CN1564550A - Access method of short packet data - Google Patents
Access method of short packet data Download PDFInfo
- Publication number
- CN1564550A CN1564550A CNA2004100338480A CN200410033848A CN1564550A CN 1564550 A CN1564550 A CN 1564550A CN A2004100338480 A CNA2004100338480 A CN A2004100338480A CN 200410033848 A CN200410033848 A CN 200410033848A CN 1564550 A CN1564550 A CN 1564550A
- Authority
- CN
- China
- Prior art keywords
- data
- short packet
- packet data
- port
- descriptor
- 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.)
- Granted
Links
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
Network processor and ATM access is adopted in the method. In receiving direction, bus conversion is carried out for short packet data arrived at main optical port by using FPGA to complete CRC check function. Data on standby optical port is merged to main optical port. Label of optical port number is printed on data of short packet. Network processor scans main optical ports to receive data. Based on optical port number, data of main port or standby optical port are distinguished. Thus data are separated and processed. Short packet data descriptor for packet length less than 48 bytes is added to sending queue. Transmitting logic optical port number is obtained through parsing, and mapped to physical sending port. In sending procedure, actual physical optical port is pointed. The method possesses favorable ratio of performance and reduces complexity of implementing protocol.
Description
Technical field
The present invention relates to a kind of high speed cut-in method that utilizes network processing unit to realize data, relate in particular to moving communicating field the ATM of short packet data is inserted at a high speed.
Background technology
For the data access technology, can consider the realization situation of its performance and function from two aspects, the one, insert the access disposal ability of processor at present, the 2nd, the access performance of data link.
Consider from inserting processor connection whether primary processor has disposal ability at a high speed also is a key that realizes that really data high-speed is handled.Mainly contain two kinds of general processor and network processing units at present, the function difference that both realized, performance also differs greatly.Computer network is normal to adopt high performance server to solve the data processing of big flow, yet studies show that for two CPU high-performance servers of PIV2.0G, because its read/write memory bottleneck of performance, the network processes performance also has only about 400Mbps.At communication field, at present high speeds that adopt network processing unit to realize data insert more, and the processor maximum of this network data special use can reach the disposal ability of G bit, can satisfy present data high-speed access demand.
Yet, network processing unit need be done a large amount of Business Processing in actual applications, and for the consideration of cost, the performance of network processing unit may be restricted, with the IXP1200 of Intel network processing unit is example, and one has 6 micro engines, and wherein 4 need are used for the post-processed after data insert, can only provide 2 micro engines for the reception and the transmission processing of data, will be that 4 light mouths receive transmission at two groups of active and standby light mouths simultaneously.Experiment showed, that the total flow that existing access processing method all can't reach design is the requirement of 310Mbps.
From the access way of data link, computer and the common access way of communication field have Ethernet access, E1 access, ATM (Asynchronous Transfer Mode) to insert or the like at present.The complexity of bottom physics realization form, performance and the realization of these access waies is had nothing in common with each other.
Ethernet inserts to be realized simply, also common, it is one of main mode of present data access, yet, this data access adapts to the access of long bag data, for short packet data, the waste of its bandwidth and transmission performances all can't reach requirement, especially when data length less than 64 bytes time performance particularly evident.Test shows, when data packet length was 64 bytes, design bandwidth was that the interface testing bandwidth of 100,000,000 performances has only about 60,000,000, and when carry in the data message when the speech data, have a large amount of Ethernet overhead bytes to be wasted in the 64 byte messages.The E1 access way is the main access way of present base station speech data access base station controller, but because the restriction of its intrinsic bandwidth inserts its high speed that can't realize data.ATM access way bottom adopts Optical Fiber Transmission, the high speed access way of short packet data is provided, but agreement realizes comparatively complicated, on adaptation layer, need to finish functions such as segmentation/reorganization, message identification, sequence error detection, greatly improved complexity, construction cycle and the fee to develop of engineering.
Summary of the invention
The technical problem to be solved in the present invention provides a kind of cut-in method of short packet data, can make full use of the network processing unit Limited resources, realizes that the high speed of short packet data inserts.
In order to solve the problems of the technologies described above, the invention provides a kind of cut-in method of short packet data, adopt ATM to insert, comprise the handling process that receives data and send data, it is characterized in that: the handling process of described reception data may further comprise the steps:
(a) will arrive the short packet data of standby light mouth and become owner of and use up mouth, and on all short packet datas, stamp light slogan label, send to network processing unit again;
(b) network processing unit detects the request signal of main port, judges whether port has short packet data, if having, carries out next step, otherwise continues to detect;
(c) send the Data Receiving request word, receive short packet data;
(d) extract the port numbers of packet and carry out that port is peeled off and validity check, verification succeeds is carried out next step, otherwise abandons this packet;
(e) send the reception instruction, short packet data is stored in the internal memory from receive FIFO, and the descriptor of structure short packet data, write the reception formation.
After adopting the inventive method, make the receiving thread of network processor micro-engine when detection port, only see main port, rather than the actual active and standby port of supporting of using, can reduce the number of times of thread inquiry bus failure widely, thereby improve performance.
Further, can adopt the handling process of following transmission data:
(o) obtain bag length to be sent descriptor from transmit queue less than the short packet data of 48 bytes;
(p) whether extraction and judgement bag transmit port and data packet length be effective, if carry out next step, otherwise abandon;
(q) select to send FIFO, carry out port arbitration and structure and send control word, after transmit port is ready, carries out data and send.
Said method also can add CRC check, and promptly described step (a) also comprises step before: carry out CRC check, obtain assay; Simultaneously, in described step (e) before, also comprise step: described network processing unit is resolved the assay that obtains, and for the execution in step (c) again of the short packet data by verification, otherwise abandons.
Said method also can have following characteristics: described step (a) also comprises step before: the short packet data that arrives active and standby light mouth is carried out the bus conversion.
Said method also can have following characteristics: described step (b) is handled short packet data by programmable gate array before.
Said method also can have following characteristics: described step (q) also comprises step afterwards: according to the message length in the described descriptor, carry out CRC check interpolation and transmission.
Said method also can have following characteristics: in the descriptor of described short packet data, comprised the buffering area base address that packet is deposited in internal memory, the side-play amount of packet effective byte in core buffer, the logic receiving port of packet or logic transmit port, fields such as length of data package.
Said method also can have following characteristics: also comprised the reserved field that is used to expand use in the descriptor of described short packet data.
Said method also can have following characteristics: when carrying out the port arbitration in the described step (q), obtain the logic light slogan that short packet data sends by analysis to the data descriptor, draw corresponding physics transmit port by mapping, in process of transmitting, specify real physics light mouth again.Compare with the form of passing through scheduling repeating query transmit port data of common employing, saved scheduling thread.
As from the foregoing, the present invention has adopted network processing unit, by active and standby logic port folding, can utilize less network processing unit resource to realize that the high speed of short packet data inserts, and guarantees the subsequent treatment ability that data insert simultaneously, possesses good cost performance.And the ATM access way of the no adaptation layer of employing has greatly been saved bandwidth, has reduced complexity and engineering cost that agreement realizes.Therefore, the present invention provides a kind of reliable, simple high-performance cut-in method for short packet data inserts.Be specially adapted to the access of speech data.
Description of drawings
Fig. 1 is the structural representation of embodiment of the invention system function module.
Fig. 2 is the descriptor of the memory of embodiment of the invention method short packet data.
Fig. 3 is the process chart that the embodiment of the invention receives short packet data.
Fig. 4 is the process chart that the embodiment of the invention sends short packet data.
Embodiment
The short packet data of present embodiment is meant the long packet less than 48 bytes of bag, because specified data (by User Defined) length is less than 48 bytes, thereby can be carried on one above the ATM cell, there is no need to carry out adaption functions such as burst reorganization, sequence error detection.Therefore, present embodiment has adopted the ATM access way of no adaptation layer (being the AAL0 mode) at special applied environment, can savings outer expense, saved bandwidth, and reduced complexity and engineering cost that agreement realizes.
As shown in Figure 1, network processing unit adds FPGA (Field Programmable Gate Array, programmable gate array) 10 and STM phy chip 20 is hardware foundations of finishing embodiment of the invention short packet data high speed cut-in method.The STM phy chip 20 of present embodiment adopts the CX29704 chip, is used for realizing the function of data physical layer and data link layer, but also can adopts the phy chip of any a support atm standard.Network processing unit adopts the aforesaid IXP1200 of Intel network processing unit, utilizes its distinctive IXBUS bus and transmitter-receiver system to finish the Business Processing of payload behind the access of short packet data and the packet header, assists to finish the CRC check function simultaneously.And FPGA is mainly used in the conversion that realizes utopia bus and IXBUS bus, the generation of 16 CRC and portion C RC verification, and standby light mouth data are incorporated into the key light mouth and add glazing slogan label to distinguish function such as light mouth data channel.From sending and receive the function of both direction, network processing unit adds FPGA10 can be divided into following functional module further:
Short packet data transmit queue 11 is used to receive network processing unit interior business processing module and finishes after the data processing descriptor of the short packet data to be sent that writes;
Send CRC generation module 13, be used for reading the message length that the short packet data descriptor is provided with, generate the CRC check field, present embodiment utilizes fpga logic to finish;
Receive CRC check module 15, be used to finish the CRC check function that receives short packet data (being ATM cell), in the present embodiment, this verifying function is to unite realization by FPGA and network processes, FPGA generates check results, is carried out the differentiation and the function corresponding operation of check results by network processing unit;
Short packet data receives formation 17, the ATM receiver module is after the reception of finishing data, construct corresponding data descriptor, write this and receive formation, read and handle by Service Processing Module, at application, the user can resolve data carried by data according to the demand of self, even adopts other agreement.
The short packet data descriptor that above-mentioned short packet data sends and receives in the formation is to be used for describing short packet data at the position of internal memory and the attribute of packet, its concrete structure as shown in Figure 2, continuous 16 byte spaces of each descriptor committed memory are by the reserved field of length and 8 bytes of the side-play amount of the port numbers of the descriptor base of 4 bytes, 1 byte, 1 byte, two bytes.Wherein, descriptor base is the buffering area base address that packet is deposited in internal memory; Side-play amount is the side-play amount of packet effective byte in core buffer, and port numbers is the receiving port of packet for receive direction, is the transmit port of packet for sending direction, is logic port; The length of length field store data bag can be used for generating the CRC check field; Reserved field can be used as in the future the expansion of consummating function and use, as service quality (Quality of Service, Qos) field etc.
Based on above system, the reception flow process of embodiment of the invention short packet data (processing procedure that does not comprise link layer and physical layer) as shown in Figure 3, may further comprise the steps:
Step 301, FPGA carries out the bus conversion to the short packet data that arrives active and standby light mouth;
Step 302 is carried out CRC check, obtains check results, simultaneously standby light mouth data is incorporated into the key light mouth, and stamps light slogan label on all short packet datas, sends to network processing unit;
Step 303, network processing unit detects the request signal of two main ports (using up mouth corresponding to two masters), promptly carry out the IXBUS bus port READY input of IXP1200, when the packet literary composition was arranged on the bus, the READY signal will be put the micro engine thread of corresponding READY position notice IXP1200;
Step 304 is judged whether port has data message, if having, carries out next step, otherwise is returned step 303, and micro engine continues testbus;
Step 305 sends the Data Receiving request word, and request word information comprises information such as receiving FIFO, receiving thread and receives the data use for IXP1200 accepting state machine;
Step 306, receiving data packets promptly after sending the request of reception, waits pending data to arrive and receives among the FIFO;
Step 307, extract port numbers and judge whether effective, because present embodiment receives the repeating query of processing of request process and detects two ports, therefore to carry out peeling off and the port validation verification of port here, if whether port numbers is effective, carry out next step, directly abandon and return step 303, restart message and receive for the data message of port numbers mistake;
Step 308, the result resolves to data message CRC check, if verification is passed through, carries out next step, otherwise dropping packets returns step 303; Present embodiment adopts the mode of additional two accepting state bytes behind valid data to obtain the CRC check result at receiving FIFO;
Step 309 for the short packet data by verification, sends and receives instruction, and packet is stored in the internal memory from receive FIFO;
Step 310, the data descriptor of structure short packet data, and it is write the reception formation, descriptor adopts the form of Fig. 2.
Above-mentioned flow process be it should be noted that especially, because present embodiment utilizes FPGA to do standby light mouth data are incorporated into the special processing of key light mouth, make the receiving thread of network processor micro-engine when detection port, only see two ports, rather than actual four ports supporting, this implementation can reduce the number of times of thread inquiry bus failure widely, thereby has improved performance.Test shows that for four light mouths, this implementation can well reach the requirement of a pair of arbitrarily active and standby light mouth design total flow 155Mbps, and existing other modes all can't realize at present.
The transmission flow of embodiment of the invention short packet data may further comprise the steps as shown in Figure 4:
Step 401 is obtained the descriptor of packet to be sent from the short packet data transmit queue, and this formation can be the interface queue of receiver module, also can be the interface queue of operational blocks which partition system;
Step 402 is extracted bag transmit port and data packet length, owing to any shared identical transmit queue of a pair of active and standby light mouth, therefore must confirm the actual port number of packet before transmission, simultaneously length of data package is done necessary inspection;
Step 403 judges whether packet length and port numbers be effective, if, carry out next step, otherwise detect failure, abandon corresponding data message, return step 401;
Step 404 is selected to send FIFO, carries out port arbitration and structure and sends control word.At sending direction, distribute according to two light mouths for the distribution that sends FIFO; At the ATM sending direction, network processing unit obtains the logic light slogan that short packet data sends in to data descriptor analytic process, draw corresponding physics transmit port by mapping, in process of transmitting, specify real physics light mouth, rather than the form of passing through scheduling repeating query transmit port data of common at present design employing, save scheduling thread, improved the transmission performance; For sending control word,, be SOP (Start ofPacket) and EOP (End of Packet) message simultaneously therefore because each short packet data less than 64 bytes, is exactly a complete packet during employing IXP1200 processor transmit status mechanism;
Step 405 judges whether transmit port is ready, if port is offhand ready, and then circular wait, otherwise carry out next step;
Step 406 is put the transmission significance bit, and the micro engine thread carries out data and sends;
Step 407, FPGA carries out CRC check interpolation and transmission according to the message length in the descriptor.
As can be seen, the foregoing description is at the ATM receive direction, the short packet data (total flow mostly is 155MBps most) that arrives active and standby light mouth is after carrying out the bus conversion through fpga logic, fpga logic will be finished the CRC check function and standby light mouth data are incorporated into the function of key light mouth simultaneously, and stamps light slogan label send to the IXP1200 network processing unit on all short packet datas.Network processing unit based on performance demands, just scans reception at the main mouth of using up when receiving data, distinguish the active and standby mouthful data of using up according to the light slogan of data afterwards and carry out data separating, receives processing then accordingly.
This processing mode can be used less micro engine to support to insert and handle, and realizes that short packet data inserts at a high speed, thereby the micro engine of more free time is provided for Business Processing.With regard to the IXP1200 network processing unit, be used for inserting processing with two micro engines, adopt said method that two light mouths performance of 310Mbps flow altogether can be provided, when supporting that four active and standby light mouths are worked simultaneously, at four ports data arrival and any two active and standby port flows being arranged all is under the designing requirement of 155Mbps, can reach any two flows that active and standby port flow is 210Mbps, has reached extraordinary performance index, resource utilization is very high, is a kind of processing method of high performance-price ratio.
On the basis of the foregoing description, the present invention can also have various conversion, for example, can select to support or not support the CRC check processing, only needs select by grand compiler toggle in reception and transmission flow.The form of adaptation layer also not necessarily is confined to AAL0.In addition, the function of FPGA also can be acted on behalf of with other Logical processing unit.
Claims (9)
1, a kind of cut-in method of short packet data adopts ATM to insert, and comprises the handling process that receives data and send data, and it is characterized in that: the handling process of described reception data may further comprise the steps:
(a) will arrive the short packet data of standby light mouth and become owner of and use up mouth, and on all short packet datas, stamp light slogan label, send to network processing unit again;
(b) network processing unit detects the request signal of main port, judges whether port has short packet data, if having, carries out next step, otherwise continues to detect;
(c) send the Data Receiving request word, receive short packet data;
(d) extract the port numbers of packet and carry out that port is peeled off and validity check, verification succeeds is carried out next step, otherwise abandons this packet;
(e) send the reception instruction, short packet data is stored in the internal memory from receive FIFO, and the descriptor of structure short packet data, write the reception formation.
2, the cut-in method of short packet data as claimed in claim 1 is characterized in that, the handling process of described transmission data may further comprise the steps:
(o) obtain bag length to be sent descriptor from transmit queue less than the short packet data of 48 bytes;
(p) whether extraction and judgement bag transmit port and data packet length be effective, if carry out next step, otherwise abandon;
(q) select to send FIFO, carry out port arbitration and structure and send control word, after transmit port is ready, carries out data and send.
3, the cut-in method of short packet data as claimed in claim 1 is characterized in that, described step (a) also comprises step before: carry out CRC check, obtain assay; Simultaneously, in described step (e) before, also comprise step: described network processing unit is resolved the assay that obtains, and for the execution in step (c) again of the short packet data by verification, otherwise abandons.
4, the cut-in method of short packet data as claimed in claim 1 is characterized in that, described step (a) also comprises step before: the short packet data that arrives active and standby light mouth is carried out the bus conversion.
As the cut-in method of claim 1,2,3 or 4 described short packet datas, it is characterized in that 5, described step (b) is handled short packet data by programmable gate array before.
6, the cut-in method of short packet data as claimed in claim 2 is characterized in that, described step (q) also comprises step afterwards: according to the message length in the described descriptor, carry out CRC check interpolation and transmission.
7, the cut-in method of short packet data as claimed in claim 1 or 2, it is characterized in that, in the descriptor of described short packet data, comprised the buffering area base address that packet is deposited in internal memory, the side-play amount of packet effective byte in core buffer, the logic receiving port of packet or logic transmit port, fields such as length of data package.
8, the cut-in method of short packet data as claimed in claim 1 or 2 is characterized in that, has also comprised the reserved field that is used to expand use in the descriptor of described short packet data.
9, the cut-in method of short packet data as claimed in claim 7, it is characterized in that, when carrying out the port arbitration in the described step (q), obtain the logic light slogan that short packet data sends by analysis to the data descriptor, draw corresponding physics transmit port by mapping, in process of transmitting, specify real physics light mouth again.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100338480A CN1306772C (en) | 2004-04-19 | 2004-04-19 | Access method of short packet data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100338480A CN1306772C (en) | 2004-04-19 | 2004-04-19 | Access method of short packet data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1564550A true CN1564550A (en) | 2005-01-12 |
| CN1306772C CN1306772C (en) | 2007-03-21 |
Family
ID=34481406
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004100338480A Expired - Fee Related CN1306772C (en) | 2004-04-19 | 2004-04-19 | Access method of short packet data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1306772C (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100420235C (en) * | 2005-03-21 | 2008-09-17 | 北京北方烽火科技有限公司 | Apparatus and method for high-speed recombination of IP wafer pack in SGSN |
| CN101568048A (en) * | 2009-05-15 | 2009-10-28 | 中兴通讯股份有限公司 | Data transmission method and system |
| CN101136852B (en) * | 2007-06-01 | 2010-05-19 | 武汉虹旭信息技术有限责任公司 | Deep pack processing method of microengine |
| CN101405699B (en) * | 2006-03-20 | 2011-11-16 | 英特尔公司 | Validating data using processor instructions |
| WO2015066849A1 (en) * | 2013-11-06 | 2015-05-14 | 华为终端有限公司 | Method and system for reducing power consumption, and modem |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1192575C (en) * | 2002-02-10 | 2005-03-09 | 华为技术有限公司 | Method of realizing quick inserting ethernet load point to point protocol using network processor |
| US20030217182A1 (en) * | 2002-05-15 | 2003-11-20 | Xiaodong Liu | Interface architecture |
| US7248594B2 (en) * | 2002-06-14 | 2007-07-24 | Intel Corporation | Efficient multi-threaded multi-processor scheduling implementation |
| US20040006633A1 (en) * | 2002-07-03 | 2004-01-08 | Intel Corporation | High-speed multi-processor, multi-thread queue implementation |
-
2004
- 2004-04-19 CN CNB2004100338480A patent/CN1306772C/en not_active Expired - Fee Related
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100420235C (en) * | 2005-03-21 | 2008-09-17 | 北京北方烽火科技有限公司 | Apparatus and method for high-speed recombination of IP wafer pack in SGSN |
| CN101405699B (en) * | 2006-03-20 | 2011-11-16 | 英特尔公司 | Validating data using processor instructions |
| CN101136852B (en) * | 2007-06-01 | 2010-05-19 | 武汉虹旭信息技术有限责任公司 | Deep pack processing method of microengine |
| CN101568048A (en) * | 2009-05-15 | 2009-10-28 | 中兴通讯股份有限公司 | Data transmission method and system |
| WO2015066849A1 (en) * | 2013-11-06 | 2015-05-14 | 华为终端有限公司 | Method and system for reducing power consumption, and modem |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1306772C (en) | 2007-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11916781B2 (en) | System and method for facilitating efficient utilization of an output buffer in a network interface controller (NIC) | |
| CN1151639C (en) | Networking systems | |
| US20080232387A1 (en) | Electronic Device and Method of Communication Resource Allocation | |
| CN103069757A (en) | Packet reassembly and resequence method, apparatus and system | |
| CN1282339C (en) | A data frame buffer memory device and method for Ethernet passive optical network | |
| CN1975680A (en) | Method for realizing interprocess asynchronous communication based on Java | |
| CN1264531A (en) | Switching fabric arrangement with time stamp function | |
| CN1268097C (en) | Method for realizing message partition packet | |
| CN1802836A (en) | Network protocol off-load engine memory management | |
| CN1567908A (en) | Converter/bridge and method of operation between Ethernet and atm interfaces | |
| CN1798094A (en) | Method of using buffer area | |
| CN1166130C (en) | Method and system for transmitting asynchronous transfer mode unit in asynchronous transfor mode chain road | |
| CN1505888A (en) | Selective admission control in a network device | |
| AU2004310639B2 (en) | Using ordered locking mechanisms to maintain sequences of items such as packets | |
| CN102891809B (en) | Multi-core network device message presses interface order-preserving method and system | |
| CN1681249A (en) | Asynchronous data segmentation/transmission method for synchronous Ethernet and data structure used therein | |
| CN102082698A (en) | Network data processing system of high performance core based on improved zero-copy technology | |
| US9274586B2 (en) | Intelligent memory interface | |
| CN1145329C (en) | Packet switching equipment and switching control method | |
| CN1826768A (en) | A scalable approach to large scale queuing through dynamic resource allocation | |
| CN1949695A (en) | Method and system for giving-up error frame in frame data transmission | |
| CN1791069A (en) | Input queue packet switch architecture and queue service discipline | |
| CN1306772C (en) | Access method of short packet data | |
| CN101848239A (en) | High-timeliness distributed service integration calling system | |
| CN114363269A (en) | Message transmission method, system, equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20070321 Termination date: 20140419 |