CN109787912A - Load-balancing method based on NAT under a kind of DPDK environment - Google Patents
Load-balancing method based on NAT under a kind of DPDK environment Download PDFInfo
- Publication number
- CN109787912A CN109787912A CN201910158676.6A CN201910158676A CN109787912A CN 109787912 A CN109787912 A CN 109787912A CN 201910158676 A CN201910158676 A CN 201910158676A CN 109787912 A CN109787912 A CN 109787912A
- Authority
- CN
- China
- Prior art keywords
- nat
- server
- load
- dpdk
- data packet
- 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
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000004891 communication Methods 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims description 6
- 229910017435 S2 In Inorganic materials 0.000 claims 1
- 230000005540 biological transmission Effects 0.000 abstract 1
- 238000012545 processing Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L61/00—Network arrangements, protocols or services for addressing or naming
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Computer And Data Communications (AREA)
Abstract
The invention discloses the load-balancing methods based on NAT under a kind of DPDK environment, approach application includes the scheduler realization of DPDK module and NAT module, comprising steps of obtaining transmitting-receiving queue port and NAT rule list and state record table to DPDK module and NAT module initialization;Source mesh IP and address using DPDK module record received data packet, calculate source mesh IP and the corresponding cryptographic Hash in address;NAT rule list is searched according to cryptographic Hash, judge whether there is server ip corresponding with obtained cryptographic Hash and port information in NAT rule list, just the source mesh IP and address are replaced if having, realize datagram transmission, otherwise it finds in state record table and loads remaining maximum service device, and the remaining maximum service device of load is sent data packets to, and record the communication information for loading remaining maximum service device and client to NAT rule list and update NAT rule list;The present invention, which can promote scheduler, has good expansion to the applicability of extensive flow, and is conducive to the equilibrium of server load.
Description
Technical field
The invention belongs to network communication fields, and in particular to a kind of DPDK (Data Plane Development Kit, number
According to plane development kit) load of NAT (Network Address Translation, network address translation) is based under environment
Equalization methods.
Background technique
Nowadays with the explosive growth of computer user, network flow also becomes increasing, for server
Load capacity also proposed bigger challenge.It is unbearable performance best single server for now to show such as
The request of modern big flow.So realizing load balancing using the thought of cluster just seems that lattice are important.For traditional
It for load balancing based on NAT, is deployed on traditional ICP/IP protocol stack, system is frequently interrupted, and context is not
Stop scheduling and undoubtedly occupies its a large amount of time.Thus traditional is realized for load balancing using NAT, scheduler itself
Have become the bottleneck of its load balancing.Due to the performance issue of scheduler, under normal circumstances for can be to 10 or so clothes
Business device is scheduled, this in most situations for can be applicable in, but for today that network flow is increasingly exploded, especially
It is the increasingly mature of DDOS technology, and attack traffic GB/S ranks up to a hundred are not problems.Tradition is based in this background
The load balancing that NAT is realized is clearly not to be competent at the work of this part, therefore find the load balancing that can resist more extensive flow
Technology urgently can not be to.
In the case where traditional NAT does non-load balanced case, for external request, scheduler generally uses poll after intercepting
Method, the address for finding out an internal server at random in address pool replaces destination address, and source address remains unchanged.
Internal server is forwarded it to again to go to handle.And the process that be just one for the internal communication with outside opposite, scheduler
The source address that server is transmitted through the data packet come is replaced with into the global address that scheduler possesses, destination address is constant, is sent out
It sees off.
Obviously the advantages of lower NAT of tradition does load balancing is that server can run any operation system for supporting TCP/IP
System, because what a IP only needed to configure on scheduler, oneself privately owned IP address is can be used in server group, has saved public affairs
The IP address of net compares flexible.
But scheduler is due to needing frequently to search data packet, replacement operation, although compared to server pair
Data packet treatment of details, this workload have been lacked very much, but get up having measured once big, and scheduler is not also subjected to such
High-frequency work.For under normal circumstances when server node number is raised to 20, scheduler itself is likely to become system
New bottleneck.
In addition, traditional when doing load balancing using NAT, the general method for considering poll, i.e., each server machine can be equal
Deng scheduled device show appreciation for somebody, using can make each server be equalized to a certain extent uses, but such case is neglected
The difference of server performance itself is omited, A server load capacity is 10 times of B server for exaggeration, for this poll
Method, when B server load reaches the upper limit, A may may just reach the 1/10 of upper loading limit.It is so not for B server
Fair, undoubtedly also reduce the performance of whole system.
Summary of the invention
For it is above-mentioned be unable to satisfy in the prior art more extensive flow and due to scheduling unfairness and cause entire
The problem of system performance reduces, the present invention in proposing the load-balancing method based on NAT under a kind of DPDK environment, in this method by
The processing module of DPDK module adapter tube scheduler, using DPDK to the fast throughput of data packet, in conjunction with the state of NAT module
Record sheet is scheduled forwarding to data packet;The limitation for overcoming scheduler performance itself improves the upper limit of load balancing, real
Show balanced forwarding, specific technical solution is as follows:
Load-balancing method based on NAT under a kind of DPDK environment, it includes DPDK module and NAT module that the method, which is used,
Scheduler realize, specifically include step:
S1, the initialization DPDK module and NAT module, the transmitting-receiving queue of the configuration completion scheduler and corresponding end
Mouthful, and set corresponding NAT rule list and state record table;
S2, source mesh IP and port that the scheduler receives data packet are recorded using DPDK module, be based on the source mesh
IP and port carry out Hash calculation and obtain the cryptographic Hash of the data packet;
S3, the NAT rule list is traversed according to the cryptographic Hash, inquired in the NAT rule list with the presence or absence of corresponding institute
State the IP and port information of the internal server of cryptographic Hash;If it exists, then the source mesh IP is carried out based on the NAT rule list
Replacement, and the data packet is sent to corresponding server, otherwise, execute step S4;
If finding institute there is no the IP and port information of the internal server of the corresponding cryptographic Hash in S4, the NAT
It states and loads remaining maximum service device in state record table, and the data packet is sent to the remaining maximum service device of the load,
And the communication information for loading remaining maximum service device and client is recorded to the NAT rule list and updates the NAT rule
Then table.
Further, in the step S2, further include judge the data packet be sent to from client server or
It is sent to customer side from server, if the data packet is sent to server from client, thens follow the steps S3~S4;Otherwise,
Execute step:
The NAT rule list is traversed according to the cryptographic Hash, inquires the source mesh IP that the NAT rule corresponds to the cryptographic Hash
And port information, and the source mesh IP and port are replaced, the data packet is sent to corresponding client.
Further, in the step S1 further include: to server in the state record table according to remaining load according to
Specified rule carries out grade classification.
Further, the data packet is received using rte_eth_tx_burst () function in the DPDK module.
Compared with prior art, the load-balancing method based on NAT has the beneficial effect that under DPDK environment of the invention
(1) with the adaptability of extensive flow: bypassing legacy protocol stack using DPDK, abandoned some processing streams of legacy protocol stack
Journey accelerates the processing of data packet, breaks through traditional limitation based on NAT load balance scheduler itself, scheduler is adjusted
More servers are spent, the load balancing under extensive flow is realized;(2) there is good harmony, promote server
The overall performance of cluster: the value obtained using the remaining load and performance comparison of state record table can quickly be found out really
The most server of remaining load, compared to the way for using poll various servers under tradition, it is contemplated that the performance of server
Difference realizes load balancing truly;(3) with scalability: based on software to the process performance of scheduler into
Row is promoted, and realizes the load balancing of extensive flow;Compared to hardware system, there are many general api interfaces to open subsequent
Hair has good support;It is also possible to further be developed to software according to actual needs.
Detailed description of the invention
Fig. 1 is the flow diagram signal that a client and server communication is realized using the method for the present invention;
Fig. 2 is the structural schematic diagram of NAT rule list described in the embodiment of the present invention;
Fig. 3 is the model structure signal that the method is completed in the embodiment of the present invention.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described.
In embodiments of the present invention, the load-balancing method based on NAT under a kind of DPDK environment is provided, refering to Fig. 3, this
DPDK module and NAT module are provided in invention in scheduler, method of the invention is completed based on the scheduler, refering to fig. 1,
It is illustrated as realizing that the flow chart of a client and server communication is illustrated using this method method, it can be seen that, side of the present invention
Method specifically includes step:
S1, initialization DPDK module and NAT module, the transmitting-receiving queue of configuration completion scheduler and corresponding port, and set
Corresponding NAT rule list and state record table are set;
In embodiments of the present invention, the transmitting-receiving process that data packet is realized by DPDK module advises NAT by NAT module
Then table and state record table maintenance, wherein NAT rule list is a Hash table, specifically sees Fig. 2, it can be seen that, Mei Gejian
Corresponding two values of value, one when being for receiving, by destination IP, port replacement is sent to internal server, one is for sending out
When sending, source address is replaced with to the IP address of external scheduler, port is then sent to client;NAT rule list is initializing
Shi Jianli, and be updated when scheduler receives data packet, every time by according to the IP of client's section, port information is breathed out
Uncommon, by the IP of internal server, port information is placed on first position, the external IP that scheduler is used, and port information is placed on
Second position;Scheduler first determines that the data packet is that external client receives according to the source port address of data packet in inquiry
Come over what still internal server needed to send, then carries out Hash further according to the IP of client and port information, determine
It is to replace destination IP, port or second port value replacing source IP using first value of key.
And for state record table in the present invention, for recording the concrete condition of the also server with remaining load: this
The internal server that invention sets a NAT module inquires the period, how can specifically be set according to the actual situation, in this regard,
The present invention is simultaneously not limited and fixes;Within an inquiry period, NAT module can periodically internally server be inquired,
The present load of current internal server is inquired, and combines the performance of each internal server, carries out comprehensive consideration;Wherein, comprehensive
Close the program implement considered are as follows: for t ← 0to n;state record[t]←server_load[t]*server_
Kind [t] * r, the performance information of each server is recorded with server_kind=[S1, S2...Sn], the numerical value the big, represents
The performance of server is stronger, and present load server_load and server_kind progress operation is obtained the true of server
Remaining load situation, wherein r be amendment constant, for limiting state_record between 1-100;The division finally obtained
Good grade record is inserted in state record table state_record, and such mode avoids traditional NAT and uses poll
In mode, because ignoring the otherness of server performance, and generation the case where cause the reduction of the overall performance of whole system;It is excellent
Choosing, the loading condition grade classification of internal server is 1-100 totally 100 grades by the present embodiment, and certainly, this is only this hair
Bright preferred embodiment is not limitation and fixation to the present invention program.
S2, source mesh IP and the port that data packet is received using DPDK module recording dispatching device are based on source mesh IP and port
It carries out Hash calculation and obtains the cryptographic Hash of data packet;
In a specific embodiment of the present invention, DPDK module uses rte_eth_tx_burst () function received data packet,
The program implement of rte_eth_tx_burst () function are as follows:
Be initially configured from rte_eth_dev_configure () and receive and dispatch queue, then by memcpy (&dev- >
Data- > dev_conf, sizeof (dev- > data- > dev_conf)) function copies configuration parameter the data field of equipment to simultaneously
Facility information is obtained with (* dev- > dev_ops- > dev_infos_get) (dev, &dev_info);Meanwhile the present invention passes through
Rte_eth_rx_queue_setup () function is to receiving queue initiating port and configures 8 queues, followed by rxq=
rte_zmalloc_socket("ethdev RX queue",sizeof(structixgbe_rx_queue),RTE_CACHE_
LINE_SIZE, socket_id) function distribution queue structure simultaneously complete to utilize rte_eth_tx_burst by interstitital texture body, configuration
() function received data packet, and record the source mesh IP and port information of corresponding data packet;Meanwhile scheduler is to the number received
Judge that data packet is to be sent to server from client or be sent to customer side from server according to packet, if data packet is from clothes
Business device is sent to client;Hash calculation is carried out according to IP and port and obtains corresponding cryptographic Hash, and NAT is traversed according to cryptographic Hash
Rule list, inquiry NAT rule corresponds to the source mesh IP and port information of the cryptographic Hash, and source mesh IP and port are replaced, and passes through
Data packet is sent to corresponding client by DPDK module by rte_eth_tx_burst () function;Otherwise, if data packet is from visitor
Family end is sent to server, equally carries out Hash calculation based on source mesh IP and port and obtains the cryptographic Hash of data packet, according to obtaining
Cryptographic Hash execute following step S3~S4.
S3, NAT rule list is searched according to cryptographic Hash, inquires the inside clothes in NAT rule list with the presence or absence of corresponding cryptographic Hash
The IP and port information of business device;If it exists, then source mesh IP is replaced based on NAT rule list, and delivered a packet to pair
Server is answered, otherwise, executes step S4;
Specifically, recording source data packet mesh IP address port after rte_eth_tx_burst () function receives data packet
Information, scheduler identifies that this is received data packet, utilizes source port, IP information carries out Hash calculation, according to cryptographic Hash
Search the IP whether NAT rule list has corresponding internal server, port information;If it exists, then based on NAT rule list to source mesh
IP is replaced, and delivers a packet to corresponding server, otherwise, executes step S4.
If there is no the IP and port information of the internal server of the corresponding cryptographic Hash in S4, NAT, state note is found
It records and loads remaining maximum service device in table, and send data packets to the remaining maximum service device of load, and it is remaining most to record load
The communication information of big server and client to NAT rule list and updates NAT rule list.
Specifically, representing this is that client is asked for the first time if not finding corresponding information in NAT rule list
Ask, then scheduler looks up state record table, find out recorded in state record table number it is maximum that, its is corresponding
The IP of internal server takes out, and then scheduler replaces the packet rs destination address, in the rte_eth_tx_ by DPDK module
Burst () function is sent to internal server, and record loads the communication information of remaining maximum service device and client to NAT
Rule list simultaneously updates NAT rule list, so that next time uses.
The above-mentioned complete process that server and client side communication is carried out using the method for the present invention to be carried out based on Fig. 1
Description, the principle and mechanism repeatedly communicated are no longer repeated herein as once communicated.
In summary, compared with prior art, the method for the present invention bypasses legacy protocol stack using DPDK, has abandoned traditional association
The some process flows for discussing stack accelerate the processing of data packet, break through traditional limitation based on NAT load balance scheduler itself,
Allow scheduler to dispatch more servers, realize the load balancing under extensive flow, so that scheduler can fit
More massive flow is answered, application field is wider;The present invention is obtained using the remaining load and performance comparison of state record table
Value can quickly find out the most server of real remaining load, compared to doing using poll various servers under tradition
Method, it is contemplated that the performance difference of server realizes load balancing truly, can be well between equalization server
Load, promote the overall performance of server cluster;The present invention can be promoted based on process performance of the software to scheduler, real
The now load balancing of extensive flow;Compared to hardware system, there are many general api interfaces that there is subsequent exploitation good
Support;It is also possible to further be developed to software according to actual needs, so that scheduler has good expand
Malleability.
The foregoing is merely a prefered embodiment of the invention, is not intended to limit the scope of the patents of the invention, although referring to aforementioned reality
Applying example, invention is explained in detail, for a person skilled in the art, still can be to aforementioned each specific
Technical solution documented by embodiment is modified, or carries out equivalence replacement to part of technical characteristic.All utilizations
The equivalent structure that description of the invention and accompanying drawing content are done directly or indirectly is used in other related technical areas, together
Reason is within the invention patent protection scope.
Claims (4)
1. the load-balancing method based on NAT under a kind of DPDK environment, which is characterized in that it includes DPDK module that the method, which is used,
It is realized with the scheduler of NAT module, specifically includes step:
S1, the initialization DPDK module and NAT module, the transmitting-receiving queue of the configuration completion scheduler and corresponding port,
And set corresponding NAT rule list and state record table;
S2, source mesh IP and port that the scheduler receives data packet are recorded using DPDK module, based on the source mesh IP and
Port carries out Hash calculation and obtains the cryptographic Hash of the data packet;
S3, the NAT rule list is traversed according to the cryptographic Hash, inquired in the NAT rule list with the presence or absence of the corresponding Kazakhstan
The IP and port information of the internal server of uncommon value;If it exists, then the source mesh IP is replaced based on the NAT rule list
It changes, and the data packet is sent to corresponding server, otherwise, execute step S4;
If finding the shape there is no the IP and port information of the internal server of the corresponding cryptographic Hash in S4, the NAT
Remaining maximum service device is loaded in state record sheet, and the data packet is sent to the remaining maximum service device of the load, and remember
The communication information for loading remaining maximum service device and client is recorded to the NAT rule list and updates the NAT rule list.
2. the load-balancing method based on NAT under DPDK environment as described in claim 1, which is characterized in that the step S2
In, it further include judging that the data packet is to be sent to server from client or be sent to customer side from server, if described
Data packet is sent to server from client, thens follow the steps S3~S4;Otherwise, step is executed:
The NAT rule list is traversed according to the cryptographic Hash, inquires source mesh IP and end that the NAT rule corresponds to the cryptographic Hash
Message breath, and the source mesh IP and port are replaced, the data packet is sent to corresponding client.
3. the load-balancing method based on NAT under DPDK environment as described in claim 1, which is characterized in that the step S1
In further include: grade classification is carried out according to specified rule according to remaining load to server in the state record table.
4. the load-balancing method based on NAT under DPDK environment as claimed in any one of claims 1 to 3, which is characterized in that institute
It states in DPDK module and the data packet is received using rte_eth_tx_burst () function.
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910158676.6A CN109787912A (en) | 2019-03-04 | 2019-03-04 | Load-balancing method based on NAT under a kind of DPDK environment |
PCT/CN2019/092690 WO2020177246A1 (en) | 2019-03-04 | 2019-06-25 | Nat-based method for load balancing in dpdk environment |
JP2021521491A JP7177548B2 (en) | 2019-03-04 | 2019-06-25 | Load Balancing Method Based on NAT in DPDK Environment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910158676.6A CN109787912A (en) | 2019-03-04 | 2019-03-04 | Load-balancing method based on NAT under a kind of DPDK environment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109787912A true CN109787912A (en) | 2019-05-21 |
Family
ID=66487429
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910158676.6A Pending CN109787912A (en) | 2019-03-04 | 2019-03-04 | Load-balancing method based on NAT under a kind of DPDK environment |
Country Status (3)
Country | Link |
---|---|
JP (1) | JP7177548B2 (en) |
CN (1) | CN109787912A (en) |
WO (1) | WO2020177246A1 (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110661904A (en) * | 2019-10-25 | 2020-01-07 | 浪潮云信息技术有限公司 | Method for realizing horizontal extension of source network address translation gateway |
CN111010342A (en) * | 2019-11-21 | 2020-04-14 | 天津卓朗科技发展有限公司 | Distributed load balancing implementation method and device |
CN111338793A (en) * | 2020-02-14 | 2020-06-26 | 拉扎斯网络科技(上海)有限公司 | Cluster task distribution method, device, equipment and storage medium |
WO2020177246A1 (en) * | 2019-03-04 | 2020-09-10 | 南京邮电大学 | Nat-based method for load balancing in dpdk environment |
CN112217919A (en) * | 2020-12-11 | 2021-01-12 | 广东省新一代通信与网络创新研究院 | Method and system for realizing network address conversion |
CN114598746A (en) * | 2022-03-07 | 2022-06-07 | 中南大学 | Method for optimizing load balancing performance between servers based on intelligent network card |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113518270A (en) * | 2021-05-24 | 2021-10-19 | 北京邮电大学 | Heterogeneous data acquisition device and method for mobile network |
CN115720222B (en) * | 2022-12-19 | 2023-06-02 | 广西大学 | Method for realizing HTTP forwarding on ARM multi-core architecture and storage medium |
CN116527410B (en) * | 2023-07-05 | 2023-09-26 | 北京亿赛通科技发展有限责任公司 | Firewall multiprocess processing method and device in reverse proxy mode |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102970242A (en) * | 2012-11-09 | 2013-03-13 | 深圳市共进电子股份有限公司 | Method for achieving load balancing |
CN103152444A (en) * | 2013-03-25 | 2013-06-12 | 华为技术有限公司 | Network address translation and message transmission method, device and system of trunking scheme |
CN107438012A (en) * | 2016-05-27 | 2017-12-05 | 北京金山云网络技术有限公司 | A kind of load balancing service retransmission method, system, balancer and host |
US20180213440A1 (en) * | 2017-01-25 | 2018-07-26 | Affirmed Networks, Inc. | Load balancing of wireless subscriber packet processing over multiple packet processing cores on a virtual machine platform |
CN108600120A (en) * | 2018-05-04 | 2018-09-28 | 广东唯网络科技有限公司 | A method of based on realization NAT under DPDK |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10757024B2 (en) | 2014-12-24 | 2020-08-25 | Ntt Communications Corporation | Load distribution apparatus, load distribution method and program |
DE112017003494T5 (en) * | 2016-07-11 | 2019-05-02 | Harmonic Inc. | MULTI-CORE SOFTWARE FORWARDING |
CN107196870B (en) * | 2017-07-20 | 2021-07-20 | 哈尔滨工业大学 | DPDK-based traffic dynamic load balancing method |
CN107786458B (en) * | 2017-11-02 | 2021-06-25 | 下一代互联网重大应用技术(北京)工程研究中心有限公司 | DPDK-based multi-port access and egress method |
CN109787912A (en) * | 2019-03-04 | 2019-05-21 | 南京邮电大学 | Load-balancing method based on NAT under a kind of DPDK environment |
-
2019
- 2019-03-04 CN CN201910158676.6A patent/CN109787912A/en active Pending
- 2019-06-25 WO PCT/CN2019/092690 patent/WO2020177246A1/en active Application Filing
- 2019-06-25 JP JP2021521491A patent/JP7177548B2/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102970242A (en) * | 2012-11-09 | 2013-03-13 | 深圳市共进电子股份有限公司 | Method for achieving load balancing |
CN103152444A (en) * | 2013-03-25 | 2013-06-12 | 华为技术有限公司 | Network address translation and message transmission method, device and system of trunking scheme |
CN107438012A (en) * | 2016-05-27 | 2017-12-05 | 北京金山云网络技术有限公司 | A kind of load balancing service retransmission method, system, balancer and host |
US20180213440A1 (en) * | 2017-01-25 | 2018-07-26 | Affirmed Networks, Inc. | Load balancing of wireless subscriber packet processing over multiple packet processing cores on a virtual machine platform |
CN108600120A (en) * | 2018-05-04 | 2018-09-28 | 广东唯网络科技有限公司 | A method of based on realization NAT under DPDK |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020177246A1 (en) * | 2019-03-04 | 2020-09-10 | 南京邮电大学 | Nat-based method for load balancing in dpdk environment |
JP2022505453A (en) * | 2019-03-04 | 2022-01-14 | 南京郵電大学 | Load balancing method based on NAT in DPDK environment |
JP7177548B2 (en) | 2019-03-04 | 2022-11-24 | 南京郵電大学 | Load Balancing Method Based on NAT in DPDK Environment |
CN110661904A (en) * | 2019-10-25 | 2020-01-07 | 浪潮云信息技术有限公司 | Method for realizing horizontal extension of source network address translation gateway |
CN111010342A (en) * | 2019-11-21 | 2020-04-14 | 天津卓朗科技发展有限公司 | Distributed load balancing implementation method and device |
CN111010342B (en) * | 2019-11-21 | 2023-04-07 | 天津卓朗科技发展有限公司 | Distributed load balancing implementation method and device |
CN111338793A (en) * | 2020-02-14 | 2020-06-26 | 拉扎斯网络科技(上海)有限公司 | Cluster task distribution method, device, equipment and storage medium |
CN112217919A (en) * | 2020-12-11 | 2021-01-12 | 广东省新一代通信与网络创新研究院 | Method and system for realizing network address conversion |
CN112217919B (en) * | 2020-12-11 | 2021-03-23 | 广东省新一代通信与网络创新研究院 | Method and system for realizing network address conversion |
CN114598746A (en) * | 2022-03-07 | 2022-06-07 | 中南大学 | Method for optimizing load balancing performance between servers based on intelligent network card |
Also Published As
Publication number | Publication date |
---|---|
JP2022505453A (en) | 2022-01-14 |
WO2020177246A1 (en) | 2020-09-10 |
JP7177548B2 (en) | 2022-11-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109787912A (en) | Load-balancing method based on NAT under a kind of DPDK environment | |
Luo et al. | Towards practical and near-optimal coflow scheduling for data center networks | |
US8977753B1 (en) | System and method for router keep-alive control | |
CN109889551A (en) | A kind of method of the Internet of Things cloud platform of Intelligent hardware access | |
US8447881B2 (en) | Load balancing for services | |
CN108989372A (en) | Method, registration center and the equipment of service discovery | |
EP2412124B1 (en) | Network topology | |
CN109547580A (en) | A kind of method and apparatus handling data message | |
CN108833462A (en) | A kind of system and method found from registration service towards micro services | |
CN108702316A (en) | A kind of resource allocation methods and device of VNF | |
CN107493191A (en) | A kind of clustered node and self scheduling container group system | |
CN103873523A (en) | Client cluster access method and device | |
CN103139157A (en) | Network communication method based on socket, device and system | |
CN112637332B (en) | Service registration discovery method and system | |
CN109510878A (en) | A kind of long connection session keeping method and device | |
US20040068729A1 (en) | Non-hierarchical collaborative computing platform | |
WO2018226146A1 (en) | Method and node for distributed network performance monitoring | |
CN102158406B (en) | Intelligent routing method for computer network links | |
CN103617021B (en) | General processor hardware abstraction layer communication system and method in software communication architectural framework | |
CN109960586A (en) | A kind of appreciable four-layer load-equalizing device of server state and equalization methods | |
CN108055296B (en) | Transaction processing method and device based on micro-service architecture | |
CN112583895A (en) | TCP communication method, system and device | |
CN106131162B (en) | A method of network service agent is realized based on IOCP mechanism | |
US8103759B2 (en) | Message redirection within a messaging infrastructure | |
CN106227545B (en) | The design method and device of development platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190521 |
|
RJ01 | Rejection of invention patent application after publication |