CN110392108A - A kind of public cloud Network Load Balance system architecture and implementation method - Google Patents
A kind of public cloud Network Load Balance system architecture and implementation method Download PDFInfo
- Publication number
- CN110392108A CN110392108A CN201910666177.8A CN201910666177A CN110392108A CN 110392108 A CN110392108 A CN 110392108A CN 201910666177 A CN201910666177 A CN 201910666177A CN 110392108 A CN110392108 A CN 110392108A
- Authority
- CN
- China
- Prior art keywords
- scheduler
- module
- manager module
- message
- configuration
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
- H04L12/46—Interconnection of networks
- H04L12/4641—Virtual LANs, VLANs, e.g. virtual private networks [VPN]
-
- 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
- H04L45/58—Association of routers
- H04L45/586—Association of routers of virtual routers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1034—Reaction to server failures by a load balancer
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/56—Provisioning of proxy services
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention is more particularly directed to a kind of public cloud Network Load Balance system architecture and implementation methods.The public cloud Network Load Balance system architecture, including northbound interface module, LB manager module, policybased routing manager module, LB configuration database, LB proxy module and LVS module；Wherein, the northbound interface module, LB manager module, policybased routing manager module and LB configuration database are located at control plane, and the LB proxy module and LVS module are located at data plane.The public cloud Network Load Balance system architecture and implementation method, the session transparency is not destroyed, client ip address is visible to rear end RS server, the configuration and rear end RS server zero configuration of modification rear end RS server are not needed simultaneously, additional agreement support and additional protocol encapsulation expense are not needed, publicly-owned cloud network four-layer load-equalizing is realized yet.
Description
Technical field
The present invention relates to cloud computings and technical field of the computer network, in particular to a kind of public cloud Network Load Balance system
System framework and implementation method.
Background technique
Publicly-owned cloud service provider one all can to its tenant provide VPC (Virtual Private Cloud), EIP
The basic networks services such as (Elastic IP), SLB (Service Load Balancer).Load balancing (i.e. SLB) is as publicly-owned
One important foundation service of cloud, plays considerable role in publicly-owned cloud network.The load balancing service of public cloud
Compared with traditional software load equilibrium and hardware load equilibrium, there is certain similitude, but also have its exclusive
Some characteristics, such as the load balancing example of public cloud are usually located inside tenant VPC, public cloud load balancing example generally with
There is (minority, which pursues high performance public cloud load balancing, may directly adopt hardware mode), public cloud in the mode of virtualization
Four-layer load-equalizing service cannot generally configure tenant back-end server (need in a transparent manner with rear end real
Server is used cooperatively).
The most typical implementation of four-layer load-equalizing is LVS (Linux Virtual Server), at present the four of industry
Layer load balancing scheme is all based on LVS or its mutation to realize.LVS there are four types of operating mode, NAT mode, TUN mode,
DR mode, FULLNAT mode.NAT mode is that scheduling is realized by the method for network address translation, first scheduler (LB)
When receiving the request data package of client (destination IP of request is VIP), which is sent the requests to according to dispatching algorithm decision
Then the real server (RS) of rear end is dispatched and just the target ip address for the request data package that client is sent and port is changed to
The IP address (RIP, Real IP) of rear end real server, such real server (RS) can receive the request of client
Data packet after real server has responded request, checks that (we need the default route RS to default route under NAT mode
It is set as LB server) data packet after response is sent to LB, after LB receives response bag again, the source address of packet is changed to void
Quasi- address (VIP) is then sent back to client.TUN mode is that transmission between LB and RS does not have to rewrite IP address, but
Client's request packet is encapsulated in inside an IP tunnel, is then sent to RS node server, node server receives it
After unlock IP tunnel after, carry out response processing, and packet is directly sent to client by the outer net address of oneself and does not have to pass through
Cross LB server.DR mode is that real server is issued in request by the destination-mac address of overwrite request message, and true
Processing result after real server response is directly returned to client user, and the same with TUN mode, DR mode can be mentioned greatly
The retractility of high group system, and DR mode does not have the expense of IP tunnel, is not necessarily to the real server in cluster yet
It must support the requirement of IP tunneling protocol, but it is same to require scheduler LB and real server RS to have one piece of network interface card to be connected to
On physics network segment, it is necessary to which, in the same LAN environment, DR mode is to use more one mode in Internet enterprises.
FULLNAT mode forwarding data packet is similar NAT mode, IN and OUT data packet is all by LVS, only difference is that rear end
RealServer or interchanger do not need to do any configuration, and the cardinal principle of FULLNAT mode is to introduce local address
(address Intranet ip), cip-vip is converted to lip- > rip, and lip and rip are IDC Intranet ip, can communicate across three layers.
Four kinds of modes of LVS can reasonably select one of which according to own service demand in enterprise applies and be disposed
Using, however for the service of the four-layer load-equalizing of publicly-owned cloud network, there is corresponding lack in above-mentioned four kinds of modes
It falls into, it can not be directly as the four-layer load-equalizing implementation of publicly-owned cloud network.NAT mode is required all rear end RS servers
Default gateway be set as the IP of LB node, this is unpractical for the load balanced product of publicly-owned cloud network；TUN mould
Formula one side will increase the encapsulation overhead of message, on the other hand since TUN mode needs back-end server to support IP
Tunneling agreement, which limits the operating systems that rear end RS server may be confined to some types；DR mode needs
VIP is configured to on all rear end RS servers and is closed the arp reply of the vip interface, this just needs to modify all rear ends
Configuration on RS server, this is also unpractical for the four-layer load-equalizing product of publicly-owned cloud network；FULLNAT mould
Although formula is not necessarily to set the default gateway of rear end RS server to the IP of LB node, since it has modified the source of message
IP, which results in the IP of client can not pass to the application on the RS server of rear end, and which limits answering for load balancing
With scene, it is also not suitable for the four-layer load-equalizing implementation as publicly-owned cloud network.
For the four-layer load-equalizing problem of implementation of current publicly-owned cloud network, it is negative that the invention proposes a kind of publicly-owned cloud networks
Carry balanced system architecture and implementation method.This method it is not necessary to modify the configuration of rear end RS server, without additional package expense and
Particular protocol requirement, and do not destroy the transparency of session, i.e. client ip is visible to rear end RS server.
Summary of the invention
In order to compensate for the shortcomings of the prior art, the present invention provides a kind of public cloud Network Load Balance systems being simple and efficient
System framework and implementation method.
The present invention is achieved through the following technical solutions:
A kind of public cloud Network Load Balance system architecture, it is characterised in that: including northbound interface module, LB manager mould
Block, policybased routing manager module, LB configuration database, LB proxy module and LVS module；Wherein, the northbound interface module,
LB manager module, policybased routing manager module and LB configuration database are located at control plane, the LB proxy module and LVS
Module is located at data plane.
The LB that the northbound interface module is used to provide restful for publicly-owned cloud platform or third party cloud pipe platform, which is managed, to be connect
Mouthful；
The LB manager module is responsible for receiving the request message from northbound interface module, provides scheduler LB example
Creation is deleted, modification and inquiry, while providing the creation of monitor, is deleted, modification and query function, in the demand of High Availabitity
It is lower that the function of health examination is provided, and empty machine where dynamic creation scheduler LB example on demand；
The policybased routing manager module is responsible for the configuration information according to scheduler LB example, calls publicly-owned cloud management flat
It is corresponding to be issued to VPC by the policybased routing management interface of platform for the policybased routing information of configuration needed for current scheduling device LB example
On virtual router qr, the reversed routing iinformation for leading to scheduler LB node is provided for the response message on the RS server of rear end；
The LB configuration database is used for the relevant configuration information of persistence recording dispatching device LB, including scheduler LB example
Itself relevant information, the relevant configuration information of monitor and the relevant configuration information of health examination；
The LB proxy module is located on the empty machine where LB example, is responsible for interacting with LB manager module, be matched according to LB
The configuration information in database is set, the relevant parameter of configuration LVS realizes four layers of LB function；
The LVS module is the LVS project increased income, and is configured to NAT mode, operates in the void where scheduler LB example
On machine, LB proxy module is according to the message of LB manager module come the relevant parameter of dynamic configuration LVS.
The LB proxy module also needs to carry out heartbeat with LB manager module to shake hands, and LB manager module is believed by heartbeat
Whether node where breath judges scheduler LB example is working properly, if discovery is abnormal, LB manager module is needed to adjust again
The empty machine of degree creation one, and relevant scheduler LB example information is reconfigured, realize the automatically restoring fault of scheduler LB.
The present invention is based on the implementation methods of the public cloud Network Load Balance system architecture, comprising the following steps:
The first step creates one for running the empty machine of LB example in VPC；
Second step, with specific reference to public cloud scheduler LB deployment mode give empty machine NIC address LIP bind one or
Multiple resilient ips, wherein a resilient ip can only be bound by exclusively enjoying mode, shared model can then bind multiple resilient ips；
Third step externally provides service using resilient ip as the virtual address VIP of scheduler LB example；
4th step adds rear end RS server to scheduler LB example, runs business, the RS server on RS server
The NIC address of NIC address and scheduler LB example is located in same VPC；
5th step, by scheduler LB exemplary configuration at the NAT mode of LVS；
6th step, configuration strategy routes on the virtual router of VPC, and the NAT reverse flow of scheduler LB is led tune
It spends on device LB node, the replacement of source IP address is carried out by scheduler LB node, realize the two-way intercommunication of LB flow.
In the first step, comprising the following steps:
(1) user creates four-layer load-equalizing example by public cloud console or OpenAPI interface, and request to create is final
Northbound interface module creation scheduler LB example is called, and passes to LB manager module, and scheduler LB essential information is entered
Library；
(2) LB manager module is checked and is currently then created in VPC first if it does not exist with the presence or absence of scheduler LB void machine
A scheduler LB void machine is built, if existing scheduler LB void machine, from current all active scheduler LB void machine nodes
Dispatch a scheduler LB void machine node；
(3) when user creates LB monitor, LB monitor configuration information is handed down to by LB manager module to be dispatched to
Scheduler LB agency on scheduler LB void machine node, while by the relevant configuration information typing LB configuration database of monitor；
(4) LB agency is responsible for that the relevant configuration of LVS is arranged according to configuration information, and four-layer load-equalizing business is made to come into force；
(5) LB manager module trigger policy routing manager module is according to LB monitor configuration information to the virtual of VPC
The routing of router distributing policy, setting completed for the configuration of four-layer load-equalizing.
In the third step, when client is initiated to communicate, source IP is the IP, i.e. CIP of client itself in message；Purpose
IP is the IP, i.e. virtual address VIP of LB monitor；The virtual address VIP is the EIP being bundled in scheduler LB example.
In 6th step, when message reaches scheduler LB node, DNAT replacement is carried out by scheduler LB node, will be reported
Destination IP in text is substituted for the IP of some rear end RS server from virtual address VIP, and then positive message can correctly be sent
Up to rear end RS server；
In the response message issued for reversed message, rear end RS server, source IP is the IP of rear end RS server itself,
Destination IP is client ip, i.e. CIP；Since CIP is not in current sub network, message can be sent to default gateway, i.e. virtual flow-line
On device qr, message can be directed on scheduler LB node by the policybased routing configured on virtual router qr, scheduler LB node
After carrying out SNAT processing to message, the source IP of message is replaced with into virtual address VIP, and sent back on virtual router qr again,
Message goes out public network by default route at this time, and is finally sent to client, and then establish two-way communication.
For different dispatching algorithms, the flow of different clients can be scheduled on different rear end RS servers, from
And realize four-layer load-equalizing, ensure the High Availabitity and high-performance of business.
The beneficial effects of the present invention are: the public cloud Network Load Balance system architecture and implementation method, do not destroy session
The transparency, client ip address is to rear end RS server as it can be seen that not needing configuration and the rear end of modification rear end RS server simultaneously
RS server zero configuration does not need additional agreement support and additional protocol encapsulation expense yet, realizes publicly-owned cloud network four
Layer load balancing.
Detailed description of the invention
Attached drawing 1 is public cloud Network Load Balance implementation method schematic diagram of the present invention.
Attached drawing 2 is public cloud Network Load Balance system architecture schematic diagram of the present invention.
Specific embodiment
In order to which technical problems, technical solutions and advantages to be solved are more clearly understood, tie below
Embodiment is closed, the present invention will be described in detail.It should be noted that specific embodiment described herein is only to explain
The present invention is not intended to limit the present invention.
The public cloud Network Load Balance system architecture, including northbound interface module, LB (Load Balancer) manager
Module, policybased routing manager module, LB configuration database, LB proxy module and LVS module；Wherein, the northbound interface mould
Block, LB manager module, policybased routing manager module and LB configuration database be located at control plane, the LB proxy module and
LVS module is located at data plane.
The LB that the northbound interface module is used to provide restful for publicly-owned cloud platform or third party cloud pipe platform, which is managed, to be connect
Mouthful；
The LB manager module is responsible for receiving the request message from northbound interface module, provides scheduler LB example
Creation is deleted, modification and inquiry, while providing the creation of monitor, is deleted, modification and query function, in the demand of High Availabitity
It is lower that the function of health examination is provided, and empty machine where dynamic creation scheduler LB example on demand；
The policybased routing manager module is responsible for the configuration information according to scheduler LB example, calls publicly-owned cloud management flat
It is corresponding to be issued to VPC by the policybased routing management interface of platform for the policybased routing information of configuration needed for current scheduling device LB example
On virtual router qr (Router of VPC), is provided for the response message on the RS server of rear end and reversely lead to scheduler LB
The routing iinformation of node；
The LB configuration database is used for the relevant configuration information of persistence recording dispatching device LB, including scheduler LB example
Itself relevant information, the relevant configuration information of monitor and the relevant configuration information of health examination；
The LB proxy module is located on the empty machine where LB example, is responsible for interacting with LB manager module, be matched according to LB
The configuration information in database is set, the relevant parameter of configuration LVS realizes four layers of LB function；
The LVS module is the LVS project increased income, and is configured to NAT mode, operates in the void where scheduler LB example
On machine, LB proxy module is according to the message of LB manager module come the relevant parameter of dynamic configuration LVS.
The LB proxy module also needs to carry out heartbeat with LB manager module to shake hands, and LB manager module is believed by heartbeat
Whether node where breath judges scheduler LB example is working properly, if discovery is abnormal, LB manager module is needed to adjust again
The empty machine of degree creation one, and relevant scheduler LB example information is reconfigured, realize the automatically restoring fault of scheduler LB.
Implementation method based on the public cloud Network Load Balance system architecture, comprising the following steps:
The first step creates one for running the empty machine of LB example in VPC；
Second step gives the NIC address LIP (Loacl IP) of empty machine with specific reference to the deployment mode of public cloud scheduler LB
One or more resilient ips are bound, wherein a resilient ip can only be bound by exclusively enjoying mode, shared model can then bind multiple
Resilient ip；
Third step externally provides service using resilient ip as the virtual address VIP of scheduler LB example；
4th step adds rear end RS (Real Server) server to scheduler LB example, runs industry on RS server
The NIC address of business, the RS server NIC address and scheduler LB example is located in same VPC；
5th step, by scheduler LB exemplary configuration at the NAT mode of LVS；
6th step, configuration strategy routes on the virtual router of VPC, and the NAT reverse flow of scheduler LB is led tune
It spends on device LB node, the replacement of source IP address is carried out by scheduler LB node, realize the two-way intercommunication of LB flow.
In the first step, comprising the following steps:
(1) user creates four-layer load-equalizing example by public cloud console or OpenAPI interface, and request to create is final
Northbound interface module creation scheduler LB example is called, and passes to LB manager module, and scheduler LB essential information is entered
Library；
(2) LB manager module is checked and is currently then created in VPC first if it does not exist with the presence or absence of scheduler LB void machine
A scheduler LB void machine is built, if existing scheduler LB void machine, from current all active scheduler LB void machine nodes
Dispatch a scheduler LB void machine node；
(3) when user creates LB monitor, LB monitor configuration information is handed down to by LB manager module to be dispatched to
Scheduler LB agency on scheduler LB void machine node, while by the relevant configuration information typing LB configuration database of monitor；
(4) LB agency is responsible for that the relevant configuration of LVS is arranged according to configuration information, and four-layer load-equalizing business is made to come into force；
(5) LB manager module trigger policy routing manager module is according to LB monitor configuration information to the virtual of VPC
The routing of router distributing policy, setting completed for the configuration of four-layer load-equalizing.
In the third step, when client is initiated to communicate, source IP is the IP, i.e. CIP of client itself in message
(Client IP)；Destination IP is the IP, i.e. virtual address VIP (Virtual IP) of LB monitor；The virtual address VIP is
The EIP being bundled in scheduler LB example.
In 6th step, when message reaches scheduler LB node, DNAT is carried out by scheduler LB node
(destination network address translation, destination network addresses conversion) replacement, by the purpose in message
IP is substituted for the IP of some rear end RS server from virtual address VIP, and (such as RIP2 typically relies on the knot of LB algorithmic dispatching
Fruit), then positive message can correctly be sent to rear end RS server；
In the response message issued for reversed message, rear end RS server, source IP is the IP of rear end RS server itself,
Destination IP is client ip, i.e. CIP；Since CIP is not in current sub network, message can be sent to default gateway, i.e. virtual flow-line
On device qr, the policybased routing that is configured on virtual router qr (srcIP=RIP1/RIP2/RIP3, srcPort=RPORT,
Nexthop=LIP) message can be directed on scheduler LB node, scheduler LB node carries out SNAT to message, and (source address turns
Change) processing after, the source IP of message is replaced with into virtual address VIP, and sent back on virtual router qr again, at this time message pass through
Default route goes out public network, and is finally sent to client, and then establishes two-way communication.
For different dispatching algorithms, the flow of different clients may be scheduled for different rear end RS servers
On, to realize four-layer load-equalizing, ensure the High Availabitity and high-performance of business.
Embodiment described above, only one kind of the specific embodiment of the invention, those skilled in the art is in this hair
The usual variations and alternatives carried out in bright technical proposal scope should be all included within the scope of the present invention.
Claims (8)
1. a kind of public cloud Network Load Balance system architecture, it is characterised in that: including northbound interface module, LB manager mould
Block, policybased routing manager module, LB configuration database, LB proxy module and LVS module；Wherein, the northbound interface module,
LB manager module, policybased routing manager module and LB configuration database are located at control plane, the LB proxy module and LVS
Module is located at data plane.
2. public cloud Network Load Balance system architecture according to claim 1, it is characterised in that:
The northbound interface module is used to provide the LB management interface of restful for publicly-owned cloud platform or third party cloud pipe platform；
The LB manager module is responsible for receiving the request message from northbound interface module, provides the wound of scheduler LB example
It builds, deletes, modification and inquiry, while the creation of monitor being provided, it deletes, modification and query function, under the demand of High Availabitity
The function of health examination is provided, and empty machine where dynamic creation scheduler LB example on demand；
The policybased routing manager module is responsible for the configuration information according to scheduler LB example, calls publicly-owned cloud management platform
It is corresponding virtual to be issued to VPC by policybased routing management interface for the policybased routing information of configuration needed for current scheduling device LB example
On router qr, the reversed routing iinformation for leading to scheduler LB node is provided for the response message on the RS server of rear end；
The LB configuration database is used for persistence recording dispatching device LB relevant configuration information, including scheduler LB example itself
Relevant information, the relevant configuration information of monitor and the relevant configuration information of health examination；
The LB proxy module is located on the empty machine where LB example, is responsible for interacting with LB manager module, according to LB configuration number
According to the configuration information in library, the relevant parameter of LVS is configured, realizes four layers of LB function；
The LVS module is the LVS project increased income, and is configured to NAT mode, operates in the empty machine where scheduler LB example
On, LB proxy module is according to the message of LB manager module come the relevant parameter of dynamic configuration LVS.
3. public cloud Network Load Balance system architecture according to claim 2, it is characterised in that: the LB proxy module
It also needs to carry out heartbeat with LB manager module to shake hands, LB manager module is judged by heartbeat message where scheduler LB example
Whether node is working properly, if discovery is abnormal, needs LB manager module to reschedule creation one empty machine, and match again
Relevant scheduler LB example information is set, realizes the automatically restoring fault of scheduler LB.
4. the implementation method of public cloud Network Load Balance system architecture according to claims 1 to 3, which is characterized in that
The following steps are included:
The first step creates one for running the empty machine of LB example in VPC；
Second step gives the NIC address LIP of empty machine to bind one or more with specific reference to the deployment mode of public cloud scheduler LB
Resilient ip, wherein a resilient ip can only be bound by exclusively enjoying mode, shared model can then bind multiple resilient ips；
Third step externally provides service using resilient ip as the virtual address VIP of scheduler LB example；
4th step adds rear end RS server to scheduler LB example, runs business, the RS server network interface card on RS server
The NIC address of address and scheduler LB example is located in same VPC；
5th step, by scheduler LB exemplary configuration at the NAT mode of LVS；
6th step, configuration strategy routes on the virtual router of VPC, and the NAT reverse flow of scheduler LB is led scheduler
On LB node, the replacement of source IP address is carried out by scheduler LB node, realizes the two-way intercommunication of LB flow.
5. the implementation method of public cloud Network Load Balance system architecture according to claim 4, which is characterized in that described
In the first step, comprising the following steps:
(1) user creates four-layer load-equalizing example by public cloud console or OpenAPI interface, and request to create is finally called
Northbound interface module creation scheduler LB example, and LB manager module is passed to, and scheduler LB essential information is put in storage；
(2) LB manager module checks currently then creates one if it does not exist with the presence or absence of scheduler LB void machine in VPC first
A scheduler LB void machine is dispatched from current all active scheduler LB void machine nodes if existing scheduler LB void machine
One scheduler LB void machine node；
(3) when user creates LB monitor, LB monitor configuration information is handed down to the scheduling being dispatched to by LB manager module
Scheduler LB agency on device LB void machine node, while by the relevant configuration information typing LB configuration database of monitor；
(4) LB agency is responsible for that the relevant configuration of LVS is arranged according to configuration information, and four-layer load-equalizing business is made to come into force；
(5) LB manager module trigger policy routing manager module is according to LB monitor configuration information to the virtual flow-line of VPC
The routing of device distributing policy, setting completed for the configuration of four-layer load-equalizing.
6. the implementation method of public cloud Network Load Balance system architecture according to claim 4, it is characterised in that: described
In third step, when client is initiated to communicate, source IP is the IP, i.e. CIP of client itself in message；Destination IP is LB monitor
IP, i.e. virtual address VIP；The virtual address VIP is the EIP being bundled in scheduler LB example.
7. the implementation method of public cloud Network Load Balance system architecture according to claim 4, it is characterised in that: described
In 6th step, when message reaches scheduler LB node, DNAT replacement is carried out by scheduler LB node, by the destination IP in message
The IP of some rear end RS server is substituted for from virtual address VIP, then positive message can correctly be sent to rear end RS service
Device；
In the response message issued for reversed message, rear end RS server, source IP is the IP of rear end RS server itself, purpose
IP is client ip, i.e. CIP；Since CIP is not in current sub network, message can be sent to default gateway, i.e. virtual router qr
On, message can be directed on scheduler LB node by the policybased routing configured on virtual router qr, and scheduler LB node is to report
After text carries out SNAT processing, the source IP of message is replaced with into virtual address VIP, and sent back on virtual router qr again, at this time
Message goes out public network by default route, and is finally sent to client, and then establish two-way communication.
8. the implementation method of public cloud Network Load Balance system architecture according to claim 7, it is characterised in that: for
Different dispatching algorithms, the flow of different clients can be scheduled on different rear end RS servers, to realize that four layers are born
Equilibrium is carried, ensures the High Availabitity and high-performance of business.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910666177.8A CN110392108A (en) | 2019-07-23 | 2019-07-23 | A kind of public cloud Network Load Balance system architecture and implementation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910666177.8A CN110392108A (en) | 2019-07-23 | 2019-07-23 | A kind of public cloud Network Load Balance system architecture and implementation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110392108A true CN110392108A (en) | 2019-10-29 |
Family
ID=68287220
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910666177.8A Pending CN110392108A (en) | 2019-07-23 | 2019-07-23 | A kind of public cloud Network Load Balance system architecture and implementation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110392108A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111010459A (en) * | 2019-12-06 | 2020-04-14 | 紫光云(南京)数字技术有限公司 | Method for solving IP address conflict when SLB is used as SNAT |
| CN111107178A (en) * | 2019-12-29 | 2020-05-05 | 苏州浪潮智能科技有限公司 | Method and equipment for assigning message to use local address |
| CN111262960A (en) * | 2020-01-15 | 2020-06-09 | 山东汇贸电子口岸有限公司 | Public cloud elastic public network IP cluster implementation method and system |
| CN111262924A (en) * | 2020-01-14 | 2020-06-09 | 安超云软件有限公司 | Software and hardware combined network load balancing implementation system and process method |
| CN111274027A (en) * | 2020-01-09 | 2020-06-12 | 山东汇贸电子口岸有限公司 | Multi-live load balancing method and system applied to openstack cloud platform |
| CN111314368A (en) * | 2020-02-27 | 2020-06-19 | 紫光云技术有限公司 | Method for realizing tube renting intercommunication by using load balancer |
| CN111371685A (en) * | 2020-02-28 | 2020-07-03 | 北京百度网讯科技有限公司 | Data processing and IPv6 mounting method and device |
| CN111756830A (en) * | 2020-06-22 | 2020-10-09 | 浪潮云信息技术股份公司 | Internal network load balancing implementation method of public cloud network |
| CN112073503A (en) * | 2020-09-03 | 2020-12-11 | 浪潮云信息技术股份公司 | High-performance load balancing method based on flow control mechanism |
| CN112398696A (en) * | 2020-11-30 | 2021-02-23 | 迈普通信技术股份有限公司 | System and method for realizing zero configuration opening |
| CN112732530A (en) * | 2021-01-12 | 2021-04-30 | 杭州萤石软件有限公司 | LVS operation and maintenance control system, method and device |
| CN113098728A (en) * | 2019-12-23 | 2021-07-09 | 华为技术有限公司 | Health check method of load balancing system and related equipment |
| CN114553678A (en) * | 2022-02-09 | 2022-05-27 | 紫光云(南京)数字技术有限公司 | Diagnosis method for soft SLB traffic problem of cloud network |
| CN114760246A (en) * | 2022-03-29 | 2022-07-15 | 浪潮云信息技术股份公司 | Service drainage method, device and medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104850450A (en) * | 2015-05-14 | 2015-08-19 | 华中科技大学 | Load balancing method and system facing mixed cloud application |
| CN106506700A (en) * | 2016-12-28 | 2017-03-15 | 北京优帆科技有限公司 | A kind of transparent proxy method of load equalizer and SiteServer LBS |
| CN109660466A (en) * | 2019-02-26 | 2019-04-19 | 浪潮软件集团有限公司 | A kind of more live load balance realizing methods towards cloud data center tenant |
-
2019
- 2019-07-23 CN CN201910666177.8A patent/CN110392108A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104850450A (en) * | 2015-05-14 | 2015-08-19 | 华中科技大学 | Load balancing method and system facing mixed cloud application |
| CN106506700A (en) * | 2016-12-28 | 2017-03-15 | 北京优帆科技有限公司 | A kind of transparent proxy method of load equalizer and SiteServer LBS |
| CN109660466A (en) * | 2019-02-26 | 2019-04-19 | 浪潮软件集团有限公司 | A kind of more live load balance realizing methods towards cloud data center tenant |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111010459A (en) * | 2019-12-06 | 2020-04-14 | 紫光云(南京)数字技术有限公司 | Method for solving IP address conflict when SLB is used as SNAT |
| CN113098728A (en) * | 2019-12-23 | 2021-07-09 | 华为技术有限公司 | Health check method of load balancing system and related equipment |
| CN113098728B (en) * | 2019-12-23 | 2023-12-19 | 华为云计算技术有限公司 | Health check method of load balancing system and related equipment |
| CN111107178A (en) * | 2019-12-29 | 2020-05-05 | 苏州浪潮智能科技有限公司 | Method and equipment for assigning message to use local address |
| CN111274027A (en) * | 2020-01-09 | 2020-06-12 | 山东汇贸电子口岸有限公司 | Multi-live load balancing method and system applied to openstack cloud platform |
| CN111262924A (en) * | 2020-01-14 | 2020-06-09 | 安超云软件有限公司 | Software and hardware combined network load balancing implementation system and process method |
| CN111262960A (en) * | 2020-01-15 | 2020-06-09 | 山东汇贸电子口岸有限公司 | Public cloud elastic public network IP cluster implementation method and system |
| CN111314368A (en) * | 2020-02-27 | 2020-06-19 | 紫光云技术有限公司 | Method for realizing tube renting intercommunication by using load balancer |
| CN111314368B (en) * | 2020-02-27 | 2022-06-07 | 紫光云技术有限公司 | Method for realizing tube renting intercommunication by using load balancer |
| CN111371685A (en) * | 2020-02-28 | 2020-07-03 | 北京百度网讯科技有限公司 | Data processing and IPv6 mounting method and device |
| CN111756830A (en) * | 2020-06-22 | 2020-10-09 | 浪潮云信息技术股份公司 | Internal network load balancing implementation method of public cloud network |
| CN112073503A (en) * | 2020-09-03 | 2020-12-11 | 浪潮云信息技术股份公司 | High-performance load balancing method based on flow control mechanism |
| CN112398696A (en) * | 2020-11-30 | 2021-02-23 | 迈普通信技术股份有限公司 | System and method for realizing zero configuration opening |
| CN112732530B (en) * | 2021-01-12 | 2022-05-24 | 杭州萤石软件有限公司 | LVS operation and maintenance control system, method and device |
| CN112732530A (en) * | 2021-01-12 | 2021-04-30 | 杭州萤石软件有限公司 | LVS operation and maintenance control system, method and device |
| CN114553678A (en) * | 2022-02-09 | 2022-05-27 | 紫光云(南京)数字技术有限公司 | Diagnosis method for soft SLB traffic problem of cloud network |
| CN114553678B (en) * | 2022-02-09 | 2024-02-13 | 紫光云(南京)数字技术有限公司 | Cloud network soft SLB flow problem diagnosis method |
| CN114760246A (en) * | 2022-03-29 | 2022-07-15 | 浪潮云信息技术股份公司 | Service drainage method, device and medium |
| CN114760246B (en) * | 2022-03-29 | 2024-05-03 | 浪潮云信息技术股份公司 | Service drainage method, device and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110392108A (en) | A kind of public cloud Network Load Balance system architecture and implementation method | |
| US10547544B2 (en) | Network fabric overlay | |
| US11025525B1 (en) | Methods and apparatus related to a virtual multi-hop network topology emulated within a data center | |
| JP5961718B2 (en) | Network architecture with middlebox | |
| TWI531185B (en) | Distributed virtual network gateways | |
| Cardellini et al. | The state of the art in locally distributed web-server systems | |
| CN104584491B (en) | Distributed virtual route and the system and method for exchanging (DVRS) are provided | |
| US7751409B1 (en) | Logical service domains for enabling network mobility | |
| CN110120934A (en) | Method, software defined network controller and the medium of application firewall strategy | |
| CN116319541A (en) | Service insertion method, device and system at logic gateway | |
| CN107947961A (en) | Kubernetes Network Management System and method based on SDN | |
| US10198338B2 (en) | System and method of generating data center alarms for missing events | |
| CN109639557A (en) | Methods, devices and systems for network communication | |
| US11956141B2 (en) | Service chaining with physical network functions and virtualized network functions | |
| WO2020209099A1 (en) | Communication system and communication method | |
| CN109474713A (en) | Message forwarding method and device | |
| US20030179775A1 (en) | Service delivery network system and method | |
| CN116743585A (en) | Multi-tenant API gateway service exposure system and method based on cloud protogenesis | |
| US20220141080A1 (en) | Availability-enhancing gateways for network traffic in virtualized computing environments | |
| AU2018204247B2 (en) | Architecture of networks with middleboxes | |
| US11516125B2 (en) | Handling packets travelling towards logical service routers (SRs) for active-active stateful service insertion | |
| Paul et al. | OpenADN: a case for open application delivery networking | |
| Hu et al. | L2OVX: an on-demand VPLS service with software-defined networks | |
| US12052171B2 (en) | Communication system and communication method | |
| US20230140555A1 (en) | Transparent network service chaining |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191029 |