CN108156034B - Message forwarding method and message forwarding system based on deep neural network assistance - Google Patents
Message forwarding method and message forwarding system based on deep neural network assistance Download PDFInfo
- Publication number
- CN108156034B CN108156034B CN201711409344.8A CN201711409344A CN108156034B CN 108156034 B CN108156034 B CN 108156034B CN 201711409344 A CN201711409344 A CN 201711409344A CN 108156034 B CN108156034 B CN 108156034B
- Authority
- CN
- China
- Prior art keywords
- message
- forwarding
- cloud host
- neural network
- forwarding node
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 72
- 238000012544 monitoring process Methods 0.000 claims description 4
- 238000004891 communication Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/74—Address processing for routing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention is suitable for the communication field, and provides a message forwarding method and a message forwarding system based on deep neural network assistance, wherein the message forwarding method comprises the following steps: the cloud host trains through a deep neural network and sends a training result to a forwarding node, wherein the training result is used for recording a receiving place of a message sent by the cloud host; the forwarding node loads the training result into a forwarding service; and the forwarding node receives a message sent by the cloud host, and sends the message to a corresponding service address from the same port according to the training result. By implementing the embodiment of the invention, messages sent by a plurality of cloud hosts can share one port to send the messages, so that the floating IP in the cloud computing environment is saved.
Description
Technical Field
The invention belongs to the field of communication, and particularly relates to a message forwarding method and a message forwarding system based on deep neural network assistance.
Background
In the current cloud computing environment, there is a common deployment scheme that physical servers providing cloud computing services are connected through a network device (e.g., a switch, a router) that exists in reality; the cloud host used as the cloud computing resource for the user is connected through the virtual network equipment; when the physical machine and the cloud host need to communicate, the network bridge equipment forwards the message. If a user wishes to create multiple cloud hosts and set up multiple network service pairings for external use in a cloud computing environment, the traditional solution is to assign a floating IP for each cloud host. But floating IP in a cloud computing environment is similar to public network IP in a real IPv4 network, and the number thereof is very limited.
Disclosure of Invention
The embodiment of the invention aims to provide a message forwarding method and a message forwarding system based on deep neural network assistance, so as to solve the problem that a plurality of cloud hosts need to use a plurality of floating IP in the prior art.
The embodiment of the invention is realized in such a way that a message forwarding method based on deep neural network assistance comprises the following steps:
the cloud host trains through a deep neural network and sends a training result to a forwarding node, wherein the training result is used for recording a receiving place of a message sent by the cloud host;
the forwarding node loads the training result into a forwarding service;
and the forwarding node receives a message sent by the cloud host, and sends the message to a corresponding service address from the same port according to the training result.
Another objective of an embodiment of the present invention is to provide a deep neural network assisted packet forwarding system, where the packet forwarding system includes:
the cloud host is used for training through a deep neural network and sending a training result to the forwarding node, and the training result is used for recording a receiving place of a message sent by the cloud host.
And the forwarding node is used for loading the training result into a forwarding service, receiving a message sent by the cloud host, and sending the message to a corresponding service address from the same port according to the training result.
According to the embodiment of the invention, the cloud host trains through the deep neural network, the training result is sent to the forwarding node, the forwarding node loads the training result into the forwarding service, the forwarding node receives the message sent by the cloud host, and the message is sent to the corresponding service address from the same port according to the training result, so that the messages sent by a plurality of cloud hosts can share one port to send the message, and the floating IP in the cloud computing environment is saved.
Drawings
Fig. 1 is a flowchart illustrating a message forwarding method based on deep neural network assistance according to an exemplary embodiment of the present invention;
fig. 2 is a structural diagram of a message forwarding system based on deep neural network assistance according to an exemplary embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In order to explain the technical means of the present invention, the following description will be given by way of specific examples.
Fig. 1 is a flowchart of a packet forwarding method based on deep neural network assistance according to an exemplary embodiment of the present invention, where the packet forwarding method includes the following steps:
step S101, a cloud host trains through a deep neural network, and sends a training result to a forwarding node, wherein the training result is used for recording a receiving place of a message sent by the cloud host.
In the embodiment of the present invention, the forwarding node is selected from the trained cloud hosts, and the cloud host serving as the forwarding node must satisfy the following conditions: all services required by the cloud host can be accessed, the hardware of the terminal is enough for forwarding services, and a specific port can be provided for the external device.
The cloud host trains through the deep neural network and sends a training result to the forwarding node, and the method comprises the following steps:
1. the cloud host forwards the message in an experimental cloud computing environment according to the configured service and records the virtual IP address of the service.
2. And monitoring the message, and removing the message which does not contain data from the message.
In the embodiment of the present invention, the inclusion of the non-data is generally: SVN message, ACK message, FIN message and RST message, therefore only need to reject: SVN message, ACK message, FIN message and RST message.
3. And classifying the eliminated messages, and training the classified messages based on the deep neural network.
In the embodiment of the invention, the classified messages can be trained based on the deep neural network, so as to obtain the training result, and the training result is used in the subsequent steps. The specific process comprises the following steps:
(1) extracting a destination port field of the TCP header from the eliminated message as a classification result;
(2) extracting the following fields from the eliminated messages as classification features: the first 20 bytes of the TCP data part, the times of occurrence of the \ r \ n character string of the TCP data part, the percentage of the 0 content bytes of the TCP data part in the total bytes of the data part and the complete content of the TCP data part;
(3) and sequentially giving high-to-low weights to the classification features, and establishing a deep neural network for training.
4. And sending the training result to a forwarding node.
Step S102, the forwarding node loads the training result into the forwarding service.
In the embodiment of the invention, the training result is a basis for the forwarding node to forward the received message, so the training result of the cloud host needs to be loaded into the forwarding node.
Step S103, the forwarding node receives a message sent by the cloud host, and sends the message to a corresponding service address from the same port according to the training result.
In the embodiment of the invention, the training result records the service addresses to which different cloud hosts need to be sent, so that after receiving the message sent by the cloud host, the forwarding node compares the cloud host with the locally loaded training result to obtain the service address to which the cloud host needs to be sent, and forwards the message to the corresponding service address.
According to the embodiment of the invention, the cloud host trains through the deep neural network, the training result is sent to the forwarding node, the forwarding node loads the training result into the forwarding service, the forwarding node receives the message sent by the cloud host, and the message is sent to the corresponding service address from the same port according to the training result, so that the messages sent by a plurality of cloud hosts can share one port to send the message, and the floating IP in the cloud computing environment is saved.
As an optional embodiment of the present invention, after the step of receiving, by the forwarding node, a packet sent by a cloud host and sending the packet from the same port to a corresponding service address according to the training result, the packet forwarding method further includes:
and the forwarding node establishes association between the cloud host and the service address and forwards a message sent by the service address to the corresponding cloud host according to the association relation.
In the embodiment of the invention, after the message sent by the cloud host is forwarded to the service address, the forwarding node establishes an association relationship between the cloud host and the service address, and in the subsequent message sending, the cloud host and the service address can communicate through the association relationship.
Fig. 2 is a structural diagram of a deep neural network-assisted packet forwarding system according to an exemplary embodiment of the present invention, where the packet forwarding system includes:
the cloud host 201 is configured to train through a deep neural network, and send a training result to a forwarding node, where the training result is used to record a receiving location of a packet sent by the cloud host.
And the forwarding node 202 is configured to load the training result into a forwarding service, receive a packet sent by the cloud host, and send the packet to a corresponding service address from the same port according to the training result.
In the embodiment of the present invention, the forwarding node is selected from the trained cloud hosts, and the cloud host serving as the forwarding node must satisfy the following conditions: all services required by the cloud host can be accessed, the hardware of the terminal is enough for forwarding services, and a specific port can be provided for the external device.
The cloud host trains through the deep neural network and sends a training result to the forwarding node, and the method comprises the following steps:
1. the cloud host forwards the message in an experimental cloud computing environment according to the configured service and records the virtual IP address of the service.
2. And monitoring the message, and removing the message which does not contain data from the message.
In the embodiment of the present invention, the inclusion of the non-data is generally: SVN message, ACK message, FIN message and RST message, therefore only need to reject: SVN message, ACK message, FIN message and RST message.
3. And classifying the eliminated messages, and training the classified messages based on the deep neural network.
In the embodiment of the invention, the classified messages can be trained based on the deep neural network, so as to obtain the training result, and the training result is used in the subsequent steps. The specific process comprises the following steps:
(1) extracting a destination port field of the TCP header from the eliminated message as a classification result;
(2) extracting the following fields from the eliminated messages as classification features: the first 20 bytes of the TCP data part, the times of occurrence of the \ r \ n character string of the TCP data part, the percentage of the 0 content bytes of the TCP data part in the total bytes of the data part and the complete content of the TCP data part;
(3) and sequentially giving high-to-low weights to the classification features, and establishing a deep neural network for training.
4. And sending the training result to a forwarding node.
The training result is a basis for the forwarding node to forward the received message, and therefore the training result of the cloud host needs to be loaded into the forwarding node.
The training result records service addresses to which different cloud hosts need to be sent, so that after receiving a message sent by the cloud host, the forwarding node compares the cloud host with the locally loaded training result to obtain the service address to which the cloud host needs to be sent, and forwards the message to the corresponding service address.
According to the embodiment of the invention, the cloud host trains through the deep neural network, the training result is sent to the forwarding node, the forwarding node loads the training result into the forwarding service, the forwarding node receives the message sent by the cloud host, and the message is sent to the corresponding service address from the same port according to the training result, so that the messages sent by a plurality of cloud hosts can share one port to send the message, and the floating IP in the cloud computing environment is saved.
As an optional embodiment of the present invention, the forwarding node is further configured to establish an association between the cloud host and the service address, and forward a packet sent by the service address to the corresponding cloud host according to the association relationship.
In the embodiment of the invention, after the message sent by the cloud host is forwarded to the service address, the forwarding node establishes an association relationship between the cloud host and the service address, and in the subsequent message sending, the cloud host and the service address can communicate through the association relationship.
Those skilled in the art can understand that each unit included in the above embodiments is only divided according to functional logic, but is not limited to the above division as long as the corresponding function can be realized; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It will be further understood by those skilled in the art that all or part of the steps in the method for implementing the above embodiments may be implemented by relevant hardware instructed by a program stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (8)
1. A message forwarding method based on deep neural network assistance is characterized by comprising the following steps:
the cloud host trains through a deep neural network, and sends a training result to a forwarding node, wherein the training result is used for recording a receiving place of a message sent by the cloud host, the forwarding node is selected from the cloud host, and the cloud host serving as the forwarding node provides a specific port to the outside;
the forwarding node loads the training result into a forwarding service;
the forwarding node receives a message sent by the cloud host, and sends the message to a corresponding service address from the same port according to the training result;
after the step of receiving, by the forwarding node, a packet sent by the cloud host and sending the packet from the same port to the corresponding service address according to the training result, the packet forwarding method further includes:
and the forwarding node establishes association between the cloud host and the service address and forwards a message sent by the service address to the corresponding cloud host according to the association relation.
2. The message forwarding method according to claim 1, wherein the training of the cloud host by the deep neural network and the sending of the training result to the forwarding node comprise:
the method comprises the following steps that a cloud host forwards a message in an experimental cloud computing environment according to configured service, and records a virtual IP address of the service;
monitoring the message, and removing the message which does not contain data from the message;
classifying the eliminated messages, and training the classified messages based on a deep neural network;
and sending the training result to a forwarding node.
3. The message forwarding method according to claim 2, wherein the classifying the rejected messages and performing deep neural network-based training on the classified messages comprises:
extracting a destination port field of the TCP header from the eliminated message as a classification result;
extracting the following fields from the eliminated messages as classification features: the first 20 bytes of the TCP data part, the times of occurrence of the \ r \ n character string of the TCP data part, the percentage of the 0 content bytes of the TCP data part in the total bytes of the data part and the complete content of the TCP data part;
and sequentially giving high-to-low weights to the classification features, and establishing a deep neural network for training.
4. The message forwarding method according to any one of claims 1 to 3, wherein the condition that the forwarding node satisfies includes: the service required by all the cloud hosts can be accessed, the hardware of the terminal is enough to carry out forwarding service, and a specific port can be provided for the external equipment.
5. A message forwarding system based on deep neural network assistance, the message forwarding system comprising:
the cloud host is used for training through a deep neural network and sending a training result to the forwarding node, the training result is used for recording a receiving place of a message sent by the cloud host, the forwarding node is selected from the cloud host, and the cloud host serving as the forwarding node provides a specific port to the outside;
the forwarding node is used for loading the training result into a forwarding service, receiving a message sent by the cloud host, and sending the message to a corresponding service address from the same port according to the training result;
and the forwarding node is also used for establishing association between the cloud host and the service address and forwarding a message sent by the service address to the corresponding cloud host according to the association relation.
6. The message forwarding system of claim 5, wherein the cloud host is trained through a deep neural network and sends the training results to the forwarding node, comprising:
the method comprises the following steps that a cloud host forwards a message in an experimental cloud computing environment according to configured service, and records a virtual IP address of the service;
monitoring the message, and removing the message which does not contain data from the message;
classifying the eliminated messages, and training the classified messages based on a deep neural network;
and sending the training result to a forwarding node.
7. The message forwarding system of claim 6 wherein the classifying the rejected messages and the deep neural network based training of the classified messages comprises:
extracting a destination port field of the TCP header from the eliminated message as a classification result;
extracting the following fields from the eliminated messages as classification features: the first 20 bytes of the TCP data part, the times of occurrence of the \ r \ n character string of the TCP data part, the percentage of the 0 content bytes of the TCP data part in the total bytes of the data part and the complete content of the TCP data part;
and sequentially giving high-to-low weights to the classification features, and establishing a deep neural network for training.
8. The message forwarding system according to any of claims 5 to 7, wherein the condition that the forwarding node satisfies includes: the service required by all the cloud hosts can be accessed, the hardware of the terminal is enough to carry out forwarding service, and a specific port can be provided for the external equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711409344.8A CN108156034B (en) | 2017-12-22 | 2017-12-22 | Message forwarding method and message forwarding system based on deep neural network assistance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711409344.8A CN108156034B (en) | 2017-12-22 | 2017-12-22 | Message forwarding method and message forwarding system based on deep neural network assistance |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108156034A CN108156034A (en) | 2018-06-12 |
| CN108156034B true CN108156034B (en) | 2021-10-01 |
Family
ID=62464331
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711409344.8A Active CN108156034B (en) | 2017-12-22 | 2017-12-22 | Message forwarding method and message forwarding system based on deep neural network assistance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108156034B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112688917A (en) * | 2020-12-10 | 2021-04-20 | 龙芯中科技术股份有限公司 | Network access method, device, electronic equipment and storage medium |
| CN112929977B (en) * | 2021-02-10 | 2022-05-31 | 山西大学 | Deep learning amplification forwarding cooperative network energy efficiency resource allocation method |
| CN115134298A (en) * | 2021-03-23 | 2022-09-30 | 华为技术有限公司 | Message processing method and network device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101877728A (en) * | 2010-06-25 | 2010-11-03 | 中兴通讯股份有限公司 | Network address translation forwarding method and device |
| CN106878482A (en) * | 2017-01-03 | 2017-06-20 | 新华三技术有限公司 | Method for network address translation and device |
| CN107241237A (en) * | 2017-05-22 | 2017-10-10 | 北京知道创宇信息技术有限公司 | A kind of method and computing device for recognizing the affiliated component of message |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101257417A (en) * | 2008-03-25 | 2008-09-03 | 浙江大学 | Detection method of TCP/IP protocol covert channel based on fuzzy neural network |
| US8902896B2 (en) * | 2012-04-16 | 2014-12-02 | International Business Machines Corporation | Packet switching without look-up table for ethernet switches |
| US9563854B2 (en) * | 2014-01-06 | 2017-02-07 | Cisco Technology, Inc. | Distributed model training |
| US10129180B2 (en) * | 2015-01-30 | 2018-11-13 | Nicira, Inc. | Transit logical switch within logical router |
-
2017
- 2017-12-22 CN CN201711409344.8A patent/CN108156034B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101877728A (en) * | 2010-06-25 | 2010-11-03 | 中兴通讯股份有限公司 | Network address translation forwarding method and device |
| CN106878482A (en) * | 2017-01-03 | 2017-06-20 | 新华三技术有限公司 | Method for network address translation and device |
| CN107241237A (en) * | 2017-05-22 | 2017-10-10 | 北京知道创宇信息技术有限公司 | A kind of method and computing device for recognizing the affiliated component of message |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108156034A (en) | 2018-06-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9871781B2 (en) | Systems and methods for path maximum transmission unit discovery | |
| US10057164B2 (en) | Apparatus and methods to aggregate FCoE (fibre channel over ethernet) filter rules of a single interface in a single or few rules on a first-hop FCoE networking element | |
| US10375193B2 (en) | Source IP address transparency systems and methods | |
| CN107231307B (en) | Method, system, and apparatus for preventing traffic switching between subnets within a data center architecture | |
| RU2464722C2 (en) | Method, device and system for distribution of messages | |
| US9912637B2 (en) | Use of packet header extension for layer-3 direct server return | |
| CN106412142A (en) | Resource device address obtaining method and device | |
| CN113326228A (en) | Message forwarding method, device and equipment based on remote direct data storage | |
| CN108156034B (en) | Message forwarding method and message forwarding system based on deep neural network assistance | |
| US8472420B2 (en) | Gateway device | |
| CN101640823B (en) | Method and equipment for shunting multi-analysis system | |
| CN107580079A (en) | A kind of message transmitting method and device | |
| US8381301B1 (en) | Split-flow attack detection | |
| CN103685032A (en) | Message forwarding method and network address conversion server | |
| CN104702709A (en) | Method and device for marking terminal type, and method and device for identifying terminal type | |
| CN104065688A (en) | Method for calling underlayer services and device thereof | |
| CN101741691A (en) | Processing method for network flow load balance | |
| CN105100300B (en) | Method and device for converting network address into NAT | |
| CN110166375A (en) | A kind of message forwarding method and device | |
| CN109660459B (en) | Physical gateway and method for multiplexing IP address | |
| CN111327543A (en) | Message forwarding method and device, storage medium, and electronic device | |
| CN108809795A (en) | Transparent shunt method and device in a kind of LAN environment | |
| CN101908982A (en) | Method, device and system for virtual local area network detection | |
| CN105634789A (en) | Method of collector for associating device and log collection system | |
| KR101794200B1 (en) | Method for transmitting and receiving protocol packet using packet processor based LINUX in distributed packet transport network |
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 |