CN101873337A - Zero-copy data capture technology based on rt8169 gigabit net card and Linux operating system - Google Patents
Zero-copy data capture technology based on rt8169 gigabit net card and Linux operating system Download PDFInfo
- Publication number
- CN101873337A CN101873337A CN200910059018A CN200910059018A CN101873337A CN 101873337 A CN101873337 A CN 101873337A CN 200910059018 A CN200910059018 A CN 200910059018A CN 200910059018 A CN200910059018 A CN 200910059018A CN 101873337 A CN101873337 A CN 101873337A
- Authority
- CN
- China
- Prior art keywords
- data
- module
- ring
- packet
- user
- 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
Images
Abstract
The invention relates to a zero-copy data capture technology based on an rt8169 gigabit net card and a Linux operating system, which is characterized in positioning data capture in the Internet backbone network and support gigabit-class network data capture. In the technical scheme, a computer can be linked in a target network in parallel through the gigabit net card to capture data flowing through the target network, and the data capture in the gigabit network can be realized through modifying the net card drive and using the zero-copy technique. Since the technology realizes the zero copy of the network data packet from the kernel to the space of the user through construction of the Linux kernel module and modification of the rt8169 gigabit net card driving program, thereby lowering the complexity of the ordinary data collection, improving the data packet capture capability, reducing the cost of the system, and achieving the purpose of low cost. The technology is realized through three modules, the modified net card driving program, the constructed kernel module and the user process interface module. The modules cooperate with each other in operation to realize the high-speed transmission of the data packet from the net card to the user program.
Description
Technical field
The present invention relates to a kind of Zero-copy data capture technology based on rt8169 PCI-Express and (SuSE) Linux OS.This technology can be applied in the fields such as distributed real-time control system, network failure analysis, intruding detection system, network monitoring system, computer forensics system, is one of key technology wherein.
Background technology
Along with popularizing of network, safety problem is just threatening each network user.Therefore very necessary to the network monitoring of computer, and wherein to network packet catch and analyze particularly important.Network message is caught in the fields such as being widely used in distributed real-time control system, network failure analysis, intruding detection system, network monitoring system, computer forensics system, is one of key technology wherein.Along with the network bandwidth constantly increases, the demand of monitoring express network data flow is more and more obvious, high performance message capturing platform all has wide practical use in fields such as the network protocol analysis under the intruding detection system of extensive broadband network, big capaciated flow network data cases, broadband network fire compartment wall, high performance communication system, high-performance router, host router.
But, only be several years ago, the bandwidth of Ethernet or 100Mbps, the develop rapidly of Internet, make Ethernet bandwidth this short several in the period of, develop into 1000Mbps from 100Mbps, arrive 10Gbps again.Believe that in the near future the Ethernet of 10Gbps will be widely used as present 1000Mbps Ethernet.But performance of processors does not but lag behind the development of the network bandwidth greatly immediately following the improvement speed of its paces, especially memory bandwidth and access delay.Show that according to CNNIC a statistical report in 2005 our Internet network development is rapid, is 139.1 times in 1997.Simultaneously, the processing of network packet consumes cpu resource very much, even the fastest current processor, the ICP/IP protocol processing can consume 80% to 90% obtained computing cycle easily, and can not reach the performance of gigabit network speed.In network, spreading a wise remark of an experienced person: under best situation, the CPU of 1MHz can handle the network bandwidth of 1Mb, according to this rule, the CPU of existing 3.9GHz can only handle the network bandwidth of 3.9Gb, the performance of this CPU when the 10Gps network is not widely used is to satisfy the demand of the network bandwidth, has verified when network is used widely faster, can find the enough fast CPU to satisfy its demand.But comparatively speaking, memory bandwidth and access delay but develop fast inadequately, and therefore the gap between cpu performance along with being out of step that correlation technique develops, only depends on the raising of CPU frequency can not solve the problem that the express network bandwidth is brought fully in continuous increase.For network data bag quickly to satisfy the needs of the current network bandwidth, and don't too much occupying system resources, the system that makes can have more resources to be used for using processing, be necessary that again network packet is handled speed technology to be studied, do not catch up with the defective that paces are brought with delayed development speed often to remedy memory bandwidth.
So the importance of catching and analyzing of our packet Network Based and the efficiency of catching and analyzing of network packet propose this technology, purpose is to solve the efficiency of catching and analyzing of network packet, what guarantee network packet catches and analyzes reliability, practicality.
Summary of the invention
The objective of the invention is by revising rt8169 PCI-Express driver, the kernel module VCDM of structure, the data capture function under three functional modules realizations of user process interface module express network.
A, modification NIC driver: in the linux system, the NIC driver major function is that the packet that network interface card is received is submitted to kernel protocol stack.The existence of the protocol stack data transmission bauds that will slow down, and can't obtain the initial data of protocal layers according to real needs.Amended NIC driver will make packet walk around protocol stack and directly be submitted in the VCDM module, obtain for user program.Also can when giving kernel module, allow packet by giving protocol stack on the traditional mode according to the actual requirements.
The kernel module VCDM:VCDM module of B, structure is the important component part of high speed packet trapping module.Receive packet and realize the direct visit of user program from network interface card these packets.The design of linux system kernel module makes that directly adding new function to kernel becomes possibility.In carrying out high speed data capture, realize memory-mapped, thereby realize the data sharing of kernel spacing, improve data packet transmission efficient greatly to user's space by registration VCDM module in kernel.
C, user process interface module: consumer process does not have conflict by the interface module access function and obtains packet from the VCDM module.This module mainly realizes following three functions.The one, allow user's space can visit mapped kernel spacing data.The 2nd and NIC driver ensuring escapement from confliction visit VCDM module in the data field, thereby obtain packet.The 3rd, the requirement that this packet is used by the subsequent captured data is submitted to, keep supplying a layer module and be for further processing.
By the mutual cooperation of three functional modules, finish the capturing function of network data, the efficient height can reach the data capture ability in the gigabit network.
Description of drawings
Fig. 1 data capture overall work flow chart;
Fig. 2 VCDM module memory distributes schematic diagram;
Fig. 3 vcdm_mm_nopage function flow process;
The ring structure and the data_buf of Fig. 4 band resource description field concern schematic diagram;
Fig. 5 application data obtains flow chart.
Embodiment
Below will be to technical scheme of the present invention is elaborated, a kind of Zero-copy data capture technology design based on rt8169 PCI-Express and (SuSE) Linux OS is as follows:
A, amended NIC driver workflow:
1, after packet arrived, network interface card can send hard interruption.
2, NIC driver is returned to the node in the rx_free ring in the data_buf buffering area chained list in the Interrupt Process function, put simultaneously ring go up corresponding by the descriptor in the description scheme of structure for not using.
3, obtain structure in the data_buf buffering area chained list, packet is filled in this structure.
4, this structure is hung into rx_busy ring, put simultaneously ring go up corresponding by the descriptor in the description scheme of structure for using.
5, last signaling application program has new data packets to obtain.
B, application work flow process:
1, application program is received signal.
2, from rx_busy ring, obtain the buf structure, put simultaneously ring go up corresponding by the descriptor in the description scheme of structure for not using.
3, take out packet and it is encapsulated as the skb_buf structure.
4, packaged skb_buf structure being submitted to the upper strata registration function handles.
5, the buf structure is hung in the rx_free ring after disposing, put simultaneously ring go up corresponding by the descriptor in the description scheme of structure for using.
6, the 2-5 step is carried out in circulation, goes up no available descriptor up to rx_buxy.
C, kernel VCDM modular design:
VCDM is a virtual unit, is created at kernel by the user, with the form loading of module.This equipment is the key of native system, comprises receiving ring Rx_busy_ring, and idle ring Rx_busy_ring sends ring Tx_ring, and the buffering area Data_buf that is used for the store data bag.This module mainly acts on being to be implemented in avoids kernel spacing and under the prerequisite of the data of user's space copy, is positioned at the application program of user's space and is positioned at data passes between the NIC driver of kernel spacing.Provide correlation function to call to user program and NIC driver.The VCDM module mainly solves following three problems:
1, the application program that is positioned at user's attitude can be visited Rx_busy_ring, three rings of Rx_busy_ring and Tx_ring and data_buf.
2, allow network interface card data can be directly passed to data buffer zone data_buf.
3, provide method, allow network interface card and application program pass through Busy_ring by ensuring escapement from confliction, Free_ring visits data_buf.
Fig. 1 is a data capture overall work workflow diagram.
The important data structures of VCDM module:
1, receive ring, idle ring and transmission ring:
typedef?struct{
Unsigned int desc; Whether // descriptor is used
Unsigned int length; // hang over the buf length on this descriptor
Unsigned int paddr; // hang over the buf station address on this descriptor
Unsigned int kaddr; // hang over the buf kernel address on this descriptor
}vcdm_bd_t；
Ring structure is used to manage the BUF buffering area, realizes that user space program and network interface card can conflict free visit buf buffering areas.
2, Buf structure:
typedef?struct{
Char packet[1520]; // packet content
Char padded[516]; // fill field, take the internal memory of 2K size for guaranteeing this structure
Unsigned int kaddr; // this buf kernel address
Unsigned int paddr; // this buf station address
Unsigned int phyaddr; // this buf physical address (not using temporarily)
}vcdm_buf_t；
The Buf structure is the real buffering area of store data bag, distributes in kernel, and is designed to the structure of one of 2k.Because in linux system, the page or leaf size is 4K, and the data block of store data message is again greater than a MTU (1514 byte).Only put 2 data messages for every page like this, just guaranteed that a data message can not cross over two pages or leaves, and only each and page gets final product alternately when therefore network interface card carries out dma operation.
3, VCDM Memory Allocation:
In kernel spacing, with RX_busy_ring, Rx_free_ring, Tx_ring, and the buf buffering area that is used to deposit the True Data bag puts together, distributes together with vmalloc, and this internal memory all is mapped to user's attitude.VCDM Memory Allocation schematic diagram as shown in Figure 2.
The method of D, the direct visit data buffer memory of network interface card data_buf:
Generally adopt, and utilize user's physical address mapping table phy-addr-table that adding defines for UserBuff specially in virtual module at user's space distribute data buffering area.Solve the problem in the direct access buffer of network interface card district.This method is carried out address UBVA after page alignment obtains aliging by the user with UserBuff initialized the time, the first address UBVA after will aliging by system call again passes to kernel state.
The VCDM module is improved original data buffer zone method of salary distribution, changes at kernel spacing distributing this piece internal memory, and does not allow operating system scheduling that it is swapped out this sheet internal memory pinning.Simultaneously again in the buf structure, increase kernel spacing first address that a unsigned int kaddr is used to deposit oneself like this, the NIC driver that is positioned at kernel spacing equally just can directly be visited this sheet internal memory, and has avoided loaded down with trivial details tabling look-up and initialization operation in the conventional method.In addition, for simplifying the operation of network interface card to this sheet internal memory, we are the form of chained list with the structure organization in this sheet internal memory.Network interface card only needs to obtain the buffering area that a node can obtain the store data bag from linked list head in due course.
E, user's attitude visit busy_ring, free_ring, the method for data_buf:
In this programme, after packet was sent to the data_buf that is positioned at kernel state, this sheet buffering area of the application access of user's attitude was taken data away.Between network interface card and the application program is to realize by the respective element of revising among busy_ring and the free_ring synchronously.Yet the direct data of access kernel space of user's attitude program in (SuSE) Linux OS.How we could allow the application access of user's attitude arrive busy_ring, free_ring so.
(SuSE) Linux OS provides a kind of technology,, this technology can map directly to image file and data file the address space of process.In this mapping, the content of file is directly connected on the process virtual address space, memory-mapped technology that Here it is.We have just realized the address space rx_busy_ring among the virtual unit VCDM in this way, and rx_free_ring is mapped to user's space with metadata cache data_buf.
In the application program operation the most at first, unpromising its of operating system distributes enough pages, yet a program can normally have been moved in the same old way.This is that system will produce page fault because along with process is constantly visited those not pages in internal memory, and kernel will distribute idle page frame for it, and the inefficacy page is called in internal memory, and this processing mode is called page fault and handles.
In order to realize this function, we have defined the vcdm-mmap function in the system call of VCDM module, when application layer is carried out mmap to the vcdm virtual unit, can call this function.This function is newly set up a vma, simultaneously vma-＞vma-start is returned to application layer as the operable virtual address of user.When user program is visited data in this virtual address space, the page or leaf at this data place never call in internal memory or call in after when having been swapped out, the function vcdm_mm_nopage that operating system will be called us processing of skipping leaf.This function is a parameter with the station address, returns the Physical Page at this place, address.Its processing procedure as shown in Figure 3.
F, network interface card and application program ensuring escapement from confliction are by rx_busy_ring, and rx_free_ring visits the method for data_buf:
Owing to visit same block cache data_buf in the time of will transmitting with network interface card DMA during application layer read data message.So we are necessary to take certain measure, making does not influence systematic function when read/write memory is synchronous.The general employing realized each mode of encircling adding read-write pointer synchronously.In this programme, we adopt more simple resource description field to realize this function.
Add resource descriptor in ring structure, after hanging up the bud node in the ring structure, resource descriptor puts 1, when not hanging node on the ring structure, and resource descriptor position 0.Be positioned at the network interface card of kernel state and the application program of user's attitude and safeguard independently bd_cur respectively, as pointing to the next ring structure pointer that needs operation.Network interface card and user program can be realized the simultaneous operation to ring by judging the resource descriptor position.
The ring structure of band resource description field and data_buf relation are as shown in Figure 4.
Application data obtains flow chart as shown in Figure 5.
Claims (5)
1. Zero-copy data capture technology based on rt8169 PCI-Express and (SuSE) Linux OS, it is characterized in that, in high speed network environment, particularly in the gigabit networking environment, by using the assembling (SuSE) Linux OS, be connected in parallel in this network with the computer of rt8169 PCI-Express, the packet in the network carried out the high efficiency data capture by using zero duplication technology to reach.Reduced bottleneck and overhead during frequently-used data is caught, improved the ability of data capture, particularly at gigabit networking.
2. a kind of Zero-copy data capture technology based on rt8169 PCI-Express and (SuSE) Linux OS as claimed in claim 1 is characterized in that, this technology realizes by three functional modules, revise NIC driver, the kernel module that makes up, the user process interface module, its method is as follows:
A, modification NIC driver: in the linux system, the NIC driver major function is that the packet that network interface card is received is submitted to kernel protocol stack.Amended NIC driver will make packet walk around protocol stack and directly be submitted in the VCDM module, obtain for user program.Thereby improved data transmission bauds, equalled and can obtain the initial data of protocal layers according to real needs.Also can when giving kernel module, allow packet by giving protocol stack on the traditional mode according to the actual requirements.
The kernel module VCMD:VCDM module of B, structure is the important component part of high speed packet trapping module.This module realizes that user program receives the direct visit of packet to these from network interface card.When carrying out the express network data capture, realize memory-mapped by registration VCDM module in linux kernel, thereby realize the data sharing of kernel spacing to user's space, improve data packet transmission efficient greatly, be the key technology that realizes zero-copy.
C, user process interface module: consumer process does not have conflict by interface module and obtains the packet that network interface card is caught from the VCDM module.This module mainly realizes following three functions.The one, allow user's space can directly visit mapped kernel spacing data.The 2nd and NIC driver ensuring escapement from confliction visit VCDM module in the data field, thereby obtain packet.The 3rd, the requirement that this packet is used by the subsequent captured data is submitted to, keep supplying a layer module and be for further processing.
3. a kind of Zero-copy data capture technology based on rt8169 PCI-Express and (SuSE) Linux OS as claimed in claim 1 or 2 is characterized in that, amended NIC driver workflow is as follows:
After A, packet arrived, network interface card can send hard interruption.
B, NIC driver are returned to the data of free time ring in the chained list of data buffer zone in the Interrupt Process function.
C, will catch packet and be filled in the chained list of data buffer zone.
D, the data in this data buffer zone chained list are suspended to receive in the ring.
E, last signaling application program have new data packets to obtain.
4. a kind of Zero-copy data capture technology as claimed in claim 1 or 2 based on rt8169 PCI-Express and (SuSE) Linux OS, it is characterized in that, the VCMD module is a virtual unit, is created at the (SuSE) Linux OS kernel by the user, with the form loading of module.This equipment is the key of native system, comprises receiving ring Rx_busy_ring, and idle ring Rx_busy_ring sends ring Tx_ring, and the buffering area Data_buf that is used for the store data bag.This module mainly acts on being to be implemented in avoids kernel spacing and under the prerequisite of the data of user's space copy, is positioned at the application program of user's space and is positioned at data passes between the NIC driver of kernel spacing.Provide correlation function to call to user program and NIC driver.It mainly finishes following three functions:
A, the application program that is positioned at user's attitude can be visited and be received ring, idle ring and send ring and data packet buffer.
B, allow network interface card data can be directly passed to the data buffer zone.
C, provide method allow network interface card and application program can ensuring escapement from confliction by receiving ring, idle ring visit data buffering area.
5. a kind of Zero-copy data capture technology based on rt8169 PCI-Express and (SuSE) Linux OS as claimed in claim 1 or 2 is characterized in that the application work flow process is as follows:
A, application program are received signal.
B, from accept ring, obtain network interface card and catch packet.
C, take out packet and with its encapsulation.
D, packaged data are submitted to the upper strata registration function handle.
E, data are hung in the idle ring ring after disposing.
The B-E step is carried out in F, circulation, goes up no data available up to receiving ring.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910059018A CN101873337A (en) | 2009-04-22 | 2009-04-22 | Zero-copy data capture technology based on rt8169 gigabit net card and Linux operating system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910059018A CN101873337A (en) | 2009-04-22 | 2009-04-22 | Zero-copy data capture technology based on rt8169 gigabit net card and Linux operating system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101873337A true CN101873337A (en) | 2010-10-27 |
Family
ID=42997998
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910059018A Pending CN101873337A (en) | 2009-04-22 | 2009-04-22 | Zero-copy data capture technology based on rt8169 gigabit net card and Linux operating system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101873337A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102497434A (en) * | 2011-12-16 | 2012-06-13 | 中国科学院计算技术研究所 | Establishing method of kernel state virtual network equipment and packet transmitting and receiving methods thereof |
| CN102510376A (en) * | 2011-10-19 | 2012-06-20 | 浙江中烟工业有限责任公司 | Multi-component security isolation concurrent processing method |
| CN102662910A (en) * | 2012-03-23 | 2012-09-12 | 浙江大学 | Network interaction system based on embedded system and network interaction method |
| 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 |
| CN103678203A (en) * | 2013-12-13 | 2014-03-26 | 国家计算机网络与信息安全管理中心 | Method and device for achieving zero copy of network card |
| CN105553889A (en) * | 2015-12-21 | 2016-05-04 | 浪潮集团有限公司 | Fast packet-receiving/packet-transmitting transferring method between physical network card and virtual network card under kernel network bridge |
| CN106850565A (en) * | 2016-12-29 | 2017-06-13 | 河北远东通信系统工程有限公司 | A kind of network data transmission method of high speed |
| CN107273213A (en) * | 2017-06-27 | 2017-10-20 | 联想(北京)有限公司 | A kind of calculation control method, network interface card and electronic equipment |
| CN107302499A (en) * | 2017-06-26 | 2017-10-27 | 北京赛特斯信息科技股份有限公司 | The NFV protocol massages receiving/transmission methods of packet buffer need not be copied |
| CN108600053A (en) * | 2018-05-10 | 2018-09-28 | 南京邮电大学 | A kind of wireless network data Packet capturing method based on zero duplication technology |
| CN110602262A (en) * | 2018-06-13 | 2019-12-20 | 网宿科技股份有限公司 | Router and method for processing data message thereof |
| CN110855610A (en) * | 2019-09-30 | 2020-02-28 | 视联动力信息技术股份有限公司 | Data packet processing method and device and storage medium |
| CN111083072A (en) * | 2019-11-29 | 2020-04-28 | 天津市滨海新区信息技术创新中心 | Switch message transmission device and method |
| CN113422771A (en) * | 2021-06-22 | 2021-09-21 | 北京华圣龙源科技有限公司 | Threat early warning method and system |
| CN116455612A (en) * | 2023-03-23 | 2023-07-18 | 京信数据科技有限公司 | Privacy calculation intermediate data stream zero-copy device and method |
| CN116915667A (en) * | 2023-07-19 | 2023-10-20 | 上海螣龙科技有限公司 | Network traffic acquisition method and device, electronic equipment and storage medium |
-
2009
- 2009-04-22 CN CN200910059018A patent/CN101873337A/en active Pending
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102510376A (en) * | 2011-10-19 | 2012-06-20 | 浙江中烟工业有限责任公司 | Multi-component security isolation concurrent processing method |
| CN102510376B (en) * | 2011-10-19 | 2014-04-30 | 浙江中烟工业有限责任公司 | Multi-component security isolation concurrent processing method |
| CN102497434A (en) * | 2011-12-16 | 2012-06-13 | 中国科学院计算技术研究所 | Establishing method of kernel state virtual network equipment and packet transmitting and receiving methods thereof |
| CN102497434B (en) * | 2011-12-16 | 2014-11-05 | 中国科学院计算技术研究所 | Establishing method of kernel state virtual network equipment and packet transmitting and receiving methods thereof |
| CN102662910B (en) * | 2012-03-23 | 2014-10-15 | 浙江大学 | Network interaction system based on embedded system and network interaction method |
| CN102662910A (en) * | 2012-03-23 | 2012-09-12 | 浙江大学 | Network interaction system based on embedded system and network interaction method |
| 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 |
| CN103678203A (en) * | 2013-12-13 | 2014-03-26 | 国家计算机网络与信息安全管理中心 | Method and device for achieving zero copy of network card |
| CN105553889A (en) * | 2015-12-21 | 2016-05-04 | 浪潮集团有限公司 | Fast packet-receiving/packet-transmitting transferring method between physical network card and virtual network card under kernel network bridge |
| CN106850565A (en) * | 2016-12-29 | 2017-06-13 | 河北远东通信系统工程有限公司 | A kind of network data transmission method of high speed |
| CN106850565B (en) * | 2016-12-29 | 2019-06-18 | 河北远东通信系统工程有限公司 | A kind of network data transmission method of high speed |
| CN107302499A (en) * | 2017-06-26 | 2017-10-27 | 北京赛特斯信息科技股份有限公司 | The NFV protocol massages receiving/transmission methods of packet buffer need not be copied |
| CN107302499B (en) * | 2017-06-26 | 2020-12-18 | 北京赛特斯信息科技股份有限公司 | NFV protocol message transmitting and receiving method without copying message buffer |
| CN107273213A (en) * | 2017-06-27 | 2017-10-20 | 联想(北京)有限公司 | A kind of calculation control method, network interface card and electronic equipment |
| CN107273213B (en) * | 2017-06-27 | 2024-04-19 | 联想(北京)有限公司 | Calculation control method, network card and electronic equipment |
| CN108600053A (en) * | 2018-05-10 | 2018-09-28 | 南京邮电大学 | A kind of wireless network data Packet capturing method based on zero duplication technology |
| CN110602262A (en) * | 2018-06-13 | 2019-12-20 | 网宿科技股份有限公司 | Router and method for processing data message thereof |
| CN110855610A (en) * | 2019-09-30 | 2020-02-28 | 视联动力信息技术股份有限公司 | Data packet processing method and device and storage medium |
| CN111083072A (en) * | 2019-11-29 | 2020-04-28 | 天津市滨海新区信息技术创新中心 | Switch message transmission device and method |
| CN113422771A (en) * | 2021-06-22 | 2021-09-21 | 北京华圣龙源科技有限公司 | Threat early warning method and system |
| CN116455612A (en) * | 2023-03-23 | 2023-07-18 | 京信数据科技有限公司 | Privacy calculation intermediate data stream zero-copy device and method |
| CN116455612B (en) * | 2023-03-23 | 2023-11-28 | 京信数据科技有限公司 | Privacy calculation intermediate data stream zero-copy device and method |
| CN116915667A (en) * | 2023-07-19 | 2023-10-20 | 上海螣龙科技有限公司 | Network traffic acquisition method and device, electronic equipment and storage medium |
| CN116915667B (en) * | 2023-07-19 | 2024-03-08 | 上海螣龙科技有限公司 | Network traffic acquisition method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101873337A (en) | Zero-copy data capture technology based on rt8169 gigabit net card and Linux operating system | |
| US20210117360A1 (en) | Network and edge acceleration tile (next) architecture | |
| US20210243247A1 (en) | Service mesh offload to network devices | |
| US11736402B2 (en) | Fast data center congestion response based on QoS of VL | |
| EP2256640B1 (en) | Managing traffic on virtualized lanes between a network switch and a virtual machine | |
| CN105706043B (en) | The list handling capacity of push type link | |
| Kapoor et al. | Bullet trains: A study of NIC burst behavior at microsecond timescales | |
| US8369334B2 (en) | Internet real-time deep packet inspection and control device and method | |
| CN105052081B (en) | Communication flows processing framework and method | |
| CN103888386B (en) | The transmission method and device, system of expansible virtual local area network packet | |
| CN108270813B (en) | Heterogeneous multi-protocol stack method, device and system | |
| CN106059821A (en) | Software-defined network (SDN)-based data center service quality assurance method | |
| CN106850565A (en) | A kind of network data transmission method of high speed | |
| Tripathi et al. | Crossbow: from hardware virtualized nics to virtualized networks | |
| Wang et al. | A multipath QUIC scheduler for mobile HTTP/2 | |
| CN103368872A (en) | Data packet forwarding system and method | |
| CN110519180A (en) | Network card virtualization queue scheduling method and system | |
| US9363209B1 (en) | Apparatus, system, and method for resequencing packets | |
| CN110351199A (en) | Flow smoothing method, server and forwarding device | |
| US20230109396A1 (en) | Load balancing and networking policy performance by a packet processing pipeline | |
| Zang et al. | PROP: Using PCIe-based RDMA to accelerate rack-scale communications in data centers | |
| CN108063737A (en) | A kind of FCoE storage area networks read request processing method and system | |
| Landau et al. | Plugging the hypervisor abstraction leaks caused by virtual networking | |
| CN106549815B (en) | Apparatus and method for real-time deep application recognition in a network | |
| US20230043461A1 (en) | Packet processing configurations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20101027 |