CN1595910A - A data packet receiving interface component of network processor and storage management method thereof - Google Patents
A data packet receiving interface component of network processor and storage management method thereof Download PDFInfo
- Publication number
- CN1595910A CN1595910A CNA2004100500047A CN200410050004A CN1595910A CN 1595910 A CN1595910 A CN 1595910A CN A2004100500047 A CNA2004100500047 A CN A2004100500047A CN 200410050004 A CN200410050004 A CN 200410050004A CN 1595910 A CN1595910 A CN 1595910A
- Authority
- CN
- China
- Prior art keywords
- data
- storage
- pointer
- data packet
- dram
- 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
- 238000003860 storage Methods 0.000 title claims abstract description 86
- 238000007726 management method Methods 0.000 title claims abstract description 46
- 238000012545 processing Methods 0.000 claims abstract description 33
- 238000013500 data storage Methods 0.000 claims abstract description 24
- 238000000034 method Methods 0.000 claims abstract description 19
- 230000015654 memory Effects 0.000 claims description 91
- 238000009826 distribution Methods 0.000 claims description 10
- 230000008520 organization Effects 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 5
- 230000003139 buffering effect Effects 0.000 claims description 4
- 238000012423 maintenance Methods 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 2
- 230000007246 mechanism Effects 0.000 claims description 2
- 238000012986 modification Methods 0.000 claims description 2
- 230000004048 modification Effects 0.000 claims description 2
- 230000005540 biological transmission Effects 0.000 abstract description 17
- 238000004891 communication Methods 0.000 abstract description 2
- 230000003068 static effect Effects 0.000 abstract 1
- 230000015572 biosynthetic process Effects 0.000 description 16
- 238000005755 formation reaction Methods 0.000 description 16
- 238000012546 transfer Methods 0.000 description 10
- 230000008859 change Effects 0.000 description 6
- 238000013461 design Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 3
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000005304 joining Methods 0.000 description 1
- 238000011017 operating method Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
本发明涉及数据通信技术领域。特别是一种用于网络处理器的数据包接收接口部件及其存储管理方法。部件包括：数据接收缓冲装置；指针存储区管理装置；DRAM(动态随机存取存储器)存储控制器；SRAM(静态随机存取存储器)存储控制器；队列管理装置；数据存储区。方法包括：使用队列表、数据包指针和存储块指针对数据存储区进行有效的组织管理；利用存储块指针和存储块位置对齐节省存储空间并提高操作效率；使用SRAM和DRAM分别存储数据包头和净荷数据来提高处理数据传输速度。本发明还通过对DRAM存储控制器进行改进，进一步提高了数据的存取速度，有助于高速网络处理器克服其存储瓶颈，实现高速数据传输和处理。
The invention relates to the technical field of data communication. In particular, a data packet receiving interface component for a network processor and a storage management method thereof. The components include: data receiving buffer device; pointer storage area management device; DRAM (dynamic random access memory) storage controller; SRAM (static random access memory) storage controller; queue management device; data storage area. The method includes: using the queue table, the data packet pointer and the storage block pointer to effectively organize and manage the data storage area; using the storage block pointer and the storage block position alignment to save storage space and improve operation efficiency; using SRAM and DRAM to store the data packet header and Payload data to improve processing data transmission speed. The invention further improves the data access speed by improving the DRAM storage controller, which helps the high-speed network processor to overcome its storage bottleneck and realize high-speed data transmission and processing.
Description
Technical field
The present invention relates to data communication technology field.Particularly a kind of packet receiving interface parts and memory management method thereof that is used for network processing unit.
Background technology
Along with rapid development of network technique, the network bandwidth has risen to present 40Gbps from several years ago 2Gbps, this needs switch and router that the faster data disposal ability is provided, in addition, be the procotol that adapts to continuous variation and the requirement of network service quality (QOS), also require the network switching equipment to possess extensibility and programmability more flexibly, and traditional GPP (general processor) and ASIC (application-specific integrated circuit (ASIC)) all can not satisfy the requirement of this two aspect simultaneously.Therefore, a kind of new-type network processing unit---network processing unit owing to have at a high speed data-handling capacity and programmability flexibly concurrently, is applied in switch and the router more and more widely.
In typical network processing unit, find by quantitative analysis, at a packet from the process that receives forwarding, nearly 2/3 time will be used for the reception of data, storage, scheduling and transmission, although in the network processing unit design, can provide data processing operation ability at a high speed as network processing unit by using a plurality of special-purpose RISC CPU (compacting instruction set processor), but the low speed transmissions ability of slow memory parts has still hindered the further raising of network processing unit performance, and storage subsystem has become the bottleneck of network processing unit.Therefore, have only by reasonably designing and receive and the transmission interface parts, improve the concurrency of Data Receiving, storage and queuing, simultaneously storage subsystem is improved, adopt rationally memory management method efficiently, improve memory transfer speed to greatest extent, just can improve the performance of network processing unit effectively.
At present, in the network processing unit design, mainly improve storage subsystem by two kinds of methods, a kind of method is to use distributed storage method, by the data of different types bag is placed on respectively in the different memories, utilize the concurrent access of memory to improve transmission speed, but this method is not improved the transmission speed of same type packet.
Another kind method is to improve storage control, utilize the more distinctive access features of DRAM memory, memory reference instruction is cushioned, predicts and reorders, postpone by hiding some read-write, improve burst (Burst) the number of transmissions, improve memory transfer speed.But the existing implementation strategy of this method need be considered concrete procotol and scheduling strategy, causes hardware prediction logic complexity and is unsuitable for disposing the network environment of frequent variation.
Summary of the invention
The object of the present invention is to provide a kind of packet receiving interface parts and memory management method thereof of network processing unit.These parts are made up of a plurality of circuit arrangements that can executed in parallel, adopt this interface unit, the packet of network processing unit receiving terminal are received, the concurrency of storage and scheduling is improved, and effectively improves the data transmission bauds of network processing unit receiving terminal.
Another object of the present invention is to realize improving the method for parallel processing of packet reception, Flow Control, storage and scheduling.
Another object of the present invention provides a kind of memory block that is used for network processing unit and organizes management method, improves the memory block is distributed and packet is lined up flexibility and speed, makes network processing unit can carry out storage area management effectively.
Another object of the present invention provides a kind of method for designing of improved DRAM storage control, and the transmission rate of DRAM memory is improved, and compares with method in the past and to have simple and better adaptability.
Description of drawings
The front has been carried out briefly bright to purpose of the present invention, below in conjunction with accompanying drawing main contents of the present invention are described, and the accompanying drawing that comprises mainly contains:
Fig. 1: be that structure chart is organized in the used memory block of the present invention.
Fig. 2: be not improved DRAM memory burst read operation sequential chart.
Fig. 3: be the used improved DRAM memory burst read operation sequential chart of the present invention.
Fig. 4: the system architecture diagram that is receiving interface parts of the present invention.
Fig. 5: the state transition graph when being interface unit of the present invention operation.
Fig. 6 is the used memory area management method flow chart of the present invention.
Embodiment
Because the organization and management method of the global design of this interface unit and memory block is closely related, below organize management method to describe to the used memory block of the present invention earlier.
In network processing unit, be convenient queuing, the memory block is generally divided into the piece of fixed size, and employing chained list way to manage, by following the tracks of the backbone network packet, and the size of the packet of being caught added up discovery, the size of about about 40% Ethernet data bag less than or near 64KB, theoretical and experience all proves, in the network switching equipment, adopt the data block of 64Byte to carry out storage administration, help reducing the memory block fragment and reduce memory accesses, so generally adopted by network processing unit.For realizing the chain type way to manage of formation, the basic thought of most memory area management method is: be provided with in each memory block inside and point to the pointer that belongs to next piece memory block of same packet, the block-chaining that this pointer is responsible for belonging to same packet arrives together.In the memory block, reserve special space simultaneously and deposit MBA memory block address pointer and various formation head and the tail pointer, be used for packet is linked into different formations.But adopt which type of specific implementation method that the access efficiency of storage area management is produced different influences with the flexibility meeting.Accompanying drawing 1 has illustrated that our used memory block organizes the memory block of management method to organize structure chart.
Organize in the management method in this memory block, one block pointer memory block is set, pointer memory block list item is made up of position field, packet pointer and memory block pointer, and position field is used for pointing out the position of data block at affiliated packet, and packet pointer is used for data packet group is made into formation.Whole storage organization is made up of three parts, queue table, pointer memory block and data storage area.Formation of in the queue table each correspondence is made up of queue head pointer territory and tail pointer territory, is used for data packet group is made into formation.The pointer in two territories is pointed out first packet and the original position of last packet in the pointer memory block in the formation respectively; In the pointer memory block each mainly comprises three territories: data block location mark domain, packet pointer territory and memory block pointer field.The position of position field indication current data block in packet mainly formed by two, and implication is as follows:
11: the first blocks of data piece in the packet.
10: the data block that is positioned at the centre position in the packet.
00: last blocks of data piece in the packet.
01: first of packet also is last blocks of data piece (illustrating that this packet only contains a blocks of data piece).
Memory block pointer and DSB data store block are alignd one by one, and the data block organisation that is used for belonging to same packet becomes chained list.
The indication of packet pointer territory belongs to the original position of next packet correspondence in the pointer memory block of same formation, is used for different packets is linked into a formation.
Comprise the division of memory block, the tissue of memory block and the method for organizing of formation.
In Fig. 1, can see, the memory block comprises a DRAM and polylith SRAM, the DRAM and the SRAM of storage data are the unit piecemeal with 64 bytes all, and each piece of DRAM all is arranged in the delegation of DRAM storage chip, the payload of packet leaves among the DRAM, and data packet head leaves among the SRAM, each deal with data that sends all is the packet header data among the SRAM, pointer memory block and queue table all are placed among the SRAM, help to accelerate memory block distribution, search, revise and releasing operation.
Each of pointer memory block and the memory block of data storage area all are to concern one to one on the position, it is the fixing DSB data store block of each unique correspondence of pointer memory block, simultaneously, in the fixing pointer memory block of the also unique correspondence of each DSB data store block one, suppose that pointer memory block list item is from 1 open numbering, the size of memory block is 64 bytes, and the initial address of whole buffering area is start_addr, and then sequence number is that the pairing DSB data store block initial address of buffer entries of N is (supposition is a unit with the byte): block_addr=start_addr+N x 64.
Adopt the mode of this aligned in position, do not need to keep pointer entry in each DSB data store block again, and belong to the correspondence position of next data block in the pointer memory block of same packet with block pointer territory indication, therefore, just the block-chaining that belongs to same packet can be arrived together by the block pointer territory, thereby saved memory space and access times.
The data storage area is divided into DRAM part and SRAM part, is mainly used in the payload of store data bag among the DRAM, is mainly used in the packet header of store data bag among the SRAM.If owing to the handled information spinner header data of network processing unit, and general of payload part is visited once when receiving and transmit, and the header part is stored in help to improve processing speed of data among the SRAM.
This memory block organizes management method can realize distribution, release and the packet of data block and data block searching, revising and operation such as deletion in formation at an easy rate, packet and searching of data block all only need a secondary index and a sequential search in the formation, the time complexity that O (n+1) only need be arranged, simultaneously, queue table, pointer memory block and data packet head all are placed among the SRAM, help further improving data block distribution, search, the speed of release and data processing.
To the improvement of DRAM storage control design mainly is by hiding the precharge time of DRAM, and utilizes repeatedly burst transfer to improve the transmission speed of memory.Because the addressing of SDRAM has exclusivity,, resend row/column address so after finishing the reading and writing operation,, the row of working originally will be closed at every turn if another row is carried out addressing.This work on hand row of closing, the operation that newline is opened in preparation is exactly precharge.Because the memory bank among the DRAM can be interfered storage capacitance because of the row gating, so precharge is a kind of process of all memory banks in the row of working being carried out data rewrite, precharge can be repaired affected data-signal, but also therefore can bring certain delay.
Fig. 2 is the sequential of a DRAM read operation, therefrom can see the influence that precharge delay is brought.
Usually the DRAM memory all can carry out a precharge at every turn after finishing a read-write operation, but if reading and writing are with the data of delegation continuously, then do not need all to carry out precharge, and it is just passable to carry out a precharge again after only needing the last time the reading and writing of this row to be finished at every turn.
Fig. 3 illustrated to same row carry out reading continuously for twice, precharge sequential chart once.
Utilize this characteristic of dram chip just, the improved method of most memories all is by continuous internal storage access instruction is cushioned and predicts, and by reordering, continuous memory access is concentrated on in the delegation as far as possible, thereby reduce the precharge number of times, reach the purpose that improves memory access speed.Fig. 3 can be regarded as improved band prediction and instructs the memory read operation sequential chart that reorders, compare with operation shown in Figure 2, remarkable advantages is arranged, but this prediction needs concrete consider procotol and scheduling strategy, ordering arbitrarily may have influence on the correctness of data processing and scheduling sometimes
In our implementation, consider that the data storage area is with 64 byte piecemeals, and most operations about agreement and scheduling are that unit carries out with data block and packet all, and, the memory capacity of DRAM storage chip delegation generally is the integral multiple of 64 bytes, in we can just be arranged in each data block with delegation when memory block is divided, in storage control, increase instruction buffer and analytic function simultaneously, for continuous memory reference instruction, we are that unit reorders to instruction with the visit to same always, connected reference is all concentrated on in delegation's piece, thereby precharge time can be reduced, simultaneously, when relating to the connected reference instruction of striding piece, also can further avoid precharge operation by judging whether to belong to, the improvement method for designing of this storage control with delegation, do not need to carry out complicated especially decision logic, realize effectively simple.As continuous visit A1, A2, A3, A4, A5.Suppose A1, A3, A5 are the visits to same, and A2, A4 are the visits to another piece, and we can reset access sequence and be A1, A3, A5, A2, A4.In original access sequence, suppose that two pieces not in delegation, then need 4 precharge at least, and in new access sequence, only need between the A2 visit, carry out a precharge at A5.
The DRAM memory all provides non-burst and happens suddenly two kinds of data-transmission modes, so-called burst transfer be meant storage chip with delegation in adjacent memory unit carry out data transmission manner continuously, promptly as long as specify initial column address and burst length, the access of addressing and data is just carried out automatically, as long as and control the gap periods of two sections burst access orders well and can accomplish continuous burst transfer, the quantity of each involved memory cell of burst transfer (row) is exactly burst length (BurstLengths is called for short BL).Fig. 2 and Fig. 3 have illustrated that burst length is 4 burst transfer process, non-burst continuous transmission mode does not adopt burst transfer but addressing separately successively, can be equivalent to BL=1 this moment, though it can allow the data be continuous transmission, but all to send column address and command information at every turn, the control resource occupation is very big, therefore should use the burst transfer pattern when reservoir designs as far as possible.
But the length of burst transfer is not to be to be the bigger the better, if the valid data of each transmission are less, and that BL (BL=1,2,4,8) establishes is excessive, and the relative time that causes transmitting invalid data is longer, and counter attending the meeting causes the decline of efficiency of transmission.Because the storage bottleneck mainly concentrates on data in network processing unit reception storage and transmission are read, so should emphasis considering, the setting of BL receives or the length of the unit data of transmit block, if be operating unit with the N bit data at every turn, the data wire width of storage control is the L position, then BL should be made as the immediate effective value that is less than or equal to N/L, for example, if N=128, L=32, then BL is taken as 4 proper.The length of BL can be configured by special register.
Organization and administration mode in conjunction with above-mentioned memory, the structured flowchart of interface unit specific implementation as shown in Figure 4, this interface unit mainly contains following each circuit arrangement and forms: Data Receiving buffer unit 1, pointer storage area management device 2, dram controller 3, SRAM controller 4, queue management device 5, SRAM pointer memory block 6, SRAM data storage area 7 and queue table 8 can external DRAM memories by dram controller 3.
A queue management device 5 is used for the organization and administration of formation and the distribution of deal with data;
Wherein, Data Receiving buffer unit 1 links to each other with multiple arrangement, be responsible for to receive the data row buffering of going forward side by side, carry out Interface status simultaneously and follow the tracks of and receive the counting of data, receiving data wire and adopt 128 bit widths, is that unit receives processing with a packet, in addition, accept buffer unit and pointer storage area management device, queue management device, DRAM storage control and SRAM storage control have interface, be used to ask each device that the data that receive are carried out concurrent processing.
Pointer storage area management device 2 is used for the organization and administration of memory block, finishes the management of memory block pointer and the distribution of maintenance and DSB data store block, is responsible for finishing certain flow control function,
The DRAM storage control, arbitration mechanism and three data access interfaces are provided: an interface is for the visit of Data Receiving parts inside to memory, an interface offers microprocessor, an interface offers the data forwarding parts, for the request from a plurality of interfaces, different priority is set, Data Receiving interface unit priority is the highest, microprocessor priority is inferior high, and it is minimum to transmit parts priority, and carries out granted access by the absolute priority strategy; The DRAM storage control also provides Instructions Cache, instruction analysis and reordering function, can many access instructions of a buffer memory, can analyze and reorder the instruction of buffer memory, to the rearrangement of instruction whether being sort criteria in same.
Except there is interface inside, storage control, pointer store trivial management devices and the queue management device also provides the external reference interface, realizes the visit to data of microprocessor and transmit block.
This interface unit uses above-mentioned memory block to organize management method and the design of improved storage control when specific implementation.
State transition graph when Fig. 5 is this interface unit operation, after device initialize is finished, when the receiving interface parts receive enough data, will ask pointer storage area management device to carry out data block distributes, carry out the packet queuing according to reception condition and pre-configured request queue management devices simultaneously, under the situation that the Flow Control inspection is passed through, by storage control storage packet data, payload for packet, be stored in the DRAM memory by dram controller, for the packet header of packet, then be stored in the SRAM memory by the SRAM controller.The queue management device is responsible for finishing the maintenance work of joining the team of packet and formation on the one hand, also be responsible for the request of response microprocessor simultaneously, when microprocessor requests is carried out new processing data packets, be responsible for by SRAM storage control read data packet header data, and send to microprocessor.
When finishing above action, storage control allows external unit, mainly be microprocessor and transmit block visit data memory block, storage control provides the arbitration controlled function, for different visits is provided with different priority, wherein, the priority of Data Receiving interface unit inter access is the highest, microprocessor priority is inferior high, and transmit block priority is minimum, and carries out granted access by the absolute priority strategy.
Pointer buffer management parts and queue management parts also allow external component simultaneously it to be conducted interviews simultaneously, and itself also has arbitrated logic, and the use absolute priority strategy identical with storage control.By the visit to pointer buffering area parts and queue management parts, external component can be finished the application and the distribution of free memory blocks, and operations such as the inquiry of data block and packet, modification, deletion, and management component itself is responsible for guaranteeing the effectively correct of operation.
Can see from above operation, this interface unit can receive in real time and cushion data, and can control other each component devices and carry out parallel work-flow, the operation that can walk abreast mainly comprises: the distribution of memory block, the storage of data, the execution of Flow Control, the tissue of formation and the distribution of deal with data.
Fig. 6 is the packet memory management method flow chart of network processing unit, and its operating procedure is as follows:
(1) preassignment DSB data store block is distributed free memory blocks and respective pointer to the application of idle queues manager;
(2) reception, buffered data;
(3), otherwise change (5) if new packet changes (4);
(4) data are write newly assigned SRAM data storage area, revise queue pointer and storage block pointer, the data that the last time is received wrap into respective queue simultaneously, change (8);
(5) judge to be header data,, otherwise change (7) if change (6);
(6) buffer data is write the SRAM data storage area;
(7) buffer data is write the DRAM data storage area;
(8) judge whether current memory block has been write completely,, otherwise change (2) if write full continuation;
(9) revise the memory block pointer and will write full memory block adding packet chained list,, change (2) to the application of idle queues manager new memory space and respective pointer.
Claims (9)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100500047A CN100440854C (en) | 2004-06-25 | 2004-06-25 | A data packet receiving interface part of a network processor and its storage management method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100500047A CN100440854C (en) | 2004-06-25 | 2004-06-25 | A data packet receiving interface part of a network processor and its storage management method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1595910A true CN1595910A (en) | 2005-03-16 |
| CN100440854C CN100440854C (en) | 2008-12-03 |
Family
ID=34665885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004100500047A Expired - Fee Related CN100440854C (en) | 2004-06-25 | 2004-06-25 | A data packet receiving interface part of a network processor and its storage management method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100440854C (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100376099C (en) * | 2005-07-04 | 2008-03-19 | 清华大学 | Integrated Queue Management Method Based on Network Processor Platform |
| CN100386752C (en) * | 2006-06-20 | 2008-05-07 | 北京飞天诚信科技有限公司 | Online updating method for USB device when communication protocol constrained |
| CN101075930B (en) * | 2006-05-16 | 2011-07-27 | 汤姆森许可贸易公司 | Network storage device |
| CN101605100B (en) * | 2009-07-15 | 2012-04-25 | 华为技术有限公司 | Method and apparatus for managing queue storage space |
| CN102567241A (en) * | 2010-12-27 | 2012-07-11 | 北京国睿中数科技股份有限公司 | Memory controller and memory access control method |
| WO2013020429A1 (en) * | 2011-08-11 | 2013-02-14 | 中兴通讯股份有限公司 | Network processor mirror implementation method and network processor |
| CN101808029B (en) * | 2009-02-13 | 2013-03-13 | 雷凌科技股份有限公司 | Method and device for preloading packet header and system using method |
| CN103314362A (en) * | 2010-12-17 | 2013-09-18 | 意法爱立信有限公司 | Vector-based matching circuit for data streams |
| CN103490939A (en) * | 2012-06-11 | 2014-01-01 | 中兴通讯股份有限公司 | Data packet processing method and data packet processing device |
| WO2014101192A1 (en) * | 2012-12-31 | 2014-07-03 | 华为技术有限公司 | Network device and message processing method |
| CN104811495A (en) * | 2015-04-27 | 2015-07-29 | 北京交通大学 | Method and module for content storage of network component of smart and cooperative network |
| WO2016019554A1 (en) * | 2014-08-07 | 2016-02-11 | 华为技术有限公司 | Queue management method and apparatus |
| CN107369473A (en) * | 2016-05-13 | 2017-11-21 | 爱思开海力士有限公司 | Storage system and its operating method |
| WO2018040600A1 (en) * | 2016-08-31 | 2018-03-08 | 深圳市中兴微电子技术有限公司 | Forwarding table-based information processing method and apparatus, and computer readable storage medium |
| CN109413122A (en) * | 2017-08-16 | 2019-03-01 | 深圳市中兴微电子技术有限公司 | Data processing method, network processor and computer storage medium |
| CN113779019A (en) * | 2021-01-14 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Current limiting method and device based on annular linked list |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR0150072B1 (en) * | 1995-11-30 | 1998-10-15 | 양승택 | Memory data path controller in parallel processing computer system |
| US6983350B1 (en) * | 1999-08-31 | 2006-01-03 | Intel Corporation | SDRAM controller for parallel processor architecture |
| US6754795B2 (en) * | 2001-12-21 | 2004-06-22 | Agere Systems Inc. | Methods and apparatus for forming linked list queue using chunk-based structure |
| CN1214541C (en) * | 2002-02-04 | 2005-08-10 | 华为技术有限公司 | Communication method between inner core and microengine inside network processor |
-
2004
- 2004-06-25 CN CNB2004100500047A patent/CN100440854C/en not_active Expired - Fee Related
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100376099C (en) * | 2005-07-04 | 2008-03-19 | 清华大学 | Integrated Queue Management Method Based on Network Processor Platform |
| CN101075930B (en) * | 2006-05-16 | 2011-07-27 | 汤姆森许可贸易公司 | Network storage device |
| CN100386752C (en) * | 2006-06-20 | 2008-05-07 | 北京飞天诚信科技有限公司 | Online updating method for USB device when communication protocol constrained |
| CN101808029B (en) * | 2009-02-13 | 2013-03-13 | 雷凌科技股份有限公司 | Method and device for preloading packet header and system using method |
| CN101605100B (en) * | 2009-07-15 | 2012-04-25 | 华为技术有限公司 | Method and apparatus for managing queue storage space |
| CN103314362B (en) * | 2010-12-17 | 2016-09-21 | 瑞典爱立信有限公司 | Match circuit based on vector for data stream |
| CN103314362A (en) * | 2010-12-17 | 2013-09-18 | 意法爱立信有限公司 | Vector-based matching circuit for data streams |
| CN102567241A (en) * | 2010-12-27 | 2012-07-11 | 北京国睿中数科技股份有限公司 | Memory controller and memory access control method |
| WO2013020429A1 (en) * | 2011-08-11 | 2013-02-14 | 中兴通讯股份有限公司 | Network processor mirror implementation method and network processor |
| CN103490939A (en) * | 2012-06-11 | 2014-01-01 | 中兴通讯股份有限公司 | Data packet processing method and data packet processing device |
| WO2014101192A1 (en) * | 2012-12-31 | 2014-07-03 | 华为技术有限公司 | Network device and message processing method |
| WO2016019554A1 (en) * | 2014-08-07 | 2016-02-11 | 华为技术有限公司 | Queue management method and apparatus |
| CN106537858A (en) * | 2014-08-07 | 2017-03-22 | 华为技术有限公司 | Queue management method and apparatus |
| CN106537858B (en) * | 2014-08-07 | 2019-07-19 | 华为技术有限公司 | A method and device for queue management |
| US10248350B2 (en) | 2014-08-07 | 2019-04-02 | Huawei Technologies Co., Ltd | Queue management method and apparatus |
| CN104811495B (en) * | 2015-04-27 | 2018-06-08 | 北京交通大学 | A kind of networking component content storage method and module for wisdom contract network |
| CN104811495A (en) * | 2015-04-27 | 2015-07-29 | 北京交通大学 | Method and module for content storage of network component of smart and cooperative network |
| CN107369473A (en) * | 2016-05-13 | 2017-11-21 | 爱思开海力士有限公司 | Storage system and its operating method |
| CN107797942A (en) * | 2016-08-31 | 2018-03-13 | 深圳市中兴微电子技术有限公司 | Reduce the method and its device of Large Copacity forward table access times |
| WO2018040600A1 (en) * | 2016-08-31 | 2018-03-08 | 深圳市中兴微电子技术有限公司 | Forwarding table-based information processing method and apparatus, and computer readable storage medium |
| CN107797942B (en) * | 2016-08-31 | 2020-11-20 | 深圳市中兴微电子技术有限公司 | Method and device for reducing access times of large-capacity forwarding table |
| CN109413122A (en) * | 2017-08-16 | 2019-03-01 | 深圳市中兴微电子技术有限公司 | Data processing method, network processor and computer storage medium |
| CN109413122B (en) * | 2017-08-16 | 2022-05-13 | 深圳市中兴微电子技术有限公司 | A data processing method, network processor and computer storage medium |
| CN113779019A (en) * | 2021-01-14 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Current limiting method and device based on annular linked list |
| CN113779019B (en) * | 2021-01-14 | 2024-05-17 | 北京沃东天骏信息技术有限公司 | Circular linked list-based current limiting method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100440854C (en) | 2008-12-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1595910A (en) | A data packet receiving interface component of network processor and storage management method thereof | |
| US7724735B2 (en) | On-chip bandwidth allocator | |
| WO2007004159A2 (en) | Method and apparatus for bandwidth efficient and bounded latency packet buffering | |
| US20050219564A1 (en) | Image forming device, pattern formation method and storage medium storing its program | |
| US20020124149A1 (en) | Efficient optimization algorithm in memory utilization for network applications | |
| US8677075B2 (en) | Memory manager for a network communications processor architecture | |
| US20050220112A1 (en) | Distributed packet processing with ordered locks to maintain requisite packet orderings | |
| CN101499956B (en) | Hierarchical buffer zone management system and method | |
| US7529224B2 (en) | Scheduler, network processor, and methods for weighted best effort scheduling | |
| US12068972B1 (en) | Shared traffic manager | |
| JP2004536515A (en) | Switch fabric with dual port memory emulation | |
| CN103345451A (en) | Data buffering method in multi-core processor | |
| Hasan et al. | Efficient use of memory bandwidth to improve network processor throughput | |
| US6697923B2 (en) | Buffer management method and a controller thereof | |
| US20160085477A1 (en) | Addressless merge command with data item identifier | |
| WO2016019554A1 (en) | Queue management method and apparatus | |
| CN1677958A (en) | A compact packet-switched node memory architecture using double-rate synchronous dynamic RAM | |
| CN1829200A (en) | Systems and methods for implementing counters in a network processor | |
| US10846225B1 (en) | Buffer read optimizations in a network device | |
| JP2003228461A (en) | Disk cache management method for disk array device | |
| US10742558B1 (en) | Traffic manager resource sharing | |
| EP1471430B1 (en) | Stream memory manager | |
| US9804959B2 (en) | In-flight packet processing | |
| CN1781079A (en) | Maintaining entity order with gate managers | |
| CN101464839B (en) | Access buffering mechanism and method |
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 | ||
| ASS | Succession or assignment of patent right |
Owner name: G-CLOUD TECHNOLOGY CO., LTD. Free format text: FORMER OWNER: INSTITUTE OF COMPUTING TECHNOLOGY, CHINESE ACADEMY OF SCIENCES Effective date: 20140514 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| COR | Change of bibliographic data |
Free format text: CORRECT: ADDRESS; FROM: 100080 HAIDIAN, BEIJING TO: 523808 DONGGUAN, GUANGDONG PROVINCE |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20140514 Address after: 523808 Guangdong province Dongguan City Songshan Lake Science and Technology Industrial Park Building No. 14 Keyuan pine Patentee after: G-CLOUD TECHNOLOGY Co.,Ltd. Address before: 100080 No. 6 South Road, Zhongguancun Academy of Sciences, Beijing Patentee before: Institute of Computing Technology, Chinese Academy of Sciences |
|
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Li Xiaowei Inventor after: Li Huawei Inventor after: Gong Shuguang Inventor after: Xu Yufeng Inventor after: Liu Tong Inventor before: Gong Shuguang Inventor before: Li Huawei Inventor before: Xu Yufeng Inventor before: Liu Tong Inventor before: Li Xiaowei |
|
| CP02 | Change in the address of a patent holder | ||
| CP02 | Change in the address of a patent holder |
Address after: 523808 19th Floor, Cloud Computing Center, Chinese Academy of Sciences, No. 1 Kehui Road, Songshan Lake Hi-tech Industrial Development Zone, Dongguan City, Guangdong Province Patentee after: G-CLOUD TECHNOLOGY Co.,Ltd. Address before: 523808 No. 14 Building, Songke Garden, Songshan Lake Science and Technology Industrial Park, Dongguan City, Guangdong Province Patentee before: G-CLOUD TECHNOLOGY Co.,Ltd. |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20081203 |