WO2003017590A2 - Enlargement of a computation rule for data packet switching - Google Patents
Enlargement of a computation rule for data packet switching Download PDFInfo
- Publication number
- WO2003017590A2 WO2003017590A2 PCT/IB2002/003226 IB0203226W WO03017590A2 WO 2003017590 A2 WO2003017590 A2 WO 2003017590A2 IB 0203226 W IB0203226 W IB 0203226W WO 03017590 A2 WO03017590 A2 WO 03017590A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- matrix
- switching device
- packet switching
- columns
- allocation unit
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/12—Shortest path evaluation
Definitions
- the invention relates to a packet switching device with a plurality of line and switching units for switching data packets on the basis of a calculation rule.
- the publication deals with a weighting-based calculation rule for linking input and output ports of the switching device.
- the weightings of the links for switching an input port to an output port are represented in a status matrix.
- the rows and columns of the status matrix correspond to the input and output port of the switching device for data packets.
- a maximum weighting is selected in each column of the matrix and marked as an accepted link between an input and an output port.
- the selected input port or the line of the status matrix corresponding to the input port is no longer taken into account.
- the marked links are recorded in a decision matrix, on the basis of which the links between the input and output ports are created.
- the invention has for its object to enable a quickly calculated switching of data packets within a packet switching device.
- the object is achieved by a packet switching device with a plurality of input and output ports and at least one switching unit consisting of a coupling matrix and an allocation unit for controlling the coupling matrix, the status unit having elements in the allocation unit, each of which weighting a link between an input port and a Output port of
- a packet switching device consists of several line units (line card) and several switching units (switch card).
- a key component of a line unit is a port control.
- Each port controller is connected to a plurality of switching units operating in parallel and has the task of arranging packets arriving at the packet switching device to be switched according to priority and desired output port of the packet switching device in queues.
- Each switching unit of the packet switching device consists of a coupling matrix (crosspoint matrix) and an allocation unit (arbiter).
- the configuration of the coupling matrix is redetermined at regular intervals by the allocation unit by means of a calculation rule and thus new links between the input and output port of the packet switching device for switching the packets are created.
- a state matrix of the packet switching device contains several elements that represent a weighting of each linkage of an input port with an output port.
- the weighting is generated by the port control, so the port control can inform the allocation unit whether and how urgently the packets in the incoming port should be conveyed.
- a weighting can contain information about the priority and class of the packets or the waiting time or size of a queue.
- One row of the state matrix corresponds to an input port and each column corresponds to an output port of the packet switching device.
- the allocation unit uses a vector to rearrange the columns within the status matrix.
- the vector consists of vector elements, each of which is assigned to a column of the state matrix. When reordering the columns, the size of the vector elements is crucial. A new The order of the columns in the status matrix depends on the vector element assigned to the column.
- the rearrangement of the columns results in a second status matrix from the first status matrix, which differs from the first status matrix in the order of the columns.
- the allocation unit uses the second status matrix to carry out a calculation rule for selecting the provisional links between the input ports and the output port.
- the links are recorded in a first decision matrix.
- the columns of the first decision matrix are rearranged in the original order of the first status matrix by the allocation unit, so that a second decision matrix is created.
- the links marked in the decision matrix are sent as accepted links to the coupling matrix for configuration and the input ports are linked to the output ports for the transmission of data packets.
- the calculation rule according to which the allocation unit works can be a
- SIMP Successive Incremental Matching over Multiple Ports
- the calculation rule can quickly determine favorable links between the input ports and the output ports and thus lead to an excellent performance of the packet switching device.
- the port control generates and sends the weightings for determining the configuration of the coupling matrix to the allocation unit at regular time intervals.
- a preferred embodiment of the status matrix contains elements which each represent a difference between a current weighting of a link generated by the port control and a weighting generated in the previous time interval.
- the data packets are temporarily stored in queues before they leave the port controller and the packet switching device. If an operating speed of the output ports is slower than the operating speed of the packet switching device, the data packets cannot leave the output ports sufficiently quickly and the queues are overcrowded with the data packets.
- a queue of the data packets reproduces a memory content of an output port.
- the vector elements required to rearrange the state matrix each represent a memory content of an output port.
- the reordering of the columns depends on the size of the vector elements.
- a column which is assigned to a vector element representing memory contents filled with data packets is arranged as the last column of the second status matrix.
- a column assigned to an empty memory content representing vector element is arranged as the first column of the second status matrix and thus used as the first by the allocation unit for processing the calculation rule.
- the vector elements each indicate the sum of all the elements arranged in a column or consist of the largest element in each case of a column of the first status matrix.
- the arrangement of the columns of the second state matrix also depends on the size of the vector element in these cases.
- the allocation unit is not only able to work on the second state matrix, it can also directly access the original state matrix, the order in which the columns of the state matrix are processed is determined by the vector.
- the invention also relates to a switching unit consisting of a switching matrix and an allocation unit for controlling the switching matrix for one
- the allocation unit works on the basis of a status matrix with elements which each represent a weighting of a link between an input port and an output port of the packet switching device and whose rows each correspond to an input port and columns each to an output port of the packet switching device.
- the allocation unit arranges the columns depending on a vector, the vector elements of which are each assigned to a column.
- the invention relates to an allocation unit for controlling the coupling matrix of a switching unit for a packet switching device with several Entry and exit ports.
- the allocation unit works on the basis of a status matrix with elements which each represent a weighting of a link between an input port and an output port of the packet switching device and whose rows each correspond to an input port and columns each correspond to an output port of the packet switching device.
- the allocation unit uses a vector to arrange the columns as a function of a vector element assigned to each column.
- the invention further relates to a switching method for a packet switching device with a plurality of input and output ports and at least one switching unit consisting of a switching matrix and an allocation unit for controlling the switching matrix.
- the allocation unit works on the basis of a status matrix with elements which each represent a weighting of a link between an input port and an output port of the packet switching device and whose rows each correspond to an input port and columns each correspond to an output port of the packet switching device.
- the columns of the state matrix are arranged by means of a vector as a function of a vector element assigned to each column.
- a packet switching device shown in Fig. 1 consists of two line units 1 and 2 and several switching units 3 to 5.
- the number of line units 1 and 2 is often much larger than two, for example 64, in the exemplary embodiment it was limited to two due to a clear representation.
- the Line units 1 and 2 each consist of different components that are not relevant to the description of the exemplary embodiment and therefore not shown, such as an optical transmission unit, frame generator, network processor, etc., and a port controller 6 and 7 that is essential for the exemplary embodiment.
- Each switching unit 3 to 5 consists of a Coupling matrix 8 and an allocation unit 9. Both the allocation unit 9 and the coupling matrix 8 of each switching unit 3 to 5 operating in parallel are each connected to the port controller 6 and 7.
- Data packets of constant length are called cells. Since cells of constant length are easier to handle than packets of varying sizes, the incoming packets within line units 1 and 2 are divided into cells of constant length and buffered in queues. After a successful placement, i.e. Accepted assignment of an input port to an output port, the cells are removed from the queue.
- the port control can use a weighting to inform the allocation unit whether and how urgently the cell located in the input port should be conveyed.
- the weighting can include information about the priority and class of the packets or the waiting time or size of a queue in the incoming port.
- the port control 6 and 7 sends the information necessary to determine the configuration of the coupling matrix, including the weighting to the allocation unit 9. After the allocation unit has determined the configuration of the switching matrix 8 and has sent the configuration both to the port control and to the switching matrix, the switching matrix 8 of each switching unit 3 to 5 is configured accordingly.
- the port control 6 and 7 sends the cells to the switching matrix 8 at regular intervals for the so-called cell periods. In FIG. 2, one is carried out by the allocation unit 9 and on one
- State matrix 10 applied calculation rule for the configuration of the coupling matrix 8 shown.
- the calculation rule is presented in the publication "Weighted Arbitration Algorithms with Priorities for Input-Queued Switches with 100% Throughput" by R. Schoenen, G. Post, G. Sander, Broadband Switching Symposium '99.
- Fig. 2 consists of a first state matrix 10, a vector 11, a second status matrix 12, the calculation rule 13, a preliminary decision matrix 14, and a final decision matrix 15.
- the status matrix 10 corresponds to a packet switching device with four input ports (input) 10 to 13 and four output ports (output) O0 to O3. It contains sixteen elements representing the weight of each link between an input port 10 through 13 and an output port O0 through 03.
- the number of columns and rows or the input and output port was limited to four in order to achieve a clear representation.
- the vector 11 consists of vector elements which are each assigned to a column of the status matrix or an output port and represent a memory content of the output port.
- a vector element with zero means that the memory content of the output port is almost empty and that a large storage capacity of the output port is available, so that data packets should be conveyed to the output port.
- the columns of the status matrix are rearranged by the allocation unit.
- the reordering of the columns creates a second status matrix 12, the columns of which are sorted according to the size of the vector elements.
- a column whose vector element represents an almost free memory content is sorted as one of the first columns.
- a column with an almost completely filled memory content of the output port is arranged as one of the last columns of the second status matrix.
- the column of the output port OI was selected as the first column of the status matrix 12.
- the allocation unit 9 sorts the following columns according to the size of the vector element, starting with the smallest vector element. This results in the column of the output port O3 as the second column, then the column of the
- the columns of the second status matrix 12 correspond to the order in which they are used by the allocation unit 9 to carry out the calculation rule 13.
- the links that were selected on the basis of the calculation rule 13 are marked in the provisional decision matrix 14.
- FIG. 3 consists of the same arrangement as FIG. 2, with the difference that the vector elements each consist of a sum of elements from a column of the state matrix 10. As a result, the columns are rearranged differently. The columns are sorted according to the size of the vector element. Since the largest vector element corresponds to the column of the output port OO, the
- the calculation rule 13 is based on the status matrix
- the links that were selected on the basis of the calculation rule 13 are marked in the provisional decision matrix 14.
- the columns of the preliminary decision matrix 14 are arranged in the final decision matrix 15 in the original order of the output ports corresponding to the columns.
- FIG. 4 also consists of the same arrangement as FIG. 2 with the difference that the vector elements of the vector 11 correspond to a largest element of the column assigned to the vector element.
- the column of the output port O2 is assigned to the largest vector element, this column is arranged as the first column in the second status matrix 12 or used as the first to execute the calculation rule. The further order of the columns depends on the size of the vector elements, so that the column of the output port O0 is determined as the next column of the status matrix 13.
- the remaining columns of the status matrix 12 are the column of the output port OI and output port O3.
- the calculation rule 13 is based on the status matrix
- the links that were selected on the basis of the calculation rule 13 are marked in the provisional decision matrix 14.
- the column of the provisional decision matrix 14 is arranged in the final state matrix 15 in the original order of the output ports corresponding to the columns.
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE10140144.2 | 2001-08-16 | ||
DE10140144A DE10140144A1 (en) | 2001-08-16 | 2001-08-16 | Extension of a calculation rule for switching data packets |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2003017590A2 true WO2003017590A2 (en) | 2003-02-27 |
WO2003017590A3 WO2003017590A3 (en) | 2004-06-10 |
Family
ID=7695585
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/IB2002/003226 WO2003017590A2 (en) | 2001-08-16 | 2002-08-02 | Enlargement of a computation rule for data packet switching |
Country Status (2)
Country | Link |
---|---|
DE (1) | DE10140144A1 (en) |
WO (1) | WO2003017590A2 (en) |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5923656A (en) * | 1996-10-22 | 1999-07-13 | Board Of Trustees Of The University Of Illinois | Scalable broad band input-queued ATM switch including weight driven cell scheduler |
-
2001
- 2001-08-16 DE DE10140144A patent/DE10140144A1/en not_active Withdrawn
-
2002
- 2002-08-02 WO PCT/IB2002/003226 patent/WO2003017590A2/en not_active Application Discontinuation
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5923656A (en) * | 1996-10-22 | 1999-07-13 | Board Of Trustees Of The University Of Illinois | Scalable broad band input-queued ATM switch including weight driven cell scheduler |
Non-Patent Citations (2)
Title |
---|
SCHOENEN R ET AL: "DISTRIBUTED CELL SCHEDULING ALGORITHMS FOR VIRTUAL-OUTPUT-QUEUED SWITCHES" 1999 IEEE GLOBAL TELECOMMUNICATIONS CONFERENCE. GLOBECOM'99. SEAMLESS INTERCONNECTION FOR UNIVERSAL SERVICES. RIO DE JANEIRO, BRAZIL, DEC. 5-9, 1999, IEEE GLOBAL TELECOMMUNICATIONS CONFERENCE, NEW YORK, NY: IEEE, US, Bd. 2, 5. Dezember 1999 (1999-12-05), Seiten 1211-1215, XP001016905 ISBN: 0-7803-5797-3 * |
SCHOENEN R: "AN ARCHITECTURE SUPPORTING QUALITY-OF-SERVICE IN VIRTUAL-OUTPUT-QUEUED SWITCHES" IEICE TRANSACTIONS ON COMMUNICATIONS, INSTITUTE OF ELECTRONICS INFORMATION AND COMM. ENG. TOKYO, JP, Bd. E83-B, Nr. 2, Februar 2000 (2000-02), Seiten 171-181, XP001063853 ISSN: 0916-8516 * |
Also Published As
Publication number | Publication date |
---|---|
WO2003017590A3 (en) | 2004-06-10 |
DE10140144A1 (en) | 2003-02-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
DE69917835T2 (en) | Data path structures with rotator switch | |
DE60128413T2 (en) | Featured priority queue scheduler | |
DE60036031T2 (en) | ALLOCATION OF PRIORITY LEVELS IN A DATA FLOW | |
DE19634492B4 (en) | Method for the optimized transmission of ATM cells over connection sections | |
DE69738386T2 (en) | Improvements in or relating to an ATM switch | |
EP0447841A2 (en) | Method for setting-up virtual circuits in an ATM (asynchronous transfer mode) switching system | |
DE19757965A1 (en) | Distributed buffer system for ATM switch | |
DE69631265T2 (en) | switching system | |
DE60119459T2 (en) | A packet switching apparatus and method for controlling the switch | |
DE60222233T2 (en) | Processor and method for maintaining the processing order of packets based on packet stream identifiers | |
DE60303444T2 (en) | PROCESS CONTROL USING QUANTUM VALUES AND DEFICIT VALUES | |
DE69817672T2 (en) | Device for sorting data elements in a binary tree and an ATM spacer with such a device | |
DE10057343A1 (en) | Packet switching device for packet data transport has coupling field provided by coupling matrices and switching controls connected in cascade | |
DE60300827T2 (en) | Communication system and method with a queuing device per service | |
DE60107828T2 (en) | Flow and blockage control in a switched network | |
WO2003017590A2 (en) | Enlargement of a computation rule for data packet switching | |
EP1283624B1 (en) | Prescribed calculation that works parallel an iterative for the avoidance of data packets | |
EP1280303B1 (en) | Packet switching device with a feedback method for its access unit | |
EP1289209A2 (en) | Repartition of weighting factors between the port control and switches of a packet switching device | |
DE102016202331A1 (en) | Network node device and method for switching data | |
DE10128686A1 (en) | Virtual output queue packet switching unit uses port controller status information | |
EP1374505B1 (en) | Circuit arrangement for data stream distribution with conflict resolution | |
EP0804044A1 (en) | Service control in an intelligent network | |
DE60225542T2 (en) | System with feedback for a packet switching apparatus with bufferless, cascaded switching matrix | |
WO1998036586A2 (en) | Method and circuit for transmitting information cells during virtual linking with various priorities |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A2 Designated state(s): CN JP US Kind code of ref document: A2 Designated state(s): CN JP |
|
AL | Designated countries for regional patents |
Kind code of ref document: A2 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FR GB GR IE IT LU MC NL PT SE SK TR Kind code of ref document: A2 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR IE IT LU MC NL PT SE SK TR |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
122 | Ep: pct application non-entry in european phase | ||
NENP | Non-entry into the national phase |
Ref country code: JP |
|
WWW | Wipo information: withdrawn in national office |
Country of ref document: JP |