (step 985), then PortJHandler 330 decreases the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere by the current Exp_Rate for the VC, and increases the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere by the new Exp_Rate for the VC, (step 990).
- FIG 10 illustrates a flow chart of the steps a PortJHandler preferably performs to determine the new sum of weight factors for active ABR VCs that are bottlenecked at an output port.
- Port_Handler 330 determines whether a VC, is currently bottlenecked elsewhere (step 1000). If PortJHandler 330 determines that the VC, is currently bottlenecked at output port 260 N (step 1005), then PortJHandler 330 determines whether, when last time output port 260 N received an RM cell for that VC persist the VC, was bottlenecked elsewhere (step 1015).
- PortJHandler 330 determines that last time the VC, was bottlenecked at output port 260 N (step 1060), then PortJHandler 330 determines that the sum of weight factors for active ABR VCs that are bottlenecked at output port 260 N remains unchanged (step 1050). If PortJHandler 330 determines that last time the VC, was bottlenecked elsewhere (step 1020), then PortJHandler 330 increases the sum of weight factors for active ABR VCs that are bottlenecked at output port 260 N by W, of VC, (step 1025).
- PortJHandler 330 determines whether, when last time output port 260 N received an RM cell for that VC,, the VC, was bottlenecked elsewhere (step 1030). If PortJHandler 330 determines that last time the VC, was bottlenecked at output port 260 N (step 1040), then PortJHandler 330 decreases the sum of weight factors for active ABR VCs that are bottlenecked elsewhere by W, of VC, (step 1045).
- PortJHandler 330 determines that the VC, is currently bottlenecked elsewhere (step 1010) and determines that, when last time output port 260 N received an RM cell for that VC shortcut the VC, was bottlenecked elsewhere (step 1035), then
- PortJHandler 330 determines that the sum of weight factors for active ABR VCs that are bottlenecked at output port 260 N remains unchanged (step 1050).
- Figure 11 illustrates a flow chart of the steps a Port_Handler preferably performs to determine the new sum of weight factors for recently active ABR VCs that are bottlenecked at an output port. If Port Handler 330 determines that the last RM cell for a VC, arrived before the most recent bandwidth allocation update (step 1105), then Port_Handler 330 determines whether the VC, is currently bottlenecked elsewhere (step 1115).
- PortJHandler 330 determines that the VC, is currently bottlenecked at output port 260 N (step 1120), then PortJHandler 330 determines that the sum of weight factors for recently active ABR VCs that are bottlenecked at output port 260 N remains unchanged (step 1165). If Port_Handler 330 determines that the last RM cell for the VC, arrived before the most recent bandwidth allocation update (step 1105) and determines that the VC, is currently bottlenecked elsewhere (step 1125), then PortJHandler 330 increases the sum of weight factors for recently active ABR VCs that are bottlenecked at output port 260 N by W, of the VC, (step 1130).
- PortJHandler 330 determines whether, when last time output port 260 N received an RM cell for that VC,, the VC, was bottlenecked elsewhere (step 1135). If Port_Handler 330 determines that last time the VC, was bottlenecked elsewhere (step 1140), then PortJHandler 330 determines whether the VC, is currently bottlenecked elsewhere (step 1115).
- PortJHandler 330 determines that the VC, is currently bottlenecked at output port 260 N (step 1120), then PortJHandler 330 determines that the sum of weight factors for recently active ABR VCs that are bottlenecked at output port 260 N remains unchanged (step 1165). If PortJHandler 330 determines that, when last time output port 260 N received an RM cell for that VC shortcut the VC, was bottlenecked elsewhere (step 1140) and determines that the VC, is currently bottlenecked elsewhere (step 1125), then PortJHandler 330 increases the sum of weight factors for recently active ABR VCs that are bottlenecked at output port 260 N by W, of VC, (step 1130).
- PortJHandler 330 determines that the last RM cell for the VC, arrived after the most recent bandwidth allocation update (step 1110) and determines that the VC, was bottlenecked at output port 260 N (step 1145), then PortJHandler 330 determines whether the VC, is currently bottlenecked elsewhere (step 1150). If PortJHandler 330 determines that the VC, is currently bottlenecked at output port 260 N (step 1155), then PortJHandler 330 determines that the sum of weight factors for recently active ABR VCs that are bottlenecked at output port 260N remains unchanged (step 1165).
- PortJHandler 330 determines that the last RM cell for the VC, arrived after the most recent bandwidth allocation update (step 1110), determines that, when last time output port 260 N received an RM cell for that VC,, the VC, was bottlenecked at output port 260 N (step 1145), and determines that the VC, is currently bottlenecked elsewhere (step 1160), then PortJHandler 330 decreases the sum of weight factors for active ABR VCs that are bottlenecked at output port 260 N by W, of VC, (step 1170).
- FIG 12 illustrates a flow chart of the steps a Port_Handler preferably performs to determine the new total bandwidth that an output port can make available to active ABR VCs that are bottlenecked at the output port.
- PortJHandler 330 determines whether a VC, is currently bottlenecked elsewhere (step 1200). If PortJHandler 330 determines that the VC, is bottlenecked at output port 260 N (step 1205), then PortJHandler 330 determines whether, when last time output port 260 N received an RM cell for that VC travers the VC, was bottlenecked elsewhere (step 1215).
- PortJHandler 330 determines that last time the VC, was bottlenecked at output port 260 N (step 1220), then Port_Handler 330 determines that the total bandwidth that output port 260 N can make available to active ABR VCs at output port 260 N remains unchanged (step 1230).
- Port JHandler 330 determines that VC, is bottlenecked at output port 260 N (step 1205) and determines that, when last time output port 260 N received an RM cell for that VC literal the VC, was bottlenecked elsewhere (step 1225), then Port_Handler 330 increases the total bandwidth that output port 260 N can make available to active ABR VCs that are bottlenecked at output port 260 N by the current Exp_Rate for the VC, (step 1235).
- Port JHandler 330 determines that the VC, is bottlenecked elsewhere (step 1210), then Port_Handler 330 determines whether, when last time output port 260 N received an RM cell for that VC,, the VC, was bottlenecked elsewhere (step 1210), then Port_Handler 330 determines whether, when last time output port 260 N received an RM cell for that VC,, the VC, was bottlenecked elsewhere (step 1210), then Port_Handler 330 determines whether, when last time output port 260 N received an RM cell for that VC, the VC, was bottlenecked elsewhere (step 1210), then Port_Handler 330 determines whether, when last time output port 260 N received an RM cell for that VC, the VC, was bottlenecked elsewhere (step 1210), then Port_Handler 330 determines whether, when last time output port 260 N received an RM cell for that VC, the VC, was bottlenecked elsewhere (step 1210), then Port
- PortJHandler 330 decreases the total bandwidth available to active ABR VCs that are bottlenecked at output port 260 N by the new Exp_Rate for the VC, (step 1255).
- PortJHandler 330 determines that the VC, is bottlenecked elsewhere (step 1210) and determines that, when last time output port 260 N received an RM cell for that VC folk the VC, was bottlenecked elsewhere (step 1250), then Port_Handler 330 increases the total bandwidth that output port 260 N can make available to active ABR VCs that are bottlenecked at output port 260 N by the current Exp_Rate for the VC, and decreases the total bandwidth that output port 260 N can make available to active ABR VCs that are bottlenecked at output port 260 N by the new ExpJRate for the V (step 1260).
- each output port invokes a corresponding Bandwidth_ Allocation J pdate program.
- output port 260 N invokes Bandwidth_Allocation_Update 340, which preferably stores the time interval between each invocation in a static memory.
- the manufacturer of a switching system may set a default for the fixed time interval, which a switching system administrator or network administrator may subsequently reconfigure.
- the time interval between each invocation may preferably be twice the upper bound for the time interval between each RM cell generated by an active virtual source.
- the upper bound may preferably be in the range of 2 "5 milliseconds to 100 milliseconds.
- the time interval between each invocation may be less than twice the upper bound for the time interval between each RM cell, in which case the time interval between each invocation must still be sufficiently large to allow for arrival of at least one RM cell between two consecutive invocations of Bandwidth .
- AllocationJ pdate 340 AllocationJ pdate 340.
- Bandwidth_Allocation_Update 340 preferably has access to the global bandwidth parameters stored in memory unit 360 by PortJHandler 330. Specifically, Bandwidth_Allocation JJpdate 340 has access to the time of the most recent bandwidth allocation update, the time of a second most recent bandwidth allocation update, the total bandwidth available to all ABR VCs (i.e., active and inactive ABR VCs), the total Exp_Rate for active ABR VCs bottlenecked elsewhere, the total Exp_Rate for recently active ABR VCs bottlenecked elsewhere, the total bandwidth available to active ABR VCs bottlenecked at output port 260 N , the sum of weight factors for active ABR VCs bottlenecked at output port 260 N , and the sum of weight factors for recently active ABR VCs bottlenecked at the output port 260 ⁇
- Figure 13 illustrates a flow chart of the steps a Bandwidth_Allocation JJpdate program preferably performs to make available a new total bandwidth
- Bandwidth_Allocation_Update 340 sets the sum of weight factors for active ABR VCs that are bottlenecked at output port 260 N equal to the sum of weight factors for recently active ABR VCs that are bottlenecked at output port 260 N (step 1305).
- Bandwidth_Allocation JJpdate 340 then sets the total bandwidth that output port 260 N can make available to active ABR VCs that are bottlenecked at output port 260 N equal to the difference between the total bandwidth available to all ABR VCs (i.e., active and inactive ABR VCs) and the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere (step 1310).
- Bandwidth_Allocation_Update 340 resets the total Exp_Rate for recently active ABR VCs that are bottlenecked elsewhere (step 1315).
- Bandwidth_Allocation JJpdate 340 then resets the sum of weight factors for recently active ABR VCs that are bottlenecked at output port 260 N (step 1320).
- Bandwidth_Allocation JJpdate 340 sets a timer for the next update to the total bandwidth that output port 260 N can make available to active ABR VCs that are bottlenecked at output port 260 N (step 1325).
- an ATM network may assign a high priority to RM cells.
- the time interval between the updates to the global bandwidth parameters i.e., invocation of Bandwidth_Allocation JJpdate
- the impact to allocation of an optimum bandwidth to the VC would be minimal.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
Description
Claims
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US6422497P | 1997-10-29 | 1997-10-29 | |
US64224P | 1997-10-29 | ||
PCT/US1998/017788 WO1999022498A1 (en) | 1997-10-29 | 1998-08-27 | Method and system for dynamic allocation of bandwidth in asynchronous transfer mode (atm) switching systems |
Publications (2)
Publication Number | Publication Date |
---|---|
EP1027789A1 true EP1027789A1 (en) | 2000-08-16 |
EP1027789A4 EP1027789A4 (en) | 2001-02-07 |
Family
ID=22054409
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP98944557A Withdrawn EP1027789A4 (en) | 1997-10-29 | 1998-08-27 | Method and system for dynamic allocation of bandwidth in asynchronous transfer mode (atm) switching systems |
Country Status (4)
Country | Link |
---|---|
EP (1) | EP1027789A4 (en) |
JP (1) | JP2001522170A (en) |
CA (1) | CA2307672A1 (en) |
WO (1) | WO1999022498A1 (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6657964B1 (en) * | 1998-05-26 | 2003-12-02 | Hitachi, Ltd. | Terminal with bandwidth control function |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5675576A (en) * | 1995-06-05 | 1997-10-07 | Lucent Technologies Inc. | Concestion control system and method for packet switched networks providing max-min fairness |
-
1998
- 1998-08-27 CA CA002307672A patent/CA2307672A1/en not_active Abandoned
- 1998-08-27 EP EP98944557A patent/EP1027789A4/en not_active Withdrawn
- 1998-08-27 JP JP2000518485A patent/JP2001522170A/en active Pending
- 1998-08-27 WO PCT/US1998/017788 patent/WO1999022498A1/en not_active Application Discontinuation
Non-Patent Citations (2)
Title |
---|
No further relevant documents disclosed * |
See also references of WO9922498A1 * |
Also Published As
Publication number | Publication date |
---|---|
WO1999022498A1 (en) | 1999-05-06 |
EP1027789A4 (en) | 2001-02-07 |
JP2001522170A (en) | 2001-11-13 |
CA2307672A1 (en) | 1999-05-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US6442138B1 (en) | Method and apparatus for controlling admission of connection requests | |
US6449253B1 (en) | Method and system for dynamic allocation of bandwidth in asynchronous transfer mode (ATM) switching systems | |
US6041039A (en) | System and method for determining network bandwidth availability using priority level feedback | |
US5408465A (en) | Flexible scheme for admission control of multimedia streams on integrated networks | |
US6167030A (en) | Buffer-based traffic measurement system and method for nominal bit rate (NBR) service | |
CA2214897C (en) | Broadband switching network | |
JP5206714B2 (en) | Transmission terminal device, network node, and relay switch | |
US6442164B1 (en) | Method and system for allocating bandwidth and buffer resources to constant bit rate (CBR) traffic | |
AU718665B2 (en) | Broadband switching system | |
AU719514B2 (en) | Broadband switching system | |
EP1183901A1 (en) | Bi-modal control system and method for partitioning a shared output buffer in a connection-oriented network | |
WO2000057669A1 (en) | Timer-based traffic measurement system and method for nominal bit rate (nbr) service | |
WO2000076148A1 (en) | Method for allocating bandwidth to real-time variable bit rate (rt-vbr) traffic | |
US6587436B1 (en) | Method and apparatus for allocation of available bandwidth | |
JP3457522B2 (en) | Packet communication method | |
US6331970B1 (en) | Dynamic generic cell rate algorithm for policing ABR traffic | |
US6163808A (en) | Network adjusts cell transfer capacity in response to a change in the actual bit rate relative to the nominal bit rate | |
EP0884923B1 (en) | Packet switching network, packet switching equipment, and network management equipment | |
EP0481447B1 (en) | Method of controlling communication network incorporating virtual channels exchange nodes and virtual paths exchange nodes, and the said communication network | |
US6542509B1 (en) | Virtual path level fairness | |
JP4238415B2 (en) | Transmission terminal device, network node, and relay switch | |
WO1999022498A1 (en) | Method and system for dynamic allocation of bandwidth in asynchronous transfer mode (atm) switching systems | |
EP0960513A2 (en) | Method and system for dynamic allocation of bandwidth in asynchronous transfer mode (atm) switching systems | |
US7130267B1 (en) | System and method for allocating bandwidth in a network node | |
JPH11136252A (en) | Atm switch and shaping method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 20000309 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): DE FR GB IT SE |
|
A4 | Supplementary search report drawn up and despatched |
Effective date: 20001229 |
|
AK | Designated contracting states |
Kind code of ref document: A4 Designated state(s): DE FR GB IT SE |
|
RIC1 | Information provided on ipc code assigned before grant |
Free format text: 7H 04L 12/56 A, 7H 04Q 11/04 B |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN |
|
18D | Application deemed to be withdrawn |
Effective date: 20030301 |