CN107743102B - efficient tcp session recombination method - Google Patents
efficient tcp session recombination method Download PDFInfo
- Publication number
- CN107743102B CN107743102B CN201711052422.3A CN201711052422A CN107743102B CN 107743102 B CN107743102 B CN 107743102B CN 201711052422 A CN201711052422 A CN 201711052422A CN 107743102 B CN107743102 B CN 107743102B
- Authority
- CN
- China
- Prior art keywords
- data
- data block
- session
- offset
- message
- 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.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L49/00—Packet switching elements
- H04L49/90—Buffering arrangements
- H04L49/9057—Arrangements for supporting packet reassembly or resequencing
-
- 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
- H04L69/16—Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
- H04L69/163—In-band adaptation of TCP data exchange; In-band control procedures
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
high-efficiency TCP session restructuring method relates to the information technology field, and the invention is completed by the flow of the session buffer process, the total flow of data restructuring, and the data block numbering calculation flow, compared with the prior message restructuring scheme, the invention directly organizes the messages into an ordered data structure through the B + tree, which can ensure that the restructuring efficiency is not reduced due to frequent data searching when the disordered messages are more, the improved algorithm can obviously improve the efficiency and the overall performance of the restructuring system under the scene of more disordered messages, and simultaneously, the scheme simplifies the comparison flow of SEQ and the restructuring process of the TCP session, so that the TCP session restructuring process and the software implementation become simple and high-efficiency.
Description
Technical Field
The invention relates to the technical field of information, in particular to the technical field of communication processing in the field of data communication.
Background
The TCP/IP protocol is widely applied to network data communication at present , when data is transmitted by using the TCP/IP protocol on the network, the data needs to be split into a plurality of messages to be transmitted, and the recombination technology of the TCP session is used in the fields of network security, DPI data acquisition, internet behavior audit and the like.
Mention is made in patent application document No. CN 201510705482.5: when a TCP message is received, inquiring a corresponding session table item according to an IP address and a port which are included in the TCP message; determining the category of the TCP message according to the direction of the TCP message, the sequence number of the TCP message and the sequence number of the TCP message received in sequence in the direction; if the TCP message is a disordered message, adding the TCP message into a disordered message linked list in the direction; if the TCP message is a sequential message, judging whether a linked list node matched with the TCP message exists in the disordered linked list of the direction; if yes, adding the linked list node and the TCP message into the sequential message linked list in the direction; otherwise, adding the TCP message into the sequential message linked list of the direction.
The scheme with the application number of CN201510705482.5 has the problems that disordered nodes stored by using a linked list can bring disordered node searching tasks, a large number of disordered node traversal searching tasks exist under the condition that disordered messages are more, excessive traversal searching can seriously affect the message recombination efficiency, in addition, the problems of complex memory management efficiency and memory hit can be caused due to the fact that the sizes of the messages are not fixed when the disordered messages are stored, and the memory with excessive application sizes not less than can easily cause memory fragmentation and reduce the recombination efficiency.
The technical scheme realizes simple and efficient TCP message recombination methods aiming at the common low efficiency problem in the TCP session recombination process, and the TCP messages are organized into data blocks with fixed size and then organized and managed in the structural form of a B + tree, so that the recombination efficiency and success rate of the TCP sessions are improved, and a high-performance TCP session recombination algorithm and a high-performance software module are realized.
The transmission principle of TCP UDP and the difference of data block, section, grouping, datagram and data packet:
with regard to the TCP output:
sending buffers are arranged on each TCP socket, when an application program calls write, a kernel copies data of the buffers of the application program to the sending buffers of the TCP, the TCP sends data blocks to an IP with the size of MSS (maximum segment size) or smaller, each data block is added with a TCP header to form a section, the IP is used for mounting the IP header to each section to form a data packet, a target IP address and a routing table item are searched to determine an outgoing interface, then the data packet is transmitted to a data link, a link output queue is added, if the queue is full, the packet is discarded, an error is returned upwards through a protocol stack, the link layer is connected to the IP layer, the IP layer is connected to the TCP layer, the TCP error is retransmitted at a certain time, when the opposite server receives the data, the local server sends ACK, receives the ACK, responds to the ACK, the data of the sending buffers are discarded, and other data can be written into the buffers again.
Regarding the UDP output:
UDP, because it is unreliable, does not have to keep a copy of the application process's data, so there is no send buffer, UDP installs an 8 byte UDP header to each data block, forms a datagram and passes to IP, IP installs an IP header to the datagram to form a packet, if the packet does not fit in MTU, fragmentation is performed, and then each fragment is added to the output queue of the data link layer. The write return of the application indicates that data has been added to the output queue. Note also that UDP has no MSS, so fragmentation for packets occurs more in UDP than TCP.
Disclosure of Invention
The invention provides efficient tcp session recombination devices based on the deficiency of the prior art, which consists of a data cache module and a data recombination module, wherein the data cache module consists of a session query module, a session starting offset recording module, a session ending offset recording module, a data block numbering calculation recording module and a B + tree generation module;
the session query module is used for querying a corresponding session table item according to the IP address and the port included in the TCP message when the TCP message is received; the session table entry comprises a session start offset, a session end marker and a B + tree structure of the storage data block;
the session initial offset recording module is used for creating a session table item according to the IP and port information of the TCP message when a first packet of the TCP message is received, recording the session initial offset and finishing the offset; the value of the start offset is SEQ value +1 of the first packet, and the end offset is initialized to be the same as the start offset; the judgment of the first packet is based on whether the message type is SYN or SYN + ACK message;
the session ending offset recording module is used for updating ending offset in the session information when the received TCP message is the last packets of the TCP message, wherein the ending offset is the SEQ value of the message;
the data block numbering calculation and recording module is used for calculating the initial offset, the ending offset and the data block numbering of a data block corresponding to a current data packet according to the SEQ value and the load length of the current TCP message and the session initial offset and the ending offset stored in the session table entry when the TCP message is received; the calculation method of the starting offset value and the ending offset value comprises the following steps: if the SEQ value of the current TCP message is smaller than the session starting offset value, the starting offset value is the maximum value of the SEQ value of the current TCP message plus a 32-bit unsigned integer, otherwise, the starting offset value is the SEQ value of the current TCP message, and the ending offset value is the starting offset value plus the packet length;
the B + tree generation module has the functions of:
1. when a TCP message is received, searching a corresponding data block from a B + tree according to the number of the data block of the TCP message just received;
2. when the data block of the TCP message just received is not in the B + tree, applying for a data block memory, initializing a data block number, and inserting the data block of the TCP message just received into the B + tree as a new application data block;
3. when the data block of the TCP message just received is in the B + tree, updating the data block found in the B + tree, specifically copying the data in the data block of the TCP message just received to the specified position of the found data block, and updating the use mark of the data block;
the data reorganization module is used for processing when a TCP session is ended, when a TCP session ending mark is detected, the data of the ordered data block cached in the B + tree is taken out for data reorganization, only the B + tree in the session table entry needs to be traversed during reorganization, the data in the corresponding block structure is output, when the B + tree in the session table entry is traversed, only the data use mark between the session starting offset and the session ending offset needs to be checked, if the data use marks are marked to be valid, the data in the corresponding block structure is output to complete session reorganization, otherwise, the session is regarded as a data incomplete session, data reorganization is abandoned, and then the memory block of each node of the B + tree is recovered.
The invention also comprises efficient tcp session restructuring methods, which are completed by a session caching step, a data restructuring step and a data block numbering step;
(1) session caching step
1) After receiving the TCP message, inquiring corresponding session table item information according to the IP and the port;
2) calculating the initial offset, the ending offset and the data block number of a data block corresponding to the current data packet according to the SEQ value and the load length of the current TCP message and the session initial offset and ending offset stored in the session table entry;
the calculation method of the starting offset value and the ending offset value comprises the following steps: if the SEQ value of the current message is smaller than the session initial offset value, the initial offset is the maximum value of the SEQ value of the current message plus a 32-bit unsigned integer, otherwise, the initial offset value is the SEQ value of the current message; the end offset should be the start offset value plus the packet length;
3) the data block information comprises the start offset and the end offset of the data block and the use state of the data block, wherein the use state of the data block is data segments in the data block, byte information is represented by each bits, and if the byte of the corresponding data bit is filled, the byte is set to be 1 to represent the filling;
4) when the data block can not be found, applying for a new data block memory, initializing the number of the data block, shifting the start and the end of the data block, and inserting the data block into a B + tree;
the data block start offset is the number of the data block multiplied by the size of the data block, and the data block end offset is the data block start offset plus the size of the data block;
5) updating the data of the searched data block and the data on the newly inserted data block, and synchronously updating the data block use mark;
the specific updating method comprises the following steps: copying the data of the data packet corresponding to the data block to the data storage position of the data block to be updated, updating the use mark of the data block, and marking the corresponding data position as 1;
(2) data reorganization step
1) Receiving TCP messages from a driver or other message capture device;
2) judging whether the message is the first packet of an uplink or downlink message, if the message is a SYN or SYN + ACK message, creating a session table item according to the IP and port information of the message, and recording the initial offset and the ending offset of the session;
the values of the starting offset and the ending offset are SEQ values +1 of the first packet;
the session table entry comprises a B + tree structure required by the session data caching process, session starting offset, session ending offset and a session ending mark;
wherein the B + tree structure will use the data block number as KEY;
3) judging whether the received message is the last packets of the session, if so, updating the ending deviation in the session information, wherein the ending deviation value is the SEQ value of the message;
meanwhile, a session ending mark is set in the session table item;
4) executing message data caching operation on the data message;
5) checking whether a session end mark exists or not, and if so, executing data recombination;
during recombination, only the B + tree in the session table entry needs to be traversed, and the data in the corresponding block structure is output;
confirming whether a data block is lost or not according to the use state mark in the data block in the recombination process, and confirming the end of the recombined data according to the initial offset and the end offset of the session;
if all data between the session starting offset and the session ending offset are recombined and no data block is lost, the recombination is completed, otherwise, the session fails, the recombination cannot be completed, and the recombined data is discarded;
(3) data block numbering step
1) Initializing the current offset value as a data packet starting offset value;
2) dividing the initial offset value by a fixed size value of a data block defined in the system to obtain a quotient which is used as the number of the current data offset data block;
3) and the end offset value of the data block corresponding to the current data block number < = the data packet represents that the current data packet belongs to more data blocks, the current offset value is updated to be +1, and the next data block numbers are calculated in an iterative manner.
Advantageous effects
Compared with the prior message reassembly scheme, the scheme directly organizes the messages into an ordered data structure through the B + tree, can ensure that the reassembly efficiency is not reduced due to frequent data search when the out-of-order messages are more, can obviously improve the efficiency and the overall performance of the reassembly system under the scene of more out-of-order messages, simplifies the comparison process of SEQ and the reassembly process of TCP sessions, and enables the TCP session reassembly process and software implementation to be simple and efficient.
Drawings
FIG. 1 is a system architecture diagram of the present invention;
FIG. 2 is a flow chart of a session caching process of the present invention;
FIG. 3 is an overall flow chart of the data reorganization of the present invention;
fig. 4 is a flow chart of the data block numbering calculation of the present invention.
Detailed Description
Example
The efficient tcp session reorganization devices realized by referring to fig. 1 comprise a data cache module A and a data reorganization module B, wherein the data cache module A consists of a session query module 11, a session starting offset recording module 12, a session ending offset recording module 13, a data block number calculating recording module 14 and a B + tree generating module 15;
the session query module 11 is used for querying a corresponding session entry according to the IP address and the port included in the TCP message 10 when receiving the TCP message 10; the session table entry comprises a session start offset, a session end marker and a B + tree structure of the storage data block;
the session initiation offset recording module 12 is used for creating a session table entry according to the IP and port information of the TCP packet 10, recording the session initiation offset, and ending the offset when receiving the first packet of the TCP packet 10; the value of the start offset is SEQ value +1 of the first packet, and the end offset is initialized to be the same as the start offset; the judgment of the first packet is based on whether the message type is SYN or SYN + ACK message;
the session ending offset recording module 13 is used for updating ending offset in the session information when the received TCP message 10 is the last packets of the TCP message 10 of this time, and the ending offset value is the SEQ value of the message;
the data block numbering and recording module 14 is used for calculating the starting offset, the ending offset and the data block numbering of the data block corresponding to the current data packet according to the SEQ value and the load length of the current TCP message 10 and the session starting offset and ending offset stored in the session table entry when the TCP message 10 is received; the calculation method of the starting offset value and the ending offset value comprises the following steps: if the SEQ value of the current TCP packet is smaller than the session start offset value, the start offset should be the maximum value of the SEQ value of the current TCP packet 10 plus a 32-bit unsigned integer, otherwise, the start offset should be the SEQ value of the current TCP packet 10, and the end offset should be the start offset plus the packet length;
the B + tree generation module 15 functions to:
1. when a TCP message 10 is received, searching a corresponding data block from a B + tree according to the number of the data block of the TCP message 10 just received;
2. when the data block of the TCP message 10 just received is not in the B + tree, applying for a data block memory, initializing a data block number, and inserting the data block of the TCP message 10 just received into the B + tree as a new application data block;
3. when the data block of the TCP packet 10 just received is in the B + tree, updating the data block found in the B + tree, specifically copying the data in the data block of the TCP packet 10 just received to the specified position of the found data block, and updating the data block usage flag;
the data reorganization module B is used for processing when the TCP session is ended, when a TCP session ending mark is detected, the data of the ordered data block cached in the B + tree is taken out for data reorganization, only the B + tree in the session table entry needs to be traversed during reorganization, the data in the corresponding block structure is output, when the B + tree in the session table entry is traversed, only the data use mark between the session starting offset and the session ending offset needs to be checked, if the data use marks are marked to be valid, the data in the corresponding block structure is output to complete session reorganization, otherwise, the session is regarded as a data incomplete session, data reorganization is abandoned, and then the memory block of each node of the B + tree is recovered.
Example two
Referring to fig. 3, the steps of the overall flow of data reorganization of the present invention are as follows:
, receiving TCP message from driver or other message capture device;
step two: judging whether the message is the first packet of the uplink or downlink message, if the message is a SYN or SYN + ACK message, creating a session table item according to the IP and port information of the message, and recording the initial offset and the ending offset of the session. The values of the start offset and the end offset are SEQ values +1 of the first packet. The session table entry contains the B + tree structure required for the session data caching process as well as the session start, end offsets and the session end marker. Wherein the B + tree structure will use the data block number as KEY;
judging whether the received message is the last packets of the session, if so, updating the ending deviation in the session information, wherein the ending deviation is the SEQ value of the message;
step four: executing message data caching operation on the data message;
step five: checking whether a session end mark exists or not, and if so, executing data reorganization. When in recombination, only the B + tree in the conversation table item needs to be traversed, and the data in the corresponding block structure is output. And in the recombination process, whether the data block is lost or not is confirmed according to the use state mark in the data block, and the end of the recombined data is confirmed according to the starting offset and the ending offset of the session. If all data between the session starting offset and the session ending offset are recombined and no data block is lost, the recombination is completed, otherwise, the session fails, the recombination cannot be completed, and the recombined data is discarded.
EXAMPLE III
Referring to fig. 2, the steps of the flow of the session caching process of the present invention are as follows:
, after receiving TCP message, inquiring the corresponding session list item information according to IP and port;
step two: calculating the initial offset, the ending offset and the data block number of a data block corresponding to the current data packet according to the SEQ value and the load length of the current message and the session initial offset and ending offset stored in the session table entry;
the calculation method of the starting offset value and the ending offset value comprises the following steps: if the SEQ value of the current message is smaller than the session initial offset value, the initial offset is the maximum value of the SEQ value of the current message plus a 32-bit unsigned integer, otherwise, the initial offset value is the SEQ value of the current message; the end offset should be the start offset value plus the packet length; the method for calculating the number of the data block is described with reference to fig. 4;
step three:
the data block information comprises a start offset and an end offset of the data block and the use state of the data block, wherein the use state of the data block is data segments in the data block, byte information is represented by each bits, and if the byte of the corresponding data bit is already filled, the byte is set to be 1 to represent filling;
step four:
if the data block can not be found, applying for a new data block memory, initializing the number of the data block, shifting the start and the end of the data block, and inserting the data block into a B + tree;
the data block start offset is the number of the data block multiplied by the size of the data block, and the data block end offset is the data block start offset plus the size of the data block;
step five:
updating the data found or newly inserted into the data block, and synchronously updating the data block use mark;
the specific updating method comprises the following steps: and copying the data of the data packet corresponding to the data block to the data storage position of the data block to be updated, updating the use mark of the data block, and setting the corresponding data position to be 1.
Example four
Referring to fig. 4, the steps of the data block number calculation flow of the present invention are as follows:
step , initializing the current offset value to the packet start offset value;
step two: dividing the initial offset value by a fixed size value of a data block defined in the system to obtain a quotient which is used as the number of the current data offset data block;
and step three, if the data block end offset value < = the end offset value of the data packet corresponding to the current data block number indicates that the current data packet belongs to more data blocks, updating the current offset value to be +1 of the data block end offset value, and iteratively calculating the numbers of the next data blocks.
Claims (1)
1, high-efficiency tcp session restructuring method, characterized in that it is completed by the steps of session buffer, data restructuring, data block numbering;
(1) session caching step
1) After receiving the TCP message, inquiring corresponding session table item information according to the IP and the port;
2) calculating the initial offset, the ending offset and the data block number of a data block corresponding to the current data packet according to the SEQ value and the load length of the current TCP message and the session initial offset and ending offset stored in the session table entry;
the calculation method of the starting offset value and the ending offset value comprises the following steps: if the SEQ value of the current message is smaller than the session initial offset value, the initial offset is the maximum value of the SEQ value of the current message plus a 32-bit unsigned integer, otherwise, the initial offset value is the SEQ value of the current message; the end offset should be the start offset value plus the packet length;
3) the data block information comprises the start offset and the end offset of the data block and the use state of the data block, wherein the use state of the data block is data segments in the data block, byte information is represented by each bits, and the byte of the corresponding data bit is set to be 1 when the byte is filled;
4) when the data block can not be found, applying for a new data block memory, initializing the number of the data block, shifting the start and the end of the data block, and inserting the data block into a B + tree;
the data block start offset is the number of the data block multiplied by the size of the data block, and the data block end offset is the data block start offset plus the size of the data block;
5) updating the data of the searched data block and the data on the newly inserted data block, and synchronously updating the data block use mark;
the specific updating method comprises the following steps: copying the data of the data packet corresponding to the data block to the data storage position of the data block to be updated, updating the use mark of the data block, and marking the corresponding data position as 1;
(2) data reorganization step
1) Receiving TCP messages from a driver or other message capture device;
2) judging whether the message is the first packet of an uplink or downlink message, if the message is a SYN or SYN + ACK message, creating a session table item according to the IP and port information of the message, and recording the initial offset and the ending offset of the session;
the values of the starting offset and the ending offset are SEQ values +1 of the first packet;
the session table entry comprises a B + tree structure required by the session data caching process, session starting offset, session ending offset and a session ending mark;
wherein the B + tree structure will use the data block number as KEY;
3) judging whether the received message is the last packets of the session, if so, updating the ending deviation in the session information, wherein the ending deviation value is the SEQ value of the message;
meanwhile, a session ending mark is set in the session table item;
4) executing message data caching operation on the data message;
5) checking whether a session end mark exists or not, and if so, executing data recombination;
during recombination, only the B + tree in the session table entry needs to be traversed, and the data in the corresponding block structure is output;
confirming whether a data block is lost or not according to the use state mark in the data block in the recombination process, and confirming the end of the recombined data according to the initial offset and the end offset of the session;
if all data between the session starting offset and the session ending offset are recombined and no data block is lost, the recombination is completed, otherwise, the session fails, the recombination cannot be completed, and the recombined data is discarded;
(3) data block numbering step
1) Initializing the current offset value as a data packet starting offset value;
2) dividing the initial offset value by a fixed size value of a data block defined in the system to obtain a quotient which is used as the number of the current data offset data block;
3) and the end offset value of the data block corresponding to the current data block number < = the data packet represents that the current data packet belongs to more data blocks, the current offset value is updated to be +1, and the next data block numbers are calculated in an iterative manner.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711052422.3A CN107743102B (en) | 2017-10-31 | 2017-10-31 | efficient tcp session recombination method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711052422.3A CN107743102B (en) | 2017-10-31 | 2017-10-31 | efficient tcp session recombination method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107743102A CN107743102A (en) | 2018-02-27 |
| CN107743102B true CN107743102B (en) | 2020-01-31 |
Family
ID=61233775
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711052422.3A Active CN107743102B (en) | 2017-10-31 | 2017-10-31 | efficient tcp session recombination method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107743102B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110855584B (en) * | 2019-10-16 | 2022-02-01 | 武汉绿色网络信息服务有限责任公司 | Method and device for TCP out-of-order recombination |
| CN113364862B (en) * | 2021-06-03 | 2022-10-11 | 上海天旦网络科技发展有限公司 | Packet decoding system and method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101795230A (en) * | 2010-02-23 | 2010-08-04 | 西安交通大学 | Network flow recovery method |
| CN105939297A (en) * | 2015-10-26 | 2016-09-14 | 杭州迪普科技有限公司 | TCP message reassembling method and TCP message reassembling device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9336040B2 (en) * | 2014-09-15 | 2016-05-10 | Intel Corporation | Techniques for remapping sessions for a multi-threaded application |
-
2017
- 2017-10-31 CN CN201711052422.3A patent/CN107743102B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101795230A (en) * | 2010-02-23 | 2010-08-04 | 西安交通大学 | Network flow recovery method |
| CN105939297A (en) * | 2015-10-26 | 2016-09-14 | 杭州迪普科技有限公司 | TCP message reassembling method and TCP message reassembling device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107743102A (en) | 2018-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7969976B2 (en) | Gateway apparatus, packet forwarding method, and program | |
| US9906630B2 (en) | Processing data packets in performance enhancing proxy (PEP) environment | |
| US9769701B2 (en) | Header compression for wireless backhaul systems | |
| US7502860B1 (en) | Method and apparatus for client-side flow control in a transport protocol | |
| CN102255972B (en) | HTTP (hyper text transport protocol)-oriented TCP (transmission control protocol) header compression method in 6LoWAPN (IPv-over low-power wireless personal area network) | |
| US7849211B2 (en) | Method and system for reliable multicast datagrams and barriers | |
| CN113055127B (en) | Data message duplicate removal and transmission method, electronic equipment and storage medium | |
| US20070237157A1 (en) | Methods of resolving datagram corruption over an internetworking protocol | |
| CN104025550B (en) | The method and device of information is obtained from data item | |
| US11271856B2 (en) | Concept for segmenting an application buffer into data packets | |
| JP2004530388A (en) | Real-time packetization and retransmission at application delivery | |
| US8140709B2 (en) | Two stage internet protocol header compression | |
| CN102035738B (en) | Method and device for acquiring routing information | |
| CN107743102B (en) | efficient tcp session recombination method | |
| US9525661B2 (en) | Efficient method of NAT without reassemling IPV4 fragments | |
| CN107786472B (en) | Efficient tcp session reorganization device | |
| EP2975819B1 (en) | Reconstructable content objects | |
| CN113726907A (en) | Routing processing method, network element equipment, device and readable storage medium | |
| JP2017538335A (en) | TCP man-in-the-middle processing method in a mode without protocol stack | |
| JP2007129452A (en) | Communication apparatus and communication method | |
| CN103746768B (en) | A kind of recognition methods of packet and equipment | |
| Mosko et al. | Secure fragmentation for content centric networking | |
| US11907773B2 (en) | Method for communicating between software entities via an API | |
| CN115174500B (en) | PISA-based transmitting node and switch for intra-network combined transmission | |
| CN118214616B (en) | WEB request proxy method based on double unidirectional isolation optical gate |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |