CN104717159A - Scheduling method based on store-and-forward switching structure - Google Patents
Scheduling method based on store-and-forward switching structure Download PDFInfo
- Publication number
- CN104717159A CN104717159A CN201310688711.8A CN201310688711A CN104717159A CN 104717159 A CN104717159 A CN 104717159A CN 201310688711 A CN201310688711 A CN 201310688711A CN 104717159 A CN104717159 A CN 104717159A
- Authority
- CN
- China
- Prior art keywords
- port
- data frame
- priority
- frame
- scheduling
- 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.)
- Pending
Links
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides a scheduling method based on a store-and-forward switching structure. The scheduling method includes the following steps that 1 data frames are filtered by a switch, data frames not passing filtration are discarded, traffic shaping inspection is conducted on the filtered data frames, and a second step is carried out; 2 two input queues and two output queues are set up at each port; 3 scheduling is received, a scheduling algorithm that RR is fair polled between ports and the high-priority queue is preferential in ports is adopted, date frame information is read from the input queues quickly, and addresses of target ports are obtained according to data frame information; 4 the data frame information is put in an output queue corresponding to an output port and a priority; 5 the scheduling is sent, and groups of queues which are stored in the output port are scheduled. According to the scheduling method, filtration and regulation are conducted before scheduling, data frames which do not conform to rules are discarded, the exchanging resource is saved, and the delaying determinacy is guaranteed during scheduling.
Description
Technical field
The present invention relates to AFDX network communication technology field, refer more particularly to the filtration of AFDX switchboard data frame, traffic shaping, storage, receiving scheduling and transmission scheduling.
Background technology
Packet switching system occupies critical role in current network application, and it is that the multi-service transport such as such as voice, video, data provide effective bearing mode.Business Streams different in onboard networks requires different quality of service guarantees, PSE must control the flow of each transport stream and PSE to the resource such as bandwidth, buffer memory of each transport stream, to provide distinguishing loss rate and transmission delay, adapt to the different requirements of multiple business to service quality.But traditional exchange scheduling mode cannot ensure the certainty of dispatching, the grade of service of application can not be ensured from switching fabric.
Summary of the invention
In order to solve technical problem existing in background technology, the present invention proposes a kind of dispatching method based on storage forwarding switching fabric, carry out before scheduling filtering and control, the Frame not meeting rule is abandoned, save exchange resource, ensure that the time delay certainty of scheduling.
Technical solution of the present invention is: a kind of dispatching method based on storage forwarding switching fabric, is characterized in that: said method comprising the steps of:
1) Frame filters through switch, for not abandoned by the Frame filtered, by the Frame filtered, carrying out traffic shaping inspection, carry out step 2);
2) each port all sets up two input ranks and two output queues;
3) receiving scheduling: adopt the dispatching algorithm that in the fair poll of RR between port, port, high-priority queue is preferential, read data frame information from input rank fast, obtains target port address according to data frame information;
4) data frame information is put in the output queue of the corresponding priority of corresponding output port;
5) scheduling is sent: the grouping being kept at output port queue is dispatched; Realize fair forwarding of packets between the ports, forward according to two-stage priority mechanism in port inside.
Above-mentioned steps 3) in data frame information comprise input rank data frame information and output queue data frame information; Input rank data frame information comprises Frame, and position, data frame length, Frame forward port, Mulitcast Queue distribution and the instruction of configuration frame in the buffer; Output queue data frame information comprises Frame, and position, data frame length, Frame forward port and the instruction of configuration frame in the buffer.
Above-mentioned input rank is different according to Frame priority, and be divided into height two priority, the high-priority data frame information of this port is only deposited in the input rank of high priority, and the input rank of low priority only exists the lower-priority data frame information of this port.
Each two output queues of the corresponding high and low priority of each port of above-mentioned output queue, only there is the data frame information from this port repeat in output queue.
Above-mentioned transmission scheduling adopts " static priority " scheduling strategy to dispatch separate output work queue's system.
Above-mentioned filtration is specifically for the filtration of minimax frame length, VL configuration, Frame completeness check aspect.
The advantageous effect that the present invention has: the scheduling method adopting Store and forword, Frame only stores portion in a switch, decreases Time and place resource overhead.Carry out before scheduling filtering and control, the Frame not meeting rule is abandoned, has saved exchange resource, ensure that the time delay certainty of scheduling.Adopt the scheduling thought based on output queue, and divide high Low Priority Queuing, for different business provides the different grade of service, solve the blockage problem caused when dispatching based on input rank, meet switch to shake and time delay qualitative requirement really.
Accompanying drawing explanation
Fig. 1 is AFDX network topology schematic diagram of the present invention;
Fig. 2 is that Frame frame of the present invention is dropped schematic diagram;
Fig. 3 is packet switching construction schematic diagram of the present invention;
Embodiment
Based on a dispatching method for storage forwarding switching fabric, comprise the following steps:
1) Frame filters through switch, for not abandoned by the Frame filtered, by the Frame filtered, carrying out traffic shaping inspection, carry out step 2);
2) each port all sets up two input ranks and two output queues;
3) receiving scheduling: adopt the dispatching algorithm that in the fair poll of RR between port, port, high-priority queue is preferential, read data frame information from input rank fast, obtains target port address according to data frame information;
4) data frame information is put in the output queue of the corresponding priority of corresponding output port;
5) scheduling is sent: the grouping being kept at output port queue is dispatched; Realize fair forwarding of packets between the ports, forward according to two-stage priority mechanism in port inside.
Switch uses a shared memory cell to carry out store frames of data, in scheduling process, is dispatched by data frame information, is finally forwarded by Frame from shared memory cell, in order to save the Time and place resource of switch.First minimax frame length, No. VL, the filtration of the aspect such as CRC check are carried out to Frame before switch scheduling, in early stage, the Frame not meeting rule is abandoned, save switch resource, ensure the scheduling time of switch valid data frame.Also traffic shaping will be carried out through the token bucket algorithm of switch, to ensure the certainty upper limit of data frames exchange time delay by the Frame filtered.
Exchange scheduling employing " queue scheduling " thought based on frame, comprise receiving scheduling and send scheduling two parts.In order to prevent the spilling of input-buffer and reduce data frame transfer time delay, in switch design, set up special input rank for each input port.Simultaneously for ensureing that different business has different dispatching priorities, high low priority is all distinguished in input rank and output queue, to support two-stage priority service.Frame is through filtering and after traffic shaping, according to its priority, putting it in the input rank of the different priorities of its correspondence.Receiving scheduling adopts the algorithm that in the fair polling dispatching of RR between port, port, high-priority queue is preferential, and read data frame information from input rank fast, obtains target port address according to data frame information.Then data frame information is put in the output queue of corresponding output port.This ensure that in same port, the Frame of high priority is always preferentially scheduled away, ensure that the grade of service of high-priority data frame.
Send scheduling to dispatch the grouping being kept at output port queue.According to ARINC664 standard, send scheduling and should realize forwarding of packets fair fast between the ports, and need to forward according to two-stage priority mechanism in port inside.Each output port queue adopts separate output work queue's system of " static priority (SP) " scheduling strategy, thus meets shake and time delay qualitative requirement really.
AFDX topology of networks as shown in Figure 1.AFDX network is a kind of hub-and-spoke configuration network centered by switch, and network comprises end system and switch two parts.End system is a part for avionics system, is connected with AFDX switch by 10Mbps or 100Mbps full-duplex link.Switch is the key equipment building avionics system data communication backbone network, two switches backup each other, for realizing high speed, the reliably exchanges data of each equipment room of avionics system, ensure the real-time of avionics network transfer of data, certainty and reliability.
Use a shared memory in switch, store the Frame arriving switch.Frame needs through the filtration of switch for aspects such as minimax frame length, VL configuration, Frame completeness checks before scheduling, for not abandoned by the Frame filtered, no longer participate in follow-up exchange scheduling, by the Frame filtered, then need to carry out traffic shaping inspection further, i.e. control.Switch is as follows based on the control algorithm of frame:
1) if byte ACi is larger than Simax, then receive this frame, and byte ACi deducts Simax;
2) if byte ACi is less than Simax, then abandon this frame, corresponding MIB parameter upgrades, and ACi is constant.
The current residue ACi of VLi is the time that the last secondary data frame of Taci, this VLi enters switch is T1, if receive again the Frame mated with this VLi in the T2 moment, then twice receives data break is Ti=T2 – T1,
1) if ACi=Ti+Taci >=Simax, then this frame is by alert pipe, then Taci is updated to min{AciMax, Ti+Taci}-Simax, PCNT is updated to T2;
2) if Ti+Taci<Simax, then this frame is not by alert pipe, and the value of current residual Taci is constant.
Wherein, Simax initial value is set to BAG, and AciMax maximum is BAG+Jitter.ACi is the residue credit of current VLi.Ti is the time that Frame i arrives switch, and PCNT is the time that a upper Frame arrives switch.
From above algorithm, do not abandon situation as shown in Figure 2 by the Frame of control.
Exchange scheduling can be carried out by the Frame of filtration and control.Not that Frame is carried out queue operation during Frame scheduling, but usage data frame information, data frame information is obtained by Frame and configuration information thereof.In switch, reality only stores a Frame, has saved space storage resources, and factor data frame information is relatively little, and the queue operation for data frame information has saved time resource.
Switch is dispatched based on queue, is divided into receiving scheduling and sends scheduling.Packet switching construction as shown in Figure 3.Receiving scheduling carries out based on input rank.For supporting the two-stage priority mechanism of ARINC664, for each port sets up height two input ranks.When a port has Frame to arrive, through to filter and after control, Frame is deposited to shared memory, and by Frame and configuration information acquisition data frame information, its input rank in corresponding port respective priority is joined the team.The data frame information form that input rank adopts is as shown in table 1.Receiving scheduling adopts the fair poll of RR between the ports, first high rear low priority scheduling algorithm in port, read data frame information from input rank fast, obtain target port address according to data frame information, then data frame information is put in the output queue of the corresponding priority of corresponding output port.This ensure that high preferential Frame is always scheduled away prior to the lower-priority data frame of same port, ensure that the grade of service of two-stage, input rank data frame information is see table 1;
Table 1
Send scheduling to carry out based on output queue, because the Frame in output queue is all forwarded to the same forwarding port of its correspondence, doing so avoids and dispatch the Frame of the queue heads caused because when forwarding that port is busy cannot be forwarded, cause the blockage problem that other port data frame follow-up also cannot forward based on input rank.
Send scheduling employing static priority scheduling mechanism, realize forwarding of packets fair fast between the ports, carry out differentiation in port inside according to two-stage priority and forward.When output queue be dispatched to some Frames to go out group time, this Frame can be found in shared memory according to its data frame information, this Frame is sent from corresponding port.The data frame information of output queue is as shown in table 2.
Table 2
Byte | Title | Describe |
2 | Position | Data are position in the buffer |
2 | Length | Data frame length |
1 | OutPort | Data retransmission port |
1 | CFG | The instruction of configuration frame |
Send scheduling forwarding data frame speed and be subject to the restriction of port transmission rate, if when the mass data frame of multiple port is forwarded to same port simultaneously, the situation that port output queue is full may be occurred.Because input rank schedule speed does not limit by link-speeds, therefore when the Frame output queue that receiving scheduling will forward is full, the discardable process of this Frame now in input rank.
Claims (6)
1. based on a dispatching method for storage forwarding switching fabric, it is characterized in that: said method comprising the steps of:
1) Frame filters through switch, for not abandoned by the Frame filtered, by the Frame filtered, carrying out traffic shaping inspection, carry out step 2);
2) each port all sets up two input ranks and two output queues;
3) receiving scheduling: adopt the dispatching algorithm that in the fair poll of RR between port, port, high-priority queue is preferential, read data frame information from input rank fast, obtains target port address according to data frame information;
4) data frame information is put in the output queue of the corresponding priority of corresponding output port;
5) scheduling is sent: the grouping being kept at output port queue is dispatched; Realize fair forwarding of packets between the ports, forward according to two-stage priority mechanism in port inside.
2. the dispatching method based on storage forwarding switching fabric according to claim 1, is characterized in that: in described step 3), data frame information comprises input rank data frame information and output queue data frame information; Input rank data frame information comprises Frame, and position, data frame length, Frame forward port, Mulitcast Queue distribution and the instruction of configuration frame in the buffer; Output queue data frame information comprises Frame, and position, data frame length, Frame forward port and the instruction of configuration frame in the buffer.
3. the dispatching method based on storage forwarding switching fabric according to claim 2, it is characterized in that: described input rank is different according to Frame priority, be divided into height two priority, the high-priority data frame information of this port is only deposited in the input rank of high priority, and the input rank of low priority only exists the lower-priority data frame information of this port.
4. the dispatching method based on storage forwarding switching fabric according to claim 3, is characterized in that: each two output queues of the corresponding high and low priority of each port of described output queue, output queue only exists the data frame information from this port repeat.
5. the dispatching method based on storage forwarding switching fabric according to claim 4, is characterized in that: described transmission scheduling adopts " static priority " scheduling strategy to dispatch separate output work queue's system.
6. the dispatching method based on storage forwarding switching fabric according to claim 5, is characterized in that: described filtration is specifically for the filtration of minimax frame length, VL configuration, Frame completeness check aspect.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310688711.8A CN104717159A (en) | 2013-12-14 | 2013-12-14 | Scheduling method based on store-and-forward switching structure |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310688711.8A CN104717159A (en) | 2013-12-14 | 2013-12-14 | Scheduling method based on store-and-forward switching structure |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104717159A true CN104717159A (en) | 2015-06-17 |
Family
ID=53416133
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310688711.8A Pending CN104717159A (en) | 2013-12-14 | 2013-12-14 | Scheduling method based on store-and-forward switching structure |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104717159A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105978824A (en) * | 2016-07-18 | 2016-09-28 | 醴陵恒茂电子科技有限公司 | Frame queue scheduling traffic shaping method and system |
CN107248963A (en) * | 2017-05-19 | 2017-10-13 | 闫晓峰 | Multichannel interchanger, passage referee method and the means of communication based on Ethernet |
CN110419204A (en) * | 2017-03-22 | 2019-11-05 | 三菱电机株式会社 | Transmit method, entity and the program of signal of communication frame |
CN110995535A (en) * | 2019-11-28 | 2020-04-10 | 中国航空工业集团公司西安航空计算技术研究所 | Airborne network communication monitoring method |
CN112073230A (en) * | 2020-08-27 | 2020-12-11 | 中国航空无线电电子研究所 | Simplified network system compatible with standard AFDX (avionics full Duplex switched Ethernet) network |
CN112751788A (en) * | 2020-12-30 | 2021-05-04 | 西安云维智联科技有限公司 | Double-plane switching method supporting multi-type frame mixed transmission |
CN113259006A (en) * | 2021-07-14 | 2021-08-13 | 北京国科天迅科技有限公司 | Optical fiber network communication system, method and device |
CN113595916A (en) * | 2020-04-30 | 2021-11-02 | 瑞昱半导体股份有限公司 | Circuit located in router or switch and frame processing method applied to router or switch |
WO2022142510A1 (en) * | 2020-12-30 | 2022-07-07 | 北京国科天迅科技有限公司 | Fc switch scheduling method and apparatus, and electronic device and storage medium |
CN116032859A (en) * | 2023-02-16 | 2023-04-28 | 之江实验室 | Fusion type rapid data exchange device and method |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020118689A1 (en) * | 2000-09-27 | 2002-08-29 | Luijten Ronald P. | Switching arrangement and method with separated output buffers |
CN101557346A (en) * | 2009-05-14 | 2009-10-14 | 中国人民解放军信息工程大学 | Polling-based packet queue output scheduling method and packet switching node |
CN101729407A (en) * | 2009-12-04 | 2010-06-09 | 西安电子科技大学 | Low delay jitter exchanging method and equipment based on unicast and multicast differentiated treatment |
CN102427426A (en) * | 2011-12-05 | 2012-04-25 | 西安电子科技大学 | Method and device for simultaneously supporting AFDX (Avionics Full-duplex Switched Ethernet) and common Ethernet switching |
CN102497310A (en) * | 2011-12-05 | 2012-06-13 | 西安电子科技大学 | Service flow shaping method and device for cascaded port of avionics full duplex switched Ethernet (AFDX) switch |
-
2013
- 2013-12-14 CN CN201310688711.8A patent/CN104717159A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020118689A1 (en) * | 2000-09-27 | 2002-08-29 | Luijten Ronald P. | Switching arrangement and method with separated output buffers |
CN101557346A (en) * | 2009-05-14 | 2009-10-14 | 中国人民解放军信息工程大学 | Polling-based packet queue output scheduling method and packet switching node |
CN101729407A (en) * | 2009-12-04 | 2010-06-09 | 西安电子科技大学 | Low delay jitter exchanging method and equipment based on unicast and multicast differentiated treatment |
CN102427426A (en) * | 2011-12-05 | 2012-04-25 | 西安电子科技大学 | Method and device for simultaneously supporting AFDX (Avionics Full-duplex Switched Ethernet) and common Ethernet switching |
CN102497310A (en) * | 2011-12-05 | 2012-06-13 | 西安电子科技大学 | Service flow shaping method and device for cascaded port of avionics full duplex switched Ethernet (AFDX) switch |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105978824A (en) * | 2016-07-18 | 2016-09-28 | 醴陵恒茂电子科技有限公司 | Frame queue scheduling traffic shaping method and system |
CN110419204A (en) * | 2017-03-22 | 2019-11-05 | 三菱电机株式会社 | Transmit method, entity and the program of signal of communication frame |
CN110419204B (en) * | 2017-03-22 | 2022-02-18 | 三菱电机株式会社 | Method, entity and program for transmitting communication signal frame |
CN107248963A (en) * | 2017-05-19 | 2017-10-13 | 闫晓峰 | Multichannel interchanger, passage referee method and the means of communication based on Ethernet |
CN107248963B (en) * | 2017-05-19 | 2020-05-22 | 闫晓峰 | Multi-channel switch based on Ethernet, channel arbitration method and communication method |
US11398974B2 (en) | 2017-05-19 | 2022-07-26 | Beijing NE-net Technology Co. Ltd. | Ethernet-based multi-channels switch, channel arbitration method and communication method thereof |
CN110995535A (en) * | 2019-11-28 | 2020-04-10 | 中国航空工业集团公司西安航空计算技术研究所 | Airborne network communication monitoring method |
CN110995535B (en) * | 2019-11-28 | 2022-04-05 | 中国航空工业集团公司西安航空计算技术研究所 | Airborne network communication monitoring method |
CN113595916A (en) * | 2020-04-30 | 2021-11-02 | 瑞昱半导体股份有限公司 | Circuit located in router or switch and frame processing method applied to router or switch |
CN112073230A (en) * | 2020-08-27 | 2020-12-11 | 中国航空无线电电子研究所 | Simplified network system compatible with standard AFDX (avionics full Duplex switched Ethernet) network |
CN112751788A (en) * | 2020-12-30 | 2021-05-04 | 西安云维智联科技有限公司 | Double-plane switching method supporting multi-type frame mixed transmission |
WO2022142510A1 (en) * | 2020-12-30 | 2022-07-07 | 北京国科天迅科技有限公司 | Fc switch scheduling method and apparatus, and electronic device and storage medium |
CN112751788B (en) * | 2020-12-30 | 2023-11-14 | 西安云维智联科技有限公司 | Double-plane switching method supporting multi-type frame mixed transmission |
CN113259006B (en) * | 2021-07-14 | 2021-11-26 | 北京国科天迅科技有限公司 | Optical fiber network communication system, method and device |
CN113259006A (en) * | 2021-07-14 | 2021-08-13 | 北京国科天迅科技有限公司 | Optical fiber network communication system, method and device |
CN116032859A (en) * | 2023-02-16 | 2023-04-28 | 之江实验室 | Fusion type rapid data exchange device and method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104717159A (en) | Scheduling method based on store-and-forward switching structure | |
US9083655B2 (en) | Internal cut-through for distributed switches | |
US8451742B2 (en) | Apparatus and method for controlling data communication | |
US7855960B2 (en) | Traffic shaping method and device | |
JP4616535B2 (en) | Network switching method using packet scheduling | |
US9608927B2 (en) | Packet exchanging device, transmission apparatus, and packet scheduling method | |
WO2019157978A1 (en) | Method for scheduling packet, first network device, and computer readable storage medium | |
TWI227080B (en) | Network switch providing congestion control and method thereof | |
CN100550852C (en) | A kind of method and device thereof of realizing mass port backpressure | |
EP3468115B1 (en) | Method to improve availabilty of real-time computer networks | |
WO2012145841A1 (en) | Hierarchical profiled scheduling and shaping | |
JP2009022038A5 (en) | ||
WO2011100878A1 (en) | Method, switching device and system for realizing flow controlling in a switching network | |
CN101692648A (en) | Method and system for queue scheduling | |
CN109842564A (en) | A kind of method, the network equipment and system that service message is sent | |
CN101478486B (en) | Method, equipment and system for switch network data scheduling | |
CN101212412A (en) | MPLS network service scheduling system | |
US8879578B2 (en) | Reducing store and forward delay in distributed systems | |
US7990873B2 (en) | Traffic shaping via internal loopback | |
US20210250301A1 (en) | Scalable deterministic services in packet networks | |
CN101771603A (en) | Data exchange method and data exchange equipment | |
CN114338544B (en) | Flow control method of FC switching network based on priority | |
CN113810305B (en) | Message forwarding method, device, forwarding node and computer readable storage medium | |
CN101277546B (en) | RPU control protocol for light synthetic switch OIS network application supporting WiNAX multi-service transmission | |
WO2023174081A1 (en) | Queue scheduling method and apparatus |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150617 |