CN108090003A - A kind of method, the system of the promotion WEB server performance based on zero-copy - Google Patents
A kind of method, the system of the promotion WEB server performance based on zero-copy Download PDFInfo
- Publication number
- CN108090003A CN108090003A CN201711153964.XA CN201711153964A CN108090003A CN 108090003 A CN108090003 A CN 108090003A CN 201711153964 A CN201711153964 A CN 201711153964A CN 108090003 A CN108090003 A CN 108090003A
- Authority
- CN
- China
- Prior art keywords
- data packet
- data
- ring
- interrupt
- interface card
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F13/14—Handling requests for interconnection or transfer
- G06F13/20—Handling requests for interconnection or transfer for access to input/output bus
- G06F13/32—Handling requests for interconnection or transfer for access to input/output bus using combination of interrupt and burst mode transfer
- G06F13/34—Handling requests for interconnection or transfer for access to input/output bus using combination of interrupt and burst mode transfer with priority control
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/50—Queue scheduling
- H04L47/62—Queue scheduling characterised by scheduling criteria
- H04L47/622—Queue service order
- H04L47/6225—Fixed service order, e.g. Round Robin
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L49/00—Packet switching elements
- H04L49/90—Buffering arrangements
- H04L49/9063—Intermediate storage in different physical parts of a node or terminal
- H04L49/9068—Intermediate storage in different physical parts of a node or terminal in the network interface card
- H04L49/9073—Early interruption upon arrival of a fraction of a packet
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer And Data Communications (AREA)
Abstract
A kind of method of the promotion WEB server performance based on zero-copy, it is characterised in that：PF_RING modules are loaded in linux kernel, the network interface card for being loaded with PF_RING modules uses the NAPI mode Catch the Data Packs based on device polling, data packet is stored in annular caching by the way of DMA, data packet is directly reached application layer by PF_RING modules by MMAP mechanism.The present invention sends and receivees rate by reducing intermediate copy process of the communication data between host storage inside, fixture block data packet, improves data communication efficiency, accelerates corresponding speed, so as to promote the performance of web server.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of promotion WEB server performances based on zero-copy
Method, system.
Background technology
Web server is also known as network server, www server, Http servers, it receives user's request from network,
And user's request is made corresponding.Web server can respond static page or picture, can also entrust to the generation of response
Some other programs enter CGI, JSP, PHP scripts etc., and the handling result of request is finally returned with form as defined in Http agreements
To browser.With the development of modern communication technology, Web system issues the important way with exchanging as various information, so
Its performance is the most important thing, in face of ten hundreds of users, improves that the performance of Web server is very urgent, this needs Web service
Utensil has higher processing data packets speed and lower transmission delay.And the factor for influencing web server performance mainly has two
Aspect：First, network interface card sends and receivees data packet by interrupt mode from network, and exist before network interface card and upper system
Data copy, interrupt operation and copy procedure will occupy substantial amounts of cpu resource, second is that the protocol stack of legacy system is in data packet
, it is necessary to carry out multiple memory copying in fractionation and encapsulation process, more system time and space resources are occupied.
The content of the invention
It is an object of the invention to propose a kind of method, the system of the promotion WEB server performance based on zero-copy, pass through
Intermediate copy process of the communication data between host storage inside is reduced, fixture block data packet sends and receivees rate, improves number
According to communication efficiency, accelerate corresponding speed, so as to promote the performance of web server.
For this purpose, the present invention uses following technical scheme：
A kind of method of the promotion WEB server performance based on zero-copy loads PF_RING modules in linux kernel,
The network interface card of PF_RING modules is loaded with using the NAPI mode Catch the Data Packs based on device polling, by number by the way of DMA
It is stored according to bag in annular caching, data packet is directly reached application layer by PF_RING modules by MMAP mechanism.
Using interrupt mechanism and the NAPI Catch the Data Packs of polling mechanism, improve data packet and send and receive rate, add
PF_RING modules are carried, the data packet of capture is copied in annular caching by dma mode, the participation of CPU is reduced, reduces
CPU handles expense, improves system performance；Directly accessing the data in annular caching by MMAP modes again reduces a number
According to copy, upper system directly accesses annular caching by way of MMAP, reduces copy number of the data in kernel, section
Save system space and resource；And the loading mode of PF_RING, it is not necessary to change trawl performance, improve network interface card versatility and can
Transplantability has not interfered with the Web server of existing deployment on line and has used.
Preferably, NAPI modes are combined using interrupt mechanism and polling mechanism, and interrupt mechanism is used when data volume is low
Data are received, when data volume is high polling mechanism are used to receive data, the NAPI modes based on device polling are used including network interface card
The process of Catch the Data Pack：
Step A：When network equipment received data packet or transmission data packet, triggering interrupt processing function performs, network interface card
Interrupt signal is sent, CPU receives interrupt signal, and opens interrupters processing starts and handles interrupt requests；
Step B：If during interrupt processing function is triggered, having data packet to reach needs to receive, then interrupt processing function
Interrupt processing can be closed to start, by network device registration into snoop queue, and the interrupt response to the network equipment is closed, touch
Feel like jelly interrupt processing；
Step C：After traps processing is triggered, start traps processing, first polled network equipment, then read all
Data packet, after reading, the network equipment is discharged from snoop queue, and checks that snoop queue is whether as sky, if it is empty, then
Traps processing terminates；If not empty, then a new wheel traps processing is restarted, until snoop queue is sky.
In a busy network, when having network packet arrival every time, it need not all trigger interruption, because in high-frequency
The disconnected whole efficiency that may influence system, under high flow capacity, the interruption that network interface card generates is likely to be breached thousands of times per second, and if
It interrupts every time and is required for system to handle, be a very big pressure, and when NAPI is using poll is the reception for prohibiting network interface card
It interrupts, can so reduce the pressure that system processing is interrupted, NAPI is a kind of raising network processes efficiency used on Linux
Technology, its key concept are exactly that data are not read by the way of interruption, and instead wake up data using interruption first and connect
The service routine of receipts, then the method for POLL carry out polling data.By using the mode that interrupt mechanism and polling mechanism combine, carry
High data packet sends and receivees rate, and then improves network processes speed, further promotes the performance of WEB server.
Preferably, the annular caching is PF_RING interface a and PF_RING interfaces b, PF_ respectively there are two interfaces
RING interfaces a writes data packet thereto for network interface card, and PF_RING interfaces b provides read data packet for application layer；Including being loaded with
Data packet is reached the process of application layer by the network interface card of PF_RING modules：
Step S1：Network interface card after the NAPI mode Catch the Data Packs of use, passes through the side of DMA by loading PF_RING modules
Formula copies data packet to annular caching；
Step S2:Data packet is write PF_RING interfaces a by network interface card by way of MMAP from ring-type caching；
Step S3:Data packet is directly transferred to application layer by PF_RING interfaces b by way of MMAP.
Be loaded with PF_RING modules Linux network interface cards employ the mode Catch the Data Pack of NAPI after, data packet is copied
To the circle queue of kernel buffers；The interface of read data packet is realized by MMAP modes.PF_RING is by data packet from kernel
Circle queue is passed directly to application layer by MMAP modes, solves the problems, such as traditional kernel processes data packet efficiency than relatively low.
Preferably, including copying data packet to annular caching by way of DMA, and the process in memory is read：
Step R1：CPU carries out dma controller initialization operation and operational order is sent to I/O interfaces；
Step R2：I/O interfaces propose DMA request, and dma controller carries out DMA request priority and judges simultaneously respond request；
Step R3：Then judge whether data packet being loaded into memory, if so, CPU directly reads data packet
In memory；If it is not, then reading data packet in annular caching, then CPU is again by digital independent into memory.
DMA refers to that, directly in memory access, hard disk accesses data using dma mode, it is possible to reduce hard disk accounts for CPU
With rate, the External Transfer Rate of hard disk is improved；MMAP is exactly memory mapping, is mapped using memory, it is possible to reduce memory copying is saved
System time and space resources, so as to promote WEB server performance.
Preferably, the MMAP mechanism only carries out a physical address map for every segment data to the behaviour of virtual address
Make, i.e., subsequent access will no longer be detected the legitimacy of address by the data that MMAP mechanism mapped；
MMAP mechanism need not copy data, and data address pointer directly is transferred to application layer.
A kind of system of the promotion WEB server performance methodology based on zero-copy, captures data, so by way of NAPI
The data packet of capture is copied in annular caching by dma mode afterwards, directly reaches data packet finally by MMAP mechanism
Application layer；Including using NAPI mode Catch the Data Pack modules, it is be combined with each other using interrupt mechanism and polling mechanism, for when number
Data are received using interrupt mechanism according to when measuring low, when data volume is high polling mechanism are used to receive data；
When network equipment received data packet or transmission data packet, triggering interrupt processing function performs, during network interface card is sent
Break signal, CPU receive interrupt signal, and opens interrupters processing starts and handles interrupt requests；
If during interrupt processing function is triggered, having data packet to reach needs to receive, then interrupt processing function can close
Interrupt processing starts, and by network device registration into snoop queue, and the interrupt response to the network equipment is closed, during triggering is soft
Disconnected processing；
After traps processing is triggered, starts traps processing, first polled network equipment, then read all data
Bag, after reading, the network equipment is discharged from snoop queue, and checks that snoop queue is whether as sky, if it is empty, then in soft
Disconnected processing terminates；If not empty, then a new wheel traps processing is restarted, until snoop queue is sky.
Preferably, including application layer received data packet module, data packet application layer will be reached for network interface card；Network interface card passes through
PF_RING modules are loaded, after the NAPI mode Catch the Data Packs of use, are copied to data packet by way of DMA annular slow
It deposits；Data packet is write PF_RING interfaces a by network interface card by way of MMAP from ring-type caching；PF_RING interfaces b is by data
Bag is directly transferred to application layer by way of MMAP.
Preferably, including copying data packet to annular cache module, data packet is copied to ring-type by way of DMA
Caching, then reads in memory；CPU carries out dma controller initialization operation and operational order is sent to I/O interfaces；
I/O interfaces propose DMA request, and dma controller carries out DMA request priority and judges simultaneously respond request；Then judge whether by
Data packet is loaded into memory, if so, CPU directly reads data packet in memory；If it is not, then data packet is read
In ring-type caching, then CPU is again by digital independent into memory.
Description of the drawings
Fig. 1 is that the network interface card of the present invention uses the flow chart of the NAPI mode Catch the Data Packs based on device polling；
Fig. 2 is that data packet is reached the flow chart of application layer by the network interface card of the present invention；
Fig. 3 is the flow chart that data packet is copied to annular caching of the present invention；
Specific embodiment
Technical solution to further illustrate the present invention below with reference to the accompanying drawings and specific embodiments.
A kind of method of promotion WEB server performance based on zero-copy of the present embodiment, loads in linux kernel
PF_RING modules are loaded with the network interface card of PF_RING modules using the NAPI mode Catch the Data Packs based on device polling, use
Data packet is stored in annular caching by the mode of DMA, and data packet is directly reached application by PF_RING modules by MMAP mechanism
Layer.
Preferably, NAPI modes are combined using interrupt mechanism and polling mechanism, and interrupt mechanism is used when data volume is low
Data are received, when data volume is high polling mechanism are used to receive data, as shown in Figure 1, being used including network interface card based on device polling
NAPI mode Catch the Data Packs process：
Step A：When network equipment received data packet or transmission data packet, triggering interrupt processing function performs, network interface card
Interrupt signal is sent, CPU receives interrupt signal, and opens interrupters processing starts and handles interrupt requests；
Step B：If during interrupt processing function is triggered, having data packet to reach needs to receive, then interrupt processing function
Interrupt processing can be closed to start, by network device registration into snoop queue, and the interrupt response to the network equipment is closed, touch
Feel like jelly interrupt processing；
Step C：After traps processing is triggered, start traps processing, first polled network equipment, then read all
Data packet, after reading, the network equipment is discharged from snoop queue, and checks that snoop queue is whether as sky, if it is empty, then
Traps processing terminates；If not empty, then a new wheel traps processing is restarted, until snoop queue is sky.
Preferably, the annular caching is PF_RING interface a and PF_RING interfaces b, PF_ respectively there are two interfaces
RING interfaces a writes data packet thereto for network interface card, and PF_RING interfaces b provides read data packet for application layer；As shown in Fig. 2,
The process of application layer is reached including being loaded with the network interface cards of PF_RING modules by data packet：
Step S1：Network interface card after the NAPI mode Catch the Data Packs of use, passes through the side of DMA by loading PF_RING modules
Formula copies data packet to annular caching；
Step S2:Data packet is write PF_RING interfaces a by network interface card by way of MMAP from ring-type caching；
Step S3:Data packet is directly transferred to application layer by PF_RING interfaces b by way of MMAP.
Preferably, as shown in figure 3, including copying data packet to annular caching by way of DMA, and memory is read
In process：
Step R1：CPU carries out dma controller initialization operation and operational order is sent to I/O interfaces；
Step R2：I/O interfaces propose DMA request, and dma controller carries out DMA request priority and judges simultaneously respond request；
Step R3：Then judge whether data packet being loaded into memory, if so, CPU directly reads data packet
In memory；If it is not, then reading data packet in annular caching, then CPU is again by digital independent into memory.
Preferably, the MMAP mechanism only carries out a physical address map for every segment data to the behaviour of virtual address
Make, i.e., subsequent access will no longer be detected the legitimacy of address by the data that MMAP mechanism mapped；
MMAP mechanism need not copy data, and data address pointer directly is transferred to application layer.
A kind of system of the promotion WEB server performance methodology based on zero-copy, captures data, so by way of NAPI
The data packet of capture is copied in annular caching by dma mode afterwards, directly reaches data packet finally by MMAP mechanism
Application layer；Including using NAPI mode Catch the Data Pack modules, it is be combined with each other using interrupt mechanism and polling mechanism, for when number
Data are received using interrupt mechanism according to when measuring low, when data volume is high polling mechanism are used to receive data；
When network equipment received data packet or transmission data packet, triggering interrupt processing function performs, during network interface card is sent
Break signal, CPU receive interrupt signal, and opens interrupters processing starts and handles interrupt requests；
If during interrupt processing function is triggered, having data packet to reach needs to receive, then interrupt processing function can close
Interrupt processing starts, and by network device registration into snoop queue, and the interrupt response to the network equipment is closed, during triggering is soft
Disconnected processing；
After traps processing is triggered, starts traps processing, first polled network equipment, then read all data
Bag, after reading, the network equipment is discharged from snoop queue, and checks that snoop queue is whether as sky, if it is empty, then in soft
Disconnected processing terminates；If not empty, then a new wheel traps processing is restarted, until snoop queue is sky.
Preferably, including application layer received data packet module, data packet application layer will be reached for network interface card；Network interface card passes through
PF_RING modules are loaded, after the NAPI mode Catch the Data Packs of use, are copied to data packet by way of DMA annular slow
It deposits；Data packet is write PF_RING interfaces a by network interface card by way of MMAP from ring-type caching；PF_RING interfaces b is by data
Bag is directly transferred to application layer by way of MMAP.
Preferably, including copying data packet to annular cache module, data packet is copied to ring-type by way of DMA
Caching, then reads in memory；CPU carries out dma controller initialization operation and operational order is sent to I/O interfaces；
I/O interfaces propose DMA request, and dma controller carries out DMA request priority and judges simultaneously respond request；Then judge whether by
Data packet is loaded into memory, if so, CPU directly reads data packet in memory；If it is not, then data packet is read
In ring-type caching, then CPU is again by digital independent into memory.
Advantageous effect：
1st, data packet is stored in annular caching using dma mode, reduces CPU processing expenses, improve system performance；
2nd, annular caching is directly accessed by way of MMAP, reduces copy number of the data in kernel, saves system
System space and resource；
3rd, using the loading mode of PF_RING, it is not necessary to change trawl performance, improve the versatility and portable of network interface card
Property.
The technical principle of the present invention is described above in association with specific embodiment.These descriptions are intended merely to explain the present invention's
Principle, and limiting the scope of the invention cannot be construed in any way.Based on explanation herein, the technology of this field
Personnel would not require any inventive effort the other specific embodiments that can associate the present invention, these modes are fallen within
Within protection scope of the present invention.
Claims (8)
- A kind of 1. method of the promotion WEB server performance based on zero-copy, it is characterised in that：PF_ is loaded in linux kernel RING modules are loaded with the network interface card of PF_RING modules using the NAPI mode Catch the Data Packs based on device polling, using DMA Mode data packet is stored in annular caching, data packet is directly reached application layer by PF_RING modules by MMAP mechanism.
- 2. a kind of method for promoting WEB server performance according to claim 1, it is characterised in that：During NAPI modes use Off line system and polling mechanism are combined, and when data volume is low interrupt mechanism are used to receive data, and poll is used when data volume is high Mechanism receives data, and the process of the NAPI mode Catch the Data Packs based on device polling is used including network interface card：Step A：When network equipment received data packet or transmission data packet, triggering interrupt processing function performs, and network interface card is sent Interrupt signal, CPU receive interrupt signal, and opens interrupters processing starts and handles interrupt requests；Step B：If during interrupt processing function is triggered, having data packet to reach needs to receive, then interrupt processing function can close It closes interrupt processing to start, by network device registration into snoop queue, and closes the interrupt response to the network equipment, triggering is soft Interrupt processing；Step C：After traps processing is triggered, starts traps processing, first polled network equipment, then read all data Bag, after reading, the network equipment is discharged from snoop queue, and checks that snoop queue is whether as sky, if it is empty, then in soft Disconnected processing terminates；If not empty, then a new wheel traps processing is restarted, until snoop queue is sky.
- 3. a kind of method of the promotion WEB server performance based on zero-copy according to claim 1, it is characterised in that：Institute State annular caching and there are two interfaces, be respectively PF_RING interface a and PF_RING interface b, PF_RING interfaces a supply network interface card to Data packet is wherein write, PF_RING interfaces b provides read data packet for application layer；Network interface card including being loaded with PF_RING modules Data packet is reached to the process of application layer：Step S1：Network interface card is by loading PF_RING modules, will by way of DMA after the NAPI mode Catch the Data Packs of use Data packet copies annular caching to；Step S2:Data packet is write PF_RING interfaces a by network interface card by way of MMAP from ring-type caching；Step S3:Data packet is directly transferred to application layer by PF_RING interfaces b by way of MMAP.
- 4. a kind of method of the promotion WEB server performance based on zero-copy according to claim 1, it is characterised in that：Bag It includes and copies data packet to annular caching by way of DMA, and read the process in memory：Step R1：CPU carries out dma controller initialization operation and operational order is sent to I/O interfaces；Step R2：I/O interfaces propose DMA request, and dma controller carries out DMA request priority and judges simultaneously respond request；Step R3：Then judge whether data packet being loaded into memory, if so, data packet is directly read memory by CPU In；If it is not, then reading data packet in annular caching, then CPU is again by digital independent into memory.
- 5. a kind of method of the promotion WEB server performance based on zero-copy according to claim 1, it is characterised in that：The MMAP mechanism only carries out operation of the physical address map to virtual address, i.e. subsequent access for every segment data The legitimacy of address will be no longer detected by the data that MMAP mechanism mapped；MMAP mechanism need not copy data, and data address pointer directly is transferred to application layer.
- 6. a kind of system of the promotion WEB server performance methodology based on zero-copy, it is characterised in that：It is caught by way of NAPI Data are caught, are then copied the data packet of capture in annular caching to by dma mode, finally by MMAP mechanism directly by number Application layer is reached according to bag；Including use NAPI mode Catch the Data Pack modules, it is be combined with each other using interrupt mechanism and polling mechanism, For receiving data using interrupt mechanism when data volume is low, when data volume is high polling mechanism is used to receive data；When network equipment received data packet or transmission data packet, triggering interrupt processing function performs, and network interface card sends interruption letter Number, CPU receives interrupt signal, and opens interrupters processing starts and handles interrupt requests；If during interrupt processing function is triggered, having data packet to reach needs to receive, then interrupt processing function can close interruption Processing starts, and by network device registration into snoop queue, and closes the interrupt response to the network equipment, triggers at traps Reason；After traps processing is triggered, starts traps processing, first polled network equipment, then read all data packets, read After taking, the network equipment is discharged from snoop queue, and checks that snoop queue is whether as sky, if it is empty, then at traps Reason terminates；If not empty, then a new wheel traps processing is restarted, until snoop queue is sky.
- 7. a kind of system of the promotion WEB server performance methodology based on zero-copy, feature exist according to claim 6 In：Including application layer received data packet module, data packet application layer will be reached for network interface card；Network interface card is by loading PF_RING After the NAPI mode Catch the Data Packs of use, data packet is copied to annular caching by way of DMA for module；Network interface card passes through Data packet is write PF_RING interfaces a by the mode of MMAP from ring-type caching；Data packet is passed through MMAP's by PF_RING interfaces b Mode is directly transferred to application layer.
- 8. a kind of system of the promotion WEB server performance methodology based on zero-copy, feature exist according to claim 6 In：Including copying data packet to annular cache module, data packet is copied to annular caching by way of DMA, is then read It gets in memory；CPU carries out dma controller initialization operation and operational order is sent to I/O interfaces；I/O interfaces propose DMA request, dma controller carry out DMA request priority and judge simultaneously respond request；Then judge whether data packet being loaded into In memory, if so, CPU directly reads data packet in memory；If it is not, then read data packet in annular caching, Then CPU is again by digital independent into memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711153964.XA CN108090003A (en) | 2017-11-20 | 2017-11-20 | A kind of method, the system of the promotion WEB server performance based on zero-copy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711153964.XA CN108090003A (en) | 2017-11-20 | 2017-11-20 | A kind of method, the system of the promotion WEB server performance based on zero-copy |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108090003A true CN108090003A (en) | 2018-05-29 |
Family
ID=62172650
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711153964.XA Pending CN108090003A (en) | 2017-11-20 | 2017-11-20 | A kind of method, the system of the promotion WEB server performance based on zero-copy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108090003A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109460671A (en) * | 2018-10-21 | 2019-03-12 | 北京亚鸿世纪科技发展有限公司 | A method of realizing that web page contents are anti-tamper based on operating system nucleus |
| CN109889403A (en) * | 2019-01-30 | 2019-06-14 | 北京云端智度科技有限公司 | A kind of CDN network flow rate testing methods based on polling mechanism |
| CN110830385A (en) * | 2019-09-25 | 2020-02-21 | 上海数荃数据科技有限公司 | Packet capturing processing method, network equipment, server and storage medium |
| CN111181736A (en) * | 2019-12-31 | 2020-05-19 | 奇安信科技集团股份有限公司 | Data transmission method, apparatus, system, medium and program product |
| CN111857731A (en) * | 2020-07-31 | 2020-10-30 | 广州锦行网络科技有限公司 | Flux storage method based on linux platform |
| CN113992608A (en) * | 2021-10-25 | 2022-01-28 | 北京科东电力控制系统有限责任公司 | Method, device and storage medium for optimizing network receiving and transmitting packet path |
| CN114189462A (en) * | 2021-12-08 | 2022-03-15 | 北京天融信网络安全技术有限公司 | Flow acquisition method and device, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101178694A (en) * | 2006-11-07 | 2008-05-14 | 莱克斯信息技术（北京）有限公司 | Intel e1000 zero copy method |
| CN101217573A (en) * | 2007-12-29 | 2008-07-09 | 厦门大学 | A Method of Accelerating Network Card Packet Capture |
| US20090225767A1 (en) * | 2008-03-05 | 2009-09-10 | Inventec Corporation | Network packet capturing method |
| CN102082698A (en) * | 2009-11-26 | 2011-06-01 | 上海大学 | Network data processing system of high performance core based on improved zero-copy technology |
| US8356299B2 (en) * | 2007-06-29 | 2013-01-15 | Lenovo (Beijing) Limited | Interrupt processing method and system |
-
2017
- 2017-11-20 CN CN201711153964.XA patent/CN108090003A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101178694A (en) * | 2006-11-07 | 2008-05-14 | 莱克斯信息技术（北京）有限公司 | Intel e1000 zero copy method |
| US8356299B2 (en) * | 2007-06-29 | 2013-01-15 | Lenovo (Beijing) Limited | Interrupt processing method and system |
| CN101217573A (en) * | 2007-12-29 | 2008-07-09 | 厦门大学 | A Method of Accelerating Network Card Packet Capture |
| US20090225767A1 (en) * | 2008-03-05 | 2009-09-10 | Inventec Corporation | Network packet capturing method |
| CN102082698A (en) * | 2009-11-26 | 2011-06-01 | 上海大学 | Network data processing system of high performance core based on improved zero-copy technology |
Non-Patent Citations (3)
| Title |
|---|
| 张令通，徐陶祎: "可编程DMA控制器8237A", 《微机原理与接口技术》 * |
| 温梨梨: "基于零拷贝的Web服务器技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 王亚，于干，张军，王浩: "PF_RING与NAPI结合的捕包性能优化和仿真", 《计算机工程与应用》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109460671A (en) * | 2018-10-21 | 2019-03-12 | 北京亚鸿世纪科技发展有限公司 | A method of realizing that web page contents are anti-tamper based on operating system nucleus |
| CN109460671B (en) * | 2018-10-21 | 2021-10-26 | 北京亚鸿世纪科技发展有限公司 | Method for realizing webpage content tamper resistance based on operating system kernel |
| CN109889403A (en) * | 2019-01-30 | 2019-06-14 | 北京云端智度科技有限公司 | A kind of CDN network flow rate testing methods based on polling mechanism |
| CN110830385A (en) * | 2019-09-25 | 2020-02-21 | 上海数荃数据科技有限公司 | Packet capturing processing method, network equipment, server and storage medium |
| CN111181736A (en) * | 2019-12-31 | 2020-05-19 | 奇安信科技集团股份有限公司 | Data transmission method, apparatus, system, medium and program product |
| CN111181736B (en) * | 2019-12-31 | 2022-04-05 | 奇安信科技集团股份有限公司 | Data transmission method, device, system and medium |
| CN111857731A (en) * | 2020-07-31 | 2020-10-30 | 广州锦行网络科技有限公司 | Flux storage method based on linux platform |
| CN113992608A (en) * | 2021-10-25 | 2022-01-28 | 北京科东电力控制系统有限责任公司 | Method, device and storage medium for optimizing network receiving and transmitting packet path |
| CN113992608B (en) * | 2021-10-25 | 2024-01-05 | 北京科东电力控制系统有限责任公司 | Network transceiver packet path optimization method, device and storage medium |
| CN114189462A (en) * | 2021-12-08 | 2022-03-15 | 北京天融信网络安全技术有限公司 | Flow acquisition method and device, electronic equipment and storage medium |
| CN114189462B (en) * | 2021-12-08 | 2024-01-23 | 北京天融信网络安全技术有限公司 | Flow acquisition method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108090003A (en) | A kind of method, the system of the promotion WEB server performance based on zero-copy | |
| US11886355B2 (en) | Emulated endpoint configuration | |
| Zhang et al. | {FlashShare}: Punching Through Server Storage Stack from Kernel to Firmware for {Ultra-Low} Latency {SSDs} | |
| US10678432B1 (en) | User space and kernel space access to memory devices through private queues | |
| CN107209681B (en) | Storage device access method, device and system | |
| US8631181B2 (en) | Validating message-signaled interrupts by tracking interrupt vectors assigned to devices | |
| US7836212B2 (en) | Reflecting bandwidth and priority in network attached storage I/O | |
| Flajslik et al. | Network Interface Design for Low Latency {Request-Response} Protocols | |
| CN102497434B (en) | Establishing method of kernel state virtual network equipment and packet transmitting and receiving methods thereof | |
| CN107992436A (en) | A kind of NVMe data read-write methods and NVMe equipment | |
| Anderson et al. | Cheating the I/O Bottleneck: Network Storage with Trapeze/Myrinet. | |
| US8996774B2 (en) | Performing emulated message signaled interrupt handling | |
| CN112463307A (en) | Data transmission method, device, equipment and readable storage medium | |
| CN109901909B (en) | Method for virtualizing system and virtualizing system | |
| WO2013048409A1 (en) | Writing message to controller memory space | |
| CN101702688A (en) | A method of sending and receiving data packets | |
| EP4439312A1 (en) | Data storage method and system, storage access configuration method and related device | |
| CN115129625A (en) | Enhanced storage protocol emulation in a peripheral device | |
| Wang et al. | Cxl over ethernet: A novel fpga-based memory disaggregation design in data centers | |
| CN117834561A (en) | Network I/O processing method and device based on SPDK | |
| US9288163B2 (en) | Low-latency packet receive method for networking devices | |
| US10977201B1 (en) | Per IO direct memory access redirection | |
| CN109120665B (en) | High-speed data packet collection method and device | |
| US11182103B1 (en) | Dedicated communications cache | |
| CN116560949B (en) | An implementation method, system, equipment and storage medium for data collection and reception |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180529 |
|
| RJ01 | Rejection of invention patent application after publication |