CN113965627A - Data sending method, low-delay data receiving method and related device - Google Patents
Data sending method, low-delay data receiving method and related device Download PDFInfo
- Publication number
- CN113965627A CN113965627A CN202010628412.5A CN202010628412A CN113965627A CN 113965627 A CN113965627 A CN 113965627A CN 202010628412 A CN202010628412 A CN 202010628412A CN 113965627 A CN113965627 A CN 113965627A
- Authority
- CN
- China
- Prior art keywords
- data
- rdma
- transmission
- memory
- transmitted
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/03—Protocol definition or specification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/16—Combinations of two or more digital computers each having at least an arithmetic unit, a program unit and a register, e.g. for a simultaneous processing of several programs
- G06F15/163—Interprocessor communication
- G06F15/173—Interprocessor communication using an interconnection network, e.g. matrix, shuffle, pyramid, star, snowflake
- G06F15/17306—Intercommunication techniques
- G06F15/17331—Distributed shared memory [DSM], e.g. remote direct memory access [RDMA]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/06—Notations for structuring of protocol data, e.g. abstract syntax notation one [ASN.1]
Abstract
The invention provides a method for sending data, a method for receiving data with low delay and a related device. The related method for sending data, the method for receiving data, the RDMA low-delay data transmission method based on active fragmentation, the related device, the computer network system and the computer readable storage medium are mainly characterized in that a transmission identifier is added to data requesting cross-end communication at a source host end, and the transmission identifier and the data are spliced in a memory with continuous block addresses; the RDMA cross-end transmission is executed through the active fragment by taking the attached transmission identification data as an RDMA transmission object, and the transmission of the fragment comprising the transmission identification is controlled; and analyzing the data with the transmission identification at the target host to obtain the transmitted data, and informing the target host of the access or call of the related application by taking the received transmission identification as a data receiving completion signal, thereby realizing timely and effective data communication.
Description
Technical Field
The present invention relates to the field of computer communication technologies, and in particular, to a method for transmitting data, a method for receiving data with low latency, and a related apparatus.
Background
With the development of electronic information technology, more and more electronic computer services are widely provided for the production and living activities of people. However, the ability of a single computer device to provide services is limited; more and more computer services, especially powerful computer services, are now implemented by computer clusters formed by a plurality of computer devices. Different computer devices are connected through a network to form a server cluster, and then various computer services are provided for people. When the service is provided to the outside, the calling and data interaction between the cross-computer equipment terminals, namely the cross-end data communication between the computers, is not required. The cross-end communication between computer devices is usually realized under the support of TCP/IP protocol.
However, in the conventional TCP/IP technology, a large amount of server resources and memory bus bandwidth are required to be occupied by an operating system and other software layers in the process of processing data packets, and data is copied and moved among a system memory, a processor cache and a network controller cache, which causes a heavy burden on a CPU and a memory of a server. Especially the severe "mismatch" of network bandwidth, processor speed and memory bandwidth, exacerbates the network latency effect.
RDMA is a remote direct data access technology. As a remote memory access technique, RDMA enables a computer to directly access the memory of other computers without processing by a processor. RDMA moves data quickly from one system to the memory of a remote system without any impact on the operating system. Due to the above advantages of RDMA communication, in more and more application scenarios, especially in the field of big data and AI, when a large amount of data remote communication requests are generated, the RDMA protocol is more and more selected to perform cross-end data transmission.
However, in order to reduce retransmission and improve communication efficiency, like the communication framework based on TCP/IP, the RDMA-based cross-end communication scheme also sets a network link MTU (Maximum Transmission Unit), and introduces a (MTU) fragmentation mechanism for data whose length exceeds the link layer MTU to improve Transmission efficiency and enhance the robustness of the communication system. However, unlike TCP/IP communication, RDMA communication bypasses the kernel of the operating system, and lacks a mechanism for timely notification after sending and receiving data packets as perfect as TCP/IP communication; that is, although in each basic semantic operation layer, the RDMA communication has a set of control message interaction mechanism at the sending and receiving ends to ensure the monitoring of the communication progress of the data segment as the operation object; however, at this granularity, even if the destination host receives all the data fragments, the destination host may not be able to be effectively notified of other related applications in a timely manner.
Disclosure of Invention
In view of the above, the present invention provides a method for transmitting data, a method for receiving data with low latency, and a related device.
In one aspect, an embodiment of the present invention provides a method for sending data, where the method is applied to a source host in a cross-end communication system of a computer, and is particularly applied to RDMA data transmission, and the method is capable of generating an RDMA packet and sending the RDMA packet.
The method for sending data comprises the following steps:
for data requesting cross-end transmission, if the length of the data exceeds a set MTU, adding a transmission identifier for the data, and splicing the transmission identifier and the data in a memory with continuous block addresses;
registering the memory corresponding to the transmission identification data as an RDMA memory; generating a corresponding RDMA data message; and sending the RDMA data message; wherein the content of the first and second substances,
before the RDMA data message is generated, dividing the RDMA memory buffer area corresponding to the attached transmission identification data into a plurality of attached transmission identification data fragments smaller than an MTU threshold value according to the memory offset and a set MTU; generating corresponding RDMA data messages corresponding to each fragment; and controlling the RDMA data message including the fragment of the transmission identification content to be finally transmitted when the RDMA data message is transmitted.
In embodiments of the first aspect, some of them are preferably implemented such that, when generating an RDMA packet, one of the RDMA data packets can include all data (content) of the transmission identification fragment. The transmission identifier, as an identifier of the transmission progress, also includes a transmission identifier fragment form in the phase of generating the RDMA packet, and the above functions can be implemented without occupying too many bytes. Therefore, by controlling the setting of the length of the transmission identifier, one RDMA data message comprises all data contents of the transmission identifier fragment, and gain effects are obtained in multiple aspects rather than on the aspect of reducing a few RDMA message sending, so that the corresponding receiving speed and the overall communication efficiency are greatly improved.
In some embodiments of the first aspect, as yet another possible implementation manner, a method for sending data includes: applying for a memory with continuous block address through a DMA engine to serve as the continuous address memory, copying the data to the memory through the DMA engine, and attaching the transmission identifier, namely splicing the transmission identifier and the data.
In some embodiments of the first aspect, as yet another possible implementation manner, a method for sending data includes: the RDMA data transmission is single-side RDMA; namely, the basic semantic operation is an RDMA read operation or an RDMA write operation, so as to reduce message transmission and interaction between two communication ends and provide overall communication efficiency.
On the other hand, an embodiment of the present invention provides a method for receiving data, which is applied to a destination host in a cross-end communication system of a computer, and is particularly applied to RDMA data transmission, and the method is capable of receiving an RDMA packet, obtaining transmitted data from the RDMA packet, and notifying related applications of the destination host in time based on a transmission identifier.
With reference to the first aspect, correspondingly, the method for receiving data includes:
receiving data sent by a source host end; specifically, all RDMA data messages are received, and complete transmitted data are obtained from the RDMA data messages; the said complete transmitted data, corresponding to the sending data of the source host, if its length exceeds the set MTU, is the attached transmission identification data sent from the source host correspondingly, including: transmitting the identification and the transmitted data; and if the received whole fragments including the transmission identifiers are regarded as the completion of the complete data transmission, removing the attached transmission identifiers, obtaining the transmitted data from the attached transmission identifiers, and informing related applications of accessing/calling.
In some embodiments of the second aspect, as a possible implementation manner, preferably, when one RDMA datagram includes all contents of the transmission identifier, then when the RDMA datagram including the fragment of the transmission identifier is received, that is, regarded as completion of the complete data transmission, the transmitted data is directly acquired from other RDMA datagrams, and the related application is notified of the access/call, so as to improve the overall communication efficiency.
In another aspect, an embodiment of the present invention provides an RDMA low-latency data transmission method based on active fragmentation, where the method is applied to RDMA data transmission of a cross-end communication system of a computer, where the system includes a source host-end device and a destination host-end device; the source host end device is used for sending data, and the destination host end device is used for receiving data.
With reference to the first and second aspects, the RDMA-based active fragmentation-based RDMA low latency data transmission method includes:
for the data requested to be transmitted across the source host side, the method for sending the data according to the first aspect is executed on the source host side, and the method for receiving the data according to the second aspect is executed on the destination host side.
In yet another aspect, an embodiment of the present invention provides an apparatus for RDMA data transfer.
In combination with the above aspects, the above apparatus includes: an RDMA network card, a processor, and a memory; wherein the memory stores a computer program operable on the processor; and the processor, when executing the computer program, can perform the method for sending data according to the first aspect together with the RDMA network card,
and/or the method of receiving data as set forth in the second aspect.
In yet another aspect, an embodiment of the present invention provides a computer network system.
In combination with the above aspects, the computer network system described above includes:
a plurality of apparatuses (not less than two) according to the fourth aspect; the devices are connected to each other through an RDMA network; these devices are used for either sending or receiving RDMA datagrams, or both.
In yet another aspect, an embodiment of the present invention provides a computer-readable storage medium.
The computer-readable storage medium stores code for performing tag data transmission, including: instructions for performing the method operations of sending data as described in the first aspect, and/or instructions for the method operations of receiving data as described in the second aspect, or instructions for the RDMA low-latency data transfer method operations based on active fragmentation as described in the third aspect.
The method for sending data, the method for receiving data, and the RDMA low-latency data transmission method based on active fragmentation, and the related device, computer network system and computer readable storage medium provided by the above embodiments mainly add a transmission identifier to data (data length exceeding a set MTU) requesting cross-end communication at a source host side, and splice the transmission identifier and the data in a memory with continuous block addresses; the RDMA cross-end transmission is executed through the active fragment by taking the attached transmission identification data as an RDMA transmission object, and the transmission of the fragment comprising the transmission identification is controlled; and analyzing the data with the transmission identification at the target host to obtain the transmitted data, and informing the target host of the access or call of the related application by taking the received transmission identification as a data receiving completion signal, thereby realizing timely and effective data communication.
The technical solution of the present invention is further described with reference to the accompanying drawings and specific embodiments.
Drawings
To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings related to a part of the embodiments of the present invention or the description in the prior art will be briefly introduced below.
Fig. 1 is a flowchart illustrating an RDMA low-latency data transmission method based on active fragmentation according to a preferred embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention is clearly and completely described below with reference to the drawings of the embodiments of the present invention. It is to be understood that the described embodiments are merely exemplary of a portion of the invention and not all embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
In the prior art, a typical TCP/IP protocol-based computer cross-end communication framework has the advantages of mature technical scheme, rich solution, reliability, easy use, and the like, but in the face of large data cross-end transmission, a communication bottleneck is formed during large data communication, especially during communication peak due to the inherent defects of the initial design of the TCP/IP protocol-based communication framework, and the overall system efficiency is affected. While RDMA, as a remote memory access technique, enables a computer to directly access the memory of other computers without the need for processing by a processor, the above-described problems are overcome to some extent. However, in order to reduce retransmission and improve communication efficiency, like the communication framework based on TCP/IP, the RDMA-based cross-end communication scheme also sets the network link MTU, and performs MTU fragmentation for data whose length exceeds the link layer MTU. However, RDMA communication bypasses the kernel of the operating system, and lacks a perfect mechanism for transmitting and receiving data messages and notifying in time after receiving the data messages as well as TCP/IP communication; that is, even if the destination host receives all data fragments, the destination host may not be able to timely and effectively notify other related applications.
Therefore, the present invention provides a method for sending data, a method for receiving data with low latency, and a related device, so as to achieve that longer data can be timely and effectively transmitted when the longer data is based on RDMA communication.
The following are some preferred embodiments of the invention. Wherein the content of the first and second substances,
some preferred embodiments of the present invention provide a method for sending data, which is applied to a source host side in a cross-terminal communication system of a computer, in particular to RDMA data transmission, and which is capable of generating an RDMA packet and sending the RDMA packet. The method comprises the steps of adding a transmission identifier to data (the data length of which exceeds a set MTU) requesting cross-end transmission, and splicing the transmission identifier and the data into a memory with continuous block addresses; and the RDMA cross-end transmission is executed by taking the data with the transmission identification as an RDMA transmission object, organized and purposeful transmission is realized by actively fragmenting and controlling the transmission of the fragments comprising the transmission identification, and the timely communication of the transmitted data is realized by combining the target host end.
When a transmission identifier (wherein the data length exceeds the set MTU) is attached to the data requesting communication, the implementation process is as follows:
applying for a free memory with continuous block address, generating a transmission identifier, copying data into the memory and splicing the data with the transmission identifier; then registering the memory corresponding to the transmission identifier and the transmission identifier attached data spliced by the data as an RDMA memory; generating a corresponding RDMA data message; and sending the RDMA data message; wherein the content of the first and second substances,
before the RDMA data message is generated, dividing the RDMA memory buffer zone corresponding to the attached transmission identification data into a plurality of attached transmission identification data fragments smaller than an MTU threshold value according to the memory offset; generating corresponding RDMA data messages corresponding to each fragment;
and controlling the RDMA data message including the fragment of the transmission identification content to be finally transmitted when the RDMA data message is transmitted.
In some of the above preferred embodiments, there is further provided a preferred implementation manner, specifically including: when the RDMA message is generated, one RDMA data message comprises the whole content of the transmission identification fragment. As the identifier of the transmission progress, the transmission identifier can completely realize the above function without occupying too many bytes. Therefore, by controlling the setting of the length of the transmission identifier, one RDMA data message comprises all data contents of the transmission identifier fragment, and gain effects are obtained in multiple aspects rather than on the aspect of reducing a few RDMA message sending, so that the corresponding receiving speed and the overall communication efficiency are greatly improved.
In some of the above preferred embodiments, there is further provided a preferred implementation manner, specifically including: and applying for a free memory by the DMA engine to serve as the address continuous memory, copying data to the memory by the DMA engine, and attaching a transmission identifier, specifically, splicing the free memory and the DMA engine in the memory, namely copying the free memory and the DMA engine to adjacent positions.
In some of the above preferred embodiments, the RDMA data transfer, among others, is single-sided RDMA; the basic semantic operations of single-sided RDMA generally include: two basic voice operation modes of RDMA read operation and RDMA read operation; therefore, the RDMA data transmission can adopt RDMA read operation and RDMA write operation. Compared with the double-side RDMA, in an effective communication process, the single-side RDMA can obviously reduce the message interaction times of two communication ends, and further improve the overall communication efficiency.
Other preferred embodiments of the present invention provide a method for receiving data, which is applied to a destination host in a cross-end computer communication system, and is particularly applied to RDMA data transmission, and is capable of receiving an RDMA packet, obtaining transmitted data therefrom, and notifying a related application at the destination host in time based on a transmission identifier. The method obtains the transmitted data by analyzing the data attached with the transmission identification at the target host, and informs the target host of the access or the call of the related application by taking the received transmission identification as a data receiving completion signal so as to realize the timely and effective transmission of the transmitted data. The data and the transmission identification are received by the destination host end in the form of data with transmission identification and then stored in the memory with continuous addresses; receiving all RDMA data messages, and obtaining complete transmitted data from the RDMA data messages; the said complete transmitted data, corresponding to the sending data of the source host, if its length exceeds the set MTU, is the transmission identification data sent by the source host, including: transmitting the identification and the transmitted data;
removing the transmission identification to obtain the transmitted data;
and acquiring the transmitted data and informing related applications to access or call by taking the received fragment comprising the transmission identifier as the completion of the complete data transmission.
In some of the foregoing preferred embodiments, a preferred implementation manner is further provided, when a transmission identifier size is transmittable through only one RDMA packet, that is, one RDMA data packet includes all contents of a transmission identifier fragment, then when an RDMA data packet including a transmission identifier fragment is received, directly acquiring transmitted data from other RDMA data packets including contents of the data fragment, and notifying an associated application of access/call, so as to improve overall communication efficiency.
Still other preferred embodiments of the present invention provide an RDMA low latency data transmission method based on active fragmentation, which is particularly applied to RDMA data transmission, and the system includes a source host-side device and a destination host-side device; the source host end device is used for sending data, and the destination host end device is used for receiving data. The method comprises the following steps:
for data that the source host side requests for cross-side transmission,
the method for transmitting data described in any of the above embodiments is performed at the source host,
and correspondingly, the method for receiving data described in any of the above embodiments is executed on the destination host.
Fig. 1 is a flow chart of a RDMA-based low-latency data transmission method provided in one of the above preferred embodiments. As shown in fig. 1, includes:
s101, adding a transmission identifier for the transmitted data:
applying for a continuous idle memory of a block address; generating a transmission identifier, copying the data into a memory, and splicing the data with the transmission identifier;
s102, registering the memory (namely, the memory corresponding to the transmission identification data) as an RDMA memory;
s103, generating a corresponding RDMA data message;
s104, sending the RDMA data message;
after registering the memory buffer area, dividing the RDMA memory buffer area corresponding to the whole transmission identification data into a plurality of transmission identification data fragments smaller than the MTU threshold value according to the memory offset; correspondingly generating a corresponding RDMA data message for each fragment; and controlling the RDMA data message including the fragment of the transmission identification content to be finally transmitted when the RDMA data message is transmitted.
At the end of the destination host computer,
s201, receiving all RDMA data messages to obtain transmission identification data;
s202, the data attached with the transmission identification is analyzed to obtain the transmitted data, and the process is just to remove the transmission identification in the data.
And in the whole receiving process, the received fragment including the transmission identifier is regarded as the whole transmission process is completed, and the related application is informed to access or call the transmitted data in time.
Still further preferred embodiments of the present invention provide an apparatus comprising: an RDMA network card, a processor, and a memory; wherein the memory stores a computer program operable on the processor; and when the processor executes the computer program, the processor and the RDMA network card can jointly execute the method for sending data and/or the method for receiving data according to any one of the above embodiments.
Still other preferred embodiments of the present invention provide a computer network system, comprising: a plurality of (not less than two) devices as described in any of the above embodiments; and they are connected to each other through RDMA networks; these devices are used either for sending RDMA datagrams (i.e., as the source host side), or for receiving RDMA datagrams (i.e., as the destination host side), or for sending RDMA datagrams and receiving RDMA datagrams from other devices.
Still other preferred embodiments of the present invention provide a computer-readable storage medium having stored therein code for performing tag data transmission, comprising: instructions for performing the method operations for sending data according to any of the above embodiments, and/or instructions for receiving data according to any of the above embodiments, or instructions for the RDMA low-latency data transfer method operations according to any of the above embodiments.
It should be noted that the memory mentioned in the above embodiments includes, but is not limited to, a memory in a narrow sense (i.e. a memory provided by a memory card in a computer system), a memory of a video card (often referred to as a video memory for short), and the like. Firstly, use
The above description is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto.
Claims (10)
1. A method for sending data for RDMA data transfer at a source host, the method comprising:
for any data requesting cross-end transmission, if the length of the data exceeds a set MTU, adding a transmission identifier for the data, and splicing the identifier and the data in a memory with continuous block addresses;
registering the memory corresponding to the transmission identification data as an RDMA memory; generating an RDMA data message; and sending the RDMA data message; wherein the content of the first and second substances,
before the RDMA data message is generated, dividing the RDMA memory buffer zone corresponding to the transmission identification data into a plurality of transmission identification data fragments smaller than an MTU threshold value according to the memory offset; generating corresponding RDMA data messages corresponding to each fragment;
and controlling the RDMA data message including the fragment of the transmission identification content to be finally transmitted when the RDMA data message is transmitted.
2. The method of claim 1, wherein the method comprises:
and one RDMA data message comprises the whole content of the transmission identification fragment.
3. The method of claim 1, wherein the method comprises:
and applying for the memory with continuous addresses through a DMA engine, and copying the data to the memory to be spliced with the additional transmission identifier.
4. The method of claim 1, wherein the method comprises:
the RDMA data transmission is single-side RDMA;
which comprises the following steps: a read operation or a write operation.
5. A method of receiving data for use in RDMA data transfer at a destination host, the method comprising:
with respect to any one of the received data,
receiving all RDMA data messages, and acquiring complete transmitted data from the RDMA data messages; wherein, if the length of the said complete transmitted data exceeds the set MTU, it will attach the transmission identification data, including: transmitting the identification and the transmitted data;
and if the received whole fragments including the transmission identifiers are regarded as the completion of the complete data transmission, removing the attached transmission identifiers, obtaining the transmitted data from the attached transmission identifiers, and informing related applications of accessing/calling.
6. The method of claim 5, wherein the method comprises:
if one RDMA data message comprises all data of the transmission identification fragment;
the RDMA data message including the transmission identification is received as the completion of the complete data transmission; and directly acquiring the transmitted data from other RDMA data messages and informing related application access/call.
7. An RDMA low-latency data transmission method based on active fragmentation, which is applied to RDMA data transmission of a cross-end communication system of computers, wherein the system comprises a source host-end device and a destination host-end device, and the method comprises the following steps:
for any data requesting to be transmitted across terminals, if the length of the data exceeds the set MTU, the data is transmitted across terminals
Method for transmitting data according to any of claims 1 to 4, carried out at the source host end
And at the destination host side, performing the method of receiving data as claimed in any one of claims 5 to 6.
8. An apparatus for RDMA data transfer, comprising:
an RDMA network card, a processor, and a memory; wherein the content of the first and second substances,
the memory stores a computer program operable on the processor;
the processor, when executing the computer program, is capable of performing any of the methods of transmitting data described in 1-4 in conjunction with the RDMA network card,
and/or a method of receiving data as claimed in any of claims 5 to 6.
9. A computer network system, comprising:
a plurality of the apparatus of claim 8; the devices are connected to each other through an RDMA network;
the apparatus is for the sending of RDMA data packets,
and/or, reception of RDMA data packets.
10. A computer-readable storage medium having stored thereon code for performing data transmission, comprising: instructions for performing the method operations of transmitting data of any of claims 1-4;
and/or instructions to perform the method operations of receiving data of any of claims 5-6;
or, the instructions of claim 7 to operate in an active fragmentation based RDMA low latency data transfer method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010628412.5A CN113965627A (en) | 2020-07-02 | 2020-07-02 | Data sending method, low-delay data receiving method and related device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010628412.5A CN113965627A (en) | 2020-07-02 | 2020-07-02 | Data sending method, low-delay data receiving method and related device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113965627A true CN113965627A (en) | 2022-01-21 |
Family
ID=79459354
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010628412.5A Pending CN113965627A (en) | 2020-07-02 | 2020-07-02 | Data sending method, low-delay data receiving method and related device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113965627A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893323A (en) * | 2016-05-23 | 2016-08-24 | 华为技术有限公司 | Data reading method and data reading equipment |
| CN107508828A (en) * | 2017-09-18 | 2017-12-22 | 南京斯坦德云科技股份有限公司 | A kind of very-long-range data interaction system and method |
| US20180307650A1 (en) * | 2017-04-25 | 2018-10-25 | Samsung Electronics Co., Ltd. | Low latency direct access block storage in nvme-of ethernet ssd |
| CN110018897A (en) * | 2018-01-09 | 2019-07-16 | 阿里巴巴集团控股有限公司 | Data processing method, device and calculating equipment |
| CN110177085A (en) * | 2019-04-26 | 2019-08-27 | 华中科技大学 | Across host dynamic stain method for tracing and system suitable for cloud environment |
-
2020
- 2020-07-02 CN CN202010628412.5A patent/CN113965627A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893323A (en) * | 2016-05-23 | 2016-08-24 | 华为技术有限公司 | Data reading method and data reading equipment |
| US20180307650A1 (en) * | 2017-04-25 | 2018-10-25 | Samsung Electronics Co., Ltd. | Low latency direct access block storage in nvme-of ethernet ssd |
| CN107508828A (en) * | 2017-09-18 | 2017-12-22 | 南京斯坦德云科技股份有限公司 | A kind of very-long-range data interaction system and method |
| CN110018897A (en) * | 2018-01-09 | 2019-07-16 | 阿里巴巴集团控股有限公司 | Data processing method, device and calculating equipment |
| CN110177085A (en) * | 2019-04-26 | 2019-08-27 | 华中科技大学 | Across host dynamic stain method for tracing and system suitable for cloud environment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7571247B2 (en) | Efficient send socket call handling by a transport layer | |
| JP4504977B2 (en) | Data processing for TCP connection using offload unit | |
| US7685287B2 (en) | Method and system for layering an infinite request/reply data stream on finite, unidirectional, time-limited transports | |
| US20150127849A1 (en) | Tcp data transmission method, tcp offload engine, and system | |
| WO2009097776A1 (en) | System, device and method for achieving service upgrade | |
| US8539089B2 (en) | System and method for vertical perimeter protection | |
| WO2024037296A1 (en) | Protocol family-based quic data transmission method and device | |
| US20230152978A1 (en) | Data Access Method and Related Device | |
| CN109120687B (en) | Data packet transmitting method, device, system, equipment and storage medium | |
| US9344363B2 (en) | Information processing system, relay device, information processing device, and information processing method | |
| WO2016095323A1 (en) | Data transmission method and apparatus based on serial server | |
| US20090157896A1 (en) | Tcp offload engine apparatus and method for system call processing for static file transmission | |
| CN111698275A (en) | Data processing method, device and equipment | |
| CN111404842B (en) | Data transmission method, device and computer storage medium | |
| CN112152940B (en) | Message processing method, device and system | |
| CN111698274B (en) | Data processing method and device | |
| CN111404986A (en) | Data transmission processing method, device and storage medium | |
| CN112165449A (en) | Control method of real-time authority of web application, electronic device and storage medium | |
| CN113965627A (en) | Data sending method, low-delay data receiving method and related device | |
| US20220286532A1 (en) | Method and apparatus for obtaining shared maximum segment size mss | |
| CN113886294A (en) | RDMA (remote direct memory Access) -based low-delay data transmission method and related device | |
| WO2019015487A1 (en) | Data retransmission method, rlc entity and mac entity | |
| WO2022151475A1 (en) | Message buffering method, memory allocator, and message forwarding system | |
| CN109688085B (en) | Transmission control protocol proxy method, storage medium and server | |
| US20050135394A1 (en) | Use of internal buffer to reduce acknowledgement related delays in acknowledgement-based reliable communication protocols |
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 |