CN102130776A - Communication method and system - Google Patents
Communication method and system Download PDFInfo
- Publication number
- CN102130776A CN102130776A CN2010100025994A CN201010002599A CN102130776A CN 102130776 A CN102130776 A CN 102130776A CN 2010100025994 A CN2010100025994 A CN 2010100025994A CN 201010002599 A CN201010002599 A CN 201010002599A CN 102130776 A CN102130776 A CN 102130776A
- Authority
- CN
- China
- Prior art keywords
- management server
- server
- application server
- switch
- user
- 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.)
- Granted
Links
Images
Abstract
The embodiment of the invention provides a communication method and system, which can solve the problems of low communication efficiency and high cost existing in the conventional communication system. The communication system comprises a switch, an application server and a management server, wherein the application server is used for receiving a user request forwarded by the switch and/or the management server, processing the request and transmitting a processing result to the switch; the switch is used for switching data; and the management server is used for monitoring the status of the application server and controlling the data forwarding of the switch according to the status of the application server and a preset rule, so that the switch can forward the user request transmitted to a failed application server to the management server when the application server fails. According to the invention, a server load balancing function, a server failure shielding function, a self-failure shielding function, a user transparent mode accessing function and a failure recovery self-healing function can be realized, and the contradiction between the improvement of switching performance and increasing network flow is thoroughly eliminated.
Description
Technical field
The present invention relates to a kind of communication technology, relate in particular to communication means and system in a kind of internet arena.
Background technology
Along with the fast development of Internet technology, the number of internet usage communication also by leaps and bounds soaring, at this moment, in the communication in internet arena, usually has many uncertain factors can cause communication failure, as, server failure etc.Safe and reliable for what guarantee to communicate by letter, in the prior art, following several scheme is arranged.
Scheme one, based on the load balancing of DNS (Domain Name System, domain name system) poll
As shown in Figure 1, in this scheme, each server in the server pools all is with an IP (Internet Protocol independently, Internet protocol) address provides service, during the domain name of the service that client computer is visited to the dns server request analysis via the Internet, core L2/L3 switch, dns server all returns to the user with the IP address of the whole servers in the server pools.Different client softwares can be selected the destination server address of an IP address as this access request according to the rule of self.Afterwards, client computer and this station server are made point-to-point communication.
Yet, in this scheme, there is following shortcoming:
1, neither one is used to check the equipment or the main frame of server service health status, and after server broke down, network manager can't be learnt this situation.
2, there is not communication-cooperation between server and the DNS service.When certain server broke down, the user still can obtain the IP address of failed server from the DNS service.Therefore, this user can't obtain normal service.
3, in normal maintenance work, from server pools, withdraw from the influence that server can produce short-term to the user.Because the Data Update of DNS service needs one than long time course with synchronous, the user may have access on the IP address of the server of having removed at this moment.
The inline of scheme two, commercial four layers of exchange scheme (work in the band) pattern
As shown in Figure 2, in this scheme, the inline pattern in commercial four layers of switching solution is to adopt the network configuration of " gateway " formula.Server uses two four layers (L4) switch, and wherein, a switch is assumed to be the L4 switch A as main usefulness; Another switch is assumed to be the L4 switch b as standby.Between two four layer switch (being L4 switch A and L4 switch b), need the cable of a special use to connect the network connection state that is used for upgrading synchronously between user terminal and the server.The IP address of four layer switch is the default gateway of self.Client computer is directed to the IP address of four layer switch to the visit of server by dns server by the Internet, core L2/L3 switch, after four layer switch are received access request, will select certain server from server pools according to certain round-robin.IP address with this server is a destination address, is source address with the user's IP address, and user's access request is transmitted to this server.Server is wanted result data is returned to the user, just must send the default gateway (four layer switch) of server to.Four layer switch receive that server will return to after the user's data, are source address with oneself IP address, and user's IP address is a destination address, and data are returned to the user.Like this, at user side, four layer switch are exactly server.
The access request of the regular analog subscriber of four layer switch meeting sends to the server in each station server pond, and the server of fault can be found by four layer switch thus.Four layer switch only can be forwarded to the request of real user on the normal server, and can be according to the health status of each server of configuration report of network manager.
Yet, in this scheme, there is following shortcoming:
Four layer switch must process user and server between the overall network flow, this requires four layer switch to have very high performance, also means very high cost.
2. the performance of four layer switch must increase along with the increase of user and server, means that also cost increases thereupon.
3. because the Single Point of Faliure risk of four layer switch A itself, so four layer switch B must remain with enough vacant performances constantly to tackle the contingent fault of four layer switch A.And if four layer switch B upward itself have four layers of Exchange Service of carrying other, four layer switch A also must keep enough vacant performances constantly to tackle the contingent fault of four layer switch B so.The total bearing capacity sum that is to say four layer switch A and four layer switch B can not surpass half of four layer switch A and four layer switch B performance sums.
4. the network traffics between user terminal and the server need twice through switch, and this makes the load of switch double.
The single armed pattern of scheme three, commercial four layers of exchange scheme
As shown in Figure 2, the single armed pattern in commercial four layers of switching solution is to adopt the network configuration that is similar to " unidirectional gateway ".The user is directed to the IP address of four layer switch by dns server to the visit of server.After four layer switch are received access request, to from server pools, select certain server according to certain round-robin, be source IP address then with the user's IP address, still be purpose IP address with original purpose IP address (IP of four layer switch itself), still the MAC Address with destination server is that destination address is transmitted this ethernet frame (frame).The special setting of server end is that the IP address of four layer switch itself (the purpose IP in the data message) is bundled on the network equipment of a non-broadcast-type of self, such as loopback (the winding network equipment) or dummy (the mute network equipment) kind equipment.Like this, because the purpose IP address in user's request originally also exists on one's body at server, server will be handled according to common normal flow after receiving the ethernet frame that four layer switch forward.Because the source IP address in the user that server the is seen request is real user's IP address itself, so the result data that server will return returns three-tier switch to core according to default route, and three-tier switch directly returns data according to routing table and gives the user.Four layers of exchange load balancing of " single armed pattern " like this, have just been realized.
The access request of the regular analog subscriber of four layer switch meeting sends to the server in each station server pond, exists the server of fault to be found by four layer switch thus.Four layer switch only can be forwarded to user's request on the server of operate as normal, and can be according to the health status of each server of configuration report of network manager.
Because in this network configuration, itself there is the Single Point of Faliure risk in four layer switch A, so need four layer switch B of another same performance specification to do reserve, also need the cable of a special use to connect the network connection state that is used for upgrading synchronously between user and the server between four layer switch A and the four layer switch B.
Yet, in this scheme, there is following shortcoming:
Four layer switch still must process user to server side to the overall network flow, this still requires four layer switch to have very high performance.
2. the performance of four layer switch still must increase along with the increase of user and server, means that also cost also increases thereupon.
3. itself still there is the Single Point of Faliure risk in four layer switch A, and four layer switch B must remain with enough vacant performances constantly to tackle the contingent fault of four layer switch A.And if four layer switch B go up four layers of Exchange Service of carrying other own, four layer switch A also must remain with enough vacant performances constantly to tackle the contingent fault of four layer switch B so.The total bearing capacity sum that is to say four layer switch A and four layer switch B can not surpass half of four layer switch A and four layer switch B performance sums.
4. as shown in Figure 2, the user to server side to network traffics need twice through switch, still make the load of switch increase.
In a word, there is the problem that communication efficiency is low, cost is high in existing communication system.
Summary of the invention
Embodiments of the invention provide a kind of communication means and device, can solve the problem that existing communication system communication efficient is low, cost is high.
Embodiments of the invention provide a kind of communication system, comprise switch, application server and management server.Described application server is used for user's request that desampler and/or management server are transmitted, and handles this request, and result is issued switch; Described switch is used for the result of application server is transmitted to the user, and under the control of management server user's request is forwarded to destination address corresponding application server or management server; Described management server, be used to monitor the state of application server, and,, make user's request that switch will mail to this fault application server be transmitted to management server so that when the application server fault according to the data forwarding that the state and the pre-defined rule of application server are controlled switch.
The embodiment of the invention also provides a kind of communication means, comprises step: switch receives user's request, and under the control of management server user's request is forwarded to destination address corresponding application server or management server; The state of management server monitoring application server, and according to the data forwarding of the state of application server and pre-defined rule control switch, so that when the application server fault, make user's request that switch will mail to this fault application server be transmitted to management server; User's request that application server desampler and/or management server are transmitted, and handle this request is issued switch with result, and by switch the result of application server is transmitted to the user.
Embodiments of the invention are by the state of management server monitoring application server, and according to the data forwarding of the state of application server and pre-defined rule control switch, so that when the application server fault, make user's request that switch will mail to this fault application server be transmitted to management server, can realize load balancing, server failure shielding, faults itself shielding, user transparent mode access, five functions of fault recovery self-healing of server; Because in communication system of the present invention, in normal course of communications, management server is communication process between participating user terminal and the application server not, only when the application server fault, just participate in the management work of communication system, therefore, the load of communication system of the present invention self has nothing to do with number of users, visit capacity, irrelevant with the quantity of application server, only relevant with the failure rate and the failure recovery time of server; So thoroughly solve the lifting of switching performance and the contradiction between the growing network traffics, opened up a novel application model.
Description of drawings
Fig. 1 shows a kind of communication system of prior art;
Fig. 2 shows the another kind of communication system of prior art;
Fig. 3 shows the flow process of the active/standby relation negotiation election module of the embodiment of the invention;
Fig. 4 shows the flow process of the ethernet data frame forwarding module of the embodiment of the invention;
Fig. 5 shows the flow process of the ARP proxy module of the embodiment of the invention;
Fig. 6 is the flow process of the health examination module of the embodiment of the invention;
Fig. 7 is the communication system of the embodiment of the invention;
Fig. 8 is the communication means of the embodiment of the invention.
Embodiment
Understand and realization the present invention the existing embodiments of the invention of describing in conjunction with the accompanying drawings for the ease of persons skilled in the art.
Embodiment one
Present embodiment provides a kind of communication system, and it comprises application server, switch and management server, introduces their structure and logical relation below respectively.
Application server is used for user's request that desampler and/or management server are transmitted, and handles this request, and result is issued switch.
Application server is at least one, and preferably, in embodiments of the present invention, application server is many, and the function of these servers is identical, is the network user service is provided.And the function of concrete service can go to realize according to the deployment corresponding application is installed.Application server is connected to other network via switch, such as INTERNET (the Internet), all is example with INTERNET below this paper.User's request of initiating from the network user on the INTERNET is transmitted to application server by switch, and the data result that server returns also returns to the user on the distal end I NTERNET via switch.
On the software level, the configuration separated into two parts of application server, a part is a service configuration, another part is an administration configuration.Service configuration is meant the software arrangements of the request return results data that server response user sends.In communication system of the present invention, the user still needs to obtain from the DNS service IP address list of the application server of respective service.We are named as ServiceIPs at this with these IP, are called for short SVCIP, and SVCIP and application server are not one to one, and for example, a plurality of or SVCIP can corresponding application server.When in the DNS service only assignment during a SVCIP, all to dispose this IP address on the application server of each in server pools; When assignment in the DNS service during a plurality of SVCIP, each application server all will dispose whole SVCIP.Need to prove that for fear of IP address conflict, application server is bundled in SVCIP (one or more) on the network interface of its non-broadcast-type, such as loopback equipment or dummy equipment.Like this, each application server can both be to being that response is made in the request of purpose with any SVCIP.The SVCIP that disposes on each application server is the same, no matter what SVCIP are arranged, in order to distinguish each application server, with regard to each application server, the IP that must dispose a difference and other application servers again is used to locate a certain application server.We are named these a few IP at this and are ServerIPs, are called for short SVRIP.Each application server all has oneself independently SVRIP, and has only one.SVRIP is bearing whole management work, must be bundled on the physical network interface of broadcast-type of application server, and application server uses this network interface to be connected to switch, and SVRIP must be in the different subnet (broadcast domain) with SVCIP.
Switch is used for the result of application server is transmitted to the user, and under the control of management server user's request is forwarded to destination address corresponding application server or management server.
Switch is the private network device with network data function of exchange.Switch links together the various network services device on the one hand, and is interconnected with other network (such as INTERNET) again on the other hand.In communication system of the present invention, switch makes that with application server and management server and INTERNET connection the three can intercommunication mutually under the prerequisite of " controlled ".The controlled destination of message transmission that is meant can be changed by switch, and concrete change strategy is to determine jointly according to the concrete operating state of initial configuration and each equipment.
On the software engineering level, the network insertion of all application servers is all finally incorporated among the same VLAN (Virtual Local Area Network, VLAN) on the switch.Be tied with two IP addresses on this VLAN,, delimit out two IP subnets according to they corresponding masks.These two subnets are used for covering respectively SVRIP and SVCIP, and we are called Server subnet to the subnet that covers SVRIP, are called for short SVRNET; The subnet that covers SVCIP is called Service subnet, is called for short SVCNET.Two IP on the VLAN also just become the default gateway address of these two subnets, are named respectively to be called Server GateWay and Service GateWay, are called for short SVRGW and SVCGW.Need to prove, though be on same VLAN, SVRNET and SVCNET without any lap, they are two independently broadcast domains.Sum up, SVRIP belongs to SVRNET, and its default gateway IP is SVRGW; SVCIP belongs to SVCNET, and its default gateway IP is SVCGW; These two IP subnets are all on same VLAN.
Management server, be used to monitor the state of application server, and,, make user's request that switch will mail to this fault application server be transmitted to management server so that when the application server fault according to the data forwarding that the state and the pre-defined rule of application server are controlled switch.
Management server is one or more server apparatus that are independent of application server, and management server is the core of whole system, and they carry out command scheduling to whole system operation.At any time, management server is all in the working condition of monitoring each equipment of whole system.When certain device fails, management server can be found this fault in the very first time, configured in one piece to system is made amendment, and utilizes self bearing capacity to remedy performance that this fault causes and the loss on the function, makes that the application service of total system is available continuously.In the view of the user, application service is not broken down or is interrupted, and for the system management guardian, can have time enough to change or repair the equipment of fault.
According to the embodiment of the invention, preferably, the quantity of management server can be selected arbitrarily between 2 to 254, typically is configured to 4.The hardware that constitutes management server can be the microcomputer of common X86 framework.The operating system software aspect, management server need be installed the operating system that disclosed core source code is provided, this requirement is to come from the kernel mode that the code of part management server applied logic need be operated in operating system, that is to say as the part of operating system kernel and moves.This need realize doing to expand revising at the partial function of existing operating system core code, can describe in detail hereinafter.
Aspect physical connection, management server uses a network interface to be connected to switch.Its residing VLAN is identical with application server.
Divide as follows according to functional module on the management server:
1, the active/standby relation between many management servers is consulted the election module
2, ethernet data frame forwarding module
3, ARP proxy module
4, health examination module
The process of each functional module is described with reference to the accompanying drawings.
As shown in Figure 3, when system adopts many management servers, management server comprises active/standby relation negotiation election module, it is used for the priority of each working resource on other this management server of management server multicast, and obtain the priority of each working resource on other management server by multicast packet, decide the operating state of each working resource on this management server according to the height of the priority of each working resource of each management server.The job step that active/standby relation is consulted the election module is as follows:
The priority of each working resource of step 301, each management server multicast oneself.Described working resource comprises: BCIP, VSIP.
Whether the priority of the working resource of step 303, each management server judgement oneself is the highest, if, execution in step 304, otherwise, execution in step 305.
As shown in Figure 4, the Frame forwarding module is used to obtain and transmit the ethernet frame that includes user's request.At first be to make up according to system's normal operating conditions to transmit vector table and connection status table, then receive health examination module in real time the more new data of transmitting vector table is also accepted its more new data to the connection status table of other management service in real time, and determine whether to carry out Frame according to the working resource state of this management server and transmit operation, also want multicast from change information simultaneously in the generation of connection status table.The same network that is suitable for other type of data forwarding mechanism of the present invention, the more new data packets of described forwarding vector table is drawn together BSVCIP, SVCIP, SVRIP and forwarding rule (promptly, corresponding relation between SVCIP and the SVRIP), the more new data packets of described connection status table is drawn together the TCP connection status between UDP session status, ICMP message question and answer state and user and the application server.Frame forwarding module job step is as follows:
User's request that step 404, desampler are transmitted, and according to certain algorithm, as poll, at random, the Hash mapping scheduling algorithm, according to transmitting vector table user's request data frame is carried out the forwarding operation, so that should be transmitted to normal application server by user's request that the fault application server is born.User's request can be determined according to SVCIP.
The renewal delta data of step 405, multicast connection status table is not so that can cause network to connect (connecting such as tcp) interruption in other management server adapter forwarding course of work when this management server fault.
As shown in Figure 5, the ARP proxy module receives the ARP query requests of other network equipment to certain IP address, and the MAC Address of this IP address correspondence returned this network equipment, described ARP inquiry is used to obtain the MAC Address corresponding with SVCIP, the described network equipment comprises switch, and the job step of ARP proxy module is as follows:
The correspondence table (SVCIP-MAC table) of step 502, structure SVCIP and MAC.
The message real-time update SVCIP-MAC table of step 503, reception health examination module.As, when SVCIP changed, the SVCIP-MAC table also will change.
As shown in Figure 6, health examination module is used to check the state of application server, to screen the application server of fault application server and operate as normal, by the ARP proxy module, originally the part IP that was corresponded to the MAC Address of fault application server among the SVCIP is replaced with the pairing MAC Address of the VSIP of management server, this replacement result finally is expressed in the ARP table of switch, and its job step is as follows:
Whole operation principles to system describes below:
Under normal (optimal) operating state, management server is not transmitted the communication data between user and the application server.Say in front and be, any application server can both provide service with the identity of any SVCIP, management server need ask guiding to cause a certain application server the user according to certain polling algorithm following of this situation to get final product so, and need not be concerned about the target ip address of user in asking.Management server uses standard A RP agreement to finish this work.When the user arrives at switch to the access request of a certain SVCIP, abide by RFC826, but exchange opportunity is broadcasted an ARP analysis request in the VLAN at the direct route port place of this SVCIP.Because SVCIP is bundled on the network interface of non-broadcast-type of application server, this interface can not received any broadcast message, so any application server can not give a response this ARP analysis request.Yet the network interface that management server is provided with a broadcast-type is connected with this VLAN, and like this, management server just can receive this ARP analysis request." ARP proxy module " on the management server started working then, abide by RFC1027, according to the corresponding relation of human configuration (relation between SVCIP and the application server) in advance, choose the MAC Address of network interface at the SVRIP place of application server and respond to switch.Switch is sent to user's request on the network interface at SVRIP place of this application server after obtaining the result.Because the purpose IP address (certain SVCIP) in user's request originally also exists on one's body at application server, application server will be handled according to common normal flow after receiving this ethernet frame.Because the source IP address in the user request that application server is seen is real user's IP itself, so the result data that application server will return returns to switch according to default route, and switch directly returns data to the user according to routing table.Like this, just realized the network service under the situation that management server does not participate in transmitting.
What need further specify is, exchange opportunity buffer memory ARP address translation table a period of time, switch buffer memory ARP address translation table for example of the present invention 4 hours, during this in, switch is broadcast polling but directly use the data cached of oneself more not.That is to say that management server inquires that at response ARP the resource consumption in this work is few, can ignore.The network interface of a broadcast-type on each management server all is connected to the VLAN at SVCNET place, and the IP that binds on this network interface belongs to SVCNET.Because the effect of this IP is to receive and respond the ARP broadcast request, we are referred to as Broadcast IP and are called for short BCIP.The BCIP of each management server can be the same or different, and explanation is arranged hereinafter.
When many management server collaborative works, respond conflict for fear of ARP, synchronization can only be handled from the ARP inquiry of switch by a management server and ask.This moment, management server solved this problem by " active/standby election negotiation module ".Active/standby election negotiation module is ethernet frame of multicast (MultiCast) at set intervals, and its recipient is configured to whole management servers, also just can receive this type of ethernet frame from other management server simultaneously.This mechanism has realized that " working resource " proprietorial election is consulted between many management servers, includes two parts information in this ethernet frame, and a part is a working resource, and another part is a priority.Typical working resource such as IP address, working resource can be to comprise an IP address, also can be to comprise a plurality of IP address.Each management server is configured to different numerical value in same working resource corresponding priorities.Like this, for some working resources, priority by making this management server and other management server multicast of receiving priority compare, management server can learn whether oneself is the management server that this working resource is had limit priority.If management server can be changed to active state with this working resource, handle all tasks relevant simultaneously with this resource; If not the highest priority, management server can be changed to resting state with this working resource, does not handle any task relevant with this resource simultaneously.
In this work of response ARP inquiry request, the working resource of management server is exactly BCIP.All be furnished with BCIP on each management server, BCIP can be the same or different.Have only the highest management server of priority just can " use " its BCIP, enter operating state, both responded ARP inquiry request.And the management server of other low priority can remain on BCIP separately " dormancy " state, neither handles any associated network information, comprises clean culture, multicast and broadcasting.The IP conflict so neither can occur, also can not produce work conflict.When the management server of limit priority breaks down certainly, it can not be again to the priority of other management server multicast oneself, originally on this resource of BCIP be time high priority management server by multicast (multicast) mode find it own present be the highest priority, so it activates the BCIP of oneself, enter operating state, in the multicast mode, management server exchanges election information, in a period of time, it may be several seconds, for certain working resource, the F8 server is received the numerical priority value that other management server multicast is come, if all less than himself priority level initializing value for this working resource, it is operating state that management server will activate or keep this working resource.
What need further specify is, active/standby election negotiation module is operated in the kernel state of operating system, and the network function module code of code and operating system is an one.The purpose of doing like this is in order to guarantee when the management server generating unit component software fault, the state that its application function part (for example responding ARP) and election funtion part can be consistent, both " can with can both using, can not with all not using ".Thereby avoided occurring management server and entered " electing winning " but state of disabler.
As mentioned before, on each management server, the network interface at BCIP place is connected to the VLAN at SVCNET place.And on this network interface, the MAC Address in BCIP and this network interface self firmware, also binding has extra IP address and MAC Address; These extra IP addresses belong to SVRNET, are named to be called Virtual Server IP, are called for short VSIP; And being named of extra MAC Address is called Virtual Server MAC, is called for short VSMAC.VSIP and VSMAC concern that one to one each is to all having constituted a working resource.Each management server all has " oneself " VSIP and VSMAC, also must have the whole VSIP and the VSMAC of other management server simultaneously.Remain use " active/standby election consult mould determine " between the management server and come these working resources of load sharing, the meaning of " oneself " in the last word is exactly that this resource correspondence priority in " active/standby election negotiation module " in default configuration is higher than being with other management server.
Each management server is all paid close attention to the operating state of each application server, and this function is called " health examination ".Management server is realized this function by health examination module is set, the health examination module of management server can be configured to termly the access request of analog subscriber is sent to each application server, and concrete access request use-case can be defined voluntarily by system operator.For the application server that can normally respond the health examination test, health examination module is aggregated into a tabulation with application server corresponding SVCIP in current configuration, and name is called Active SVCIP, is called for short ASVCIP.For out of order application server, health examination module also is aggregated into a tabulation with application server corresponding SVCIP in current configuration, and name is called Bad SVCIP, is called for short BSVCIP.Obviously:
∈ASVCIP∪∈BSVCIP＝∈SVCIP
According to preamble, " the ARP proxy module on the management server is according to the corresponding relation of human configuration in advance, chooses the MAC Address of network interface at the SVRIP place of application server and responds the ARP analysis request of switch to SVCIP.”。But when BSVCIP occurred, meaning that this application server breaks down can not provide service, meaned that also visit BSVCIP user is about to run into the mistake of " can't Connection Service device ".Run into this situation for fear of the user, the health examination module of management server must be handled this problem in the very first time, concrete processing method is: according to predetermined, a balanced as far as possible allocation order, BSVCIP is corresponded to VSMAC successively, and the BSVCIP quantity that its effect trends towards each VSMAC correspondence is the same; Afterwards, the new correspondence relationship information of this group is sent to the ARP proxy module; Meanwhile, the ASVCIP tabulation is sent to the Frame forwarding module.After the ARP proxy module is received lastest imformation, immediately its covering is merged in the SVCIP-MAC correspondence table of acquiescence, and, its objective is that the notice switch is that the network traffics of BSVCIP redirect to management server itself with the access destination address with the whole new SVCIP-MAC correspondence table of the form of ARP Advertising broadcasting.Meanwhile, the Frame forwarding module of management server has been received new ASVCIP tabulation, covers original available application services device tabulation with it immediately, according to the forwarding rule continuation forwarding work of himself.So far, health examination module has been finished a complete work period, and the next work period of circulation beginning.
The action of Frame forwarding module is to be to have (normally VSMAC) on own this machine and Target IP is the ethernet frame of SVCIP (normally BSVCIP) with all Destination MACs, and poll is transmitted to the server (application server) at the pairing MAC Address of ASVCIP place.The words is indigestion relatively, and popular saying is exactly " being transmitted to normal application server by the network service request that the fault application server is born ".As mentioned before, each application server all has been configured to can be with the state of any one SVCIP service, so application server can " not take notice of " that network access request is from the user or from management server, can " not take notice of " which SVCIP its Target IP is yet.The network traffics of returning then are to walk default route according to user's source IP directly to have given switch, and without management server.Like this, the Frame forwarding module only needs mechanically to transmit, and does not need to write down the state (as the state of TCP) that each network connects, and does not also need to handle the network traffics of returning, so the performance cost of management server seldom.What need further specify is, the Frame forwarding module is in running order always, and whether its source network traffics that will transmit exist fully by the ARP proxy module controls.
So, user capture has been forwarded on the server of ASVCIP equably to the network traffics of BSVCIP, service to the user is guaranteed, and management server self is also remaining minimum load---the network traffics of the approach axis of a handling failure server.
As shown in Figure 7, intactly explain orally the transition of each operating state below again.Suppose that pre-defined rule is:
Work catcher when default four management servers run into several application server faults is in proper order for D＞C＞B＞A＞D＞C... direction circulation, to reach load balancing as much as possible;
When default four management servers ran into faults itself, for the handover sequence (election priority) of overall this work of scheduling be: A met D, and B meets A, and C meets B, and D meets C;
When default four management servers ran into faults itself, for the work of the processing application server fault of originally bearing, its handover sequence (election priority) is: A met B, and B meets C, and C meets D, and D meets A.
In above-mentioned pre-defined rule, the work of accomplishing the processing server fault as far as possible is distributed uniform as far as possible, and during faults itself, overall traffic control joins backward, handles the then handing-over forward of work of application server fault.Intactly explain orally the transition of each operating state below.
One, normal condition
If initial condition is: armamentarium is working properly.The communication system course of work of the embodiment of the invention is as follows:
The scheduling inquiry of management server A response switch;
Switch directly is transmitted to application server to the network traffics of user terminal;
Application server returns the network traffics of returning to the user via switch;
Management server B, C, D is in idle condition.
Two, malfunction 1
If initial condition is: application server 1 breaks down.The communication system course of work of the embodiment of the invention is as follows:
Four management servers all monitor application server 1 and break down;
Be responsible for the overall work scheduling according to pre-defined rule management server A;
Management server A broadcasts corresponding ethernet address modification information on network, according to pre-defined rule, switch is forwarded to management server D with being about to the access request of user to fault application server 1;
Switch directly is transmitted to user capture to the source network traffics of normal use server the application server of operate as normal;
The user's that management server D transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch;
The application server of operate as normal returns the network traffics of returning to the user via switch;
Management server B, C is in idle condition.
Three, malfunction 2
If initial condition is: application server 1 breaks down; Management server A breaks down.The communication system course of work of the embodiment of the invention is as follows:
Management server B, C, D all monitor application server 1 and break down;
Replace the scheduling of management server A catcher overall work according to pre-defined rule management server B;
According to the work of the handling failure application server on the pre-defined rule management server D catcher management server A, this work comprises work unborn and that may be assigned with from now on; Management server D broadcasts corresponding ethernet address modification information on network, switch is transmitted to management server D with the network traffics that are about to user capture fault application server;
Switch still is forwarded to management server D with the network traffics of user capture fault application server 1;
Switch directly is transmitted to the flow of user capture normal use server the application server of operate as normal;
The user's that management server D transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch
The application server of operate as normal returns the network traffics of returning to the user via switch;
Management server C is in idle condition;
Four, malfunction 3
If initial condition is: application server 1,2 breaks down; Management server A breaks down.The communication system course of work of the embodiment of the invention is as follows:
Management server B, C, D all monitor application server 1,2 and break down;
Continue to be responsible for the overall work scheduling according to pre-defined rule management server B;
Take over the work that management server A comes the handling failure application server according to pre-defined rule management server D, though there is not operating load at present;
Switch is forwarded to management server D with the network traffics of user capture fault application server 1;
Management server B broadcasts corresponding ethernet address modification information on network, switch is forwarded to idle management server C with the network traffics that are about to user capture fault application server 2;
Switch directly is transmitted to user capture to the source network traffics of normal use server the application server of operate as normal;
The user's that management server D transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch;
The user's that management server C transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch;
The application server of operate as normal returns the network traffics of returning to the user via switch;
The management server of non-fault all is in operating duty, but the load of management server B is very light.
Five, malfunction 4:
If initial condition is: application server 1,2,3 break down; Management server A breaks down.The communication system course of work of the embodiment of the invention is as follows:
Management server B, C, D all monitor application server 1,2, and 3 break down;
Continue to be responsible for the overall work scheduling according to pre-defined rule management server B;
Take over the work that management server A comes the handling failure application server according to pre-defined rule management server D, though there is not operating load at present;
Switch is forwarded to management server D with the network traffics of user capture fault application server 1;
Switch is forwarded to management server C with the network traffics of user capture fault application server 2;
Management server B broadcasts corresponding ethernet address modification information on network, switch is transmitted on the very low management server B of relative load with the network traffics that are about to user capture fault application server 3;
Switch directly is transmitted to user capture to the flow of normal use server the application server of operate as normal;
The user's that management server D transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch;
The user's that management server C transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch;
The other work of management server B is user's that switch is transferred network access traffic is transmitted to other operate as normal once more via switch application server;
The application server of operate as normal returns the network traffics of returning to the user via switch;
The management server of non-fault all is in operating duty;
Six, malfunction 5
If initial condition is: application server 1,2,3,4 break down; Management server A breaks down.The communication system course of work of the embodiment of the invention is as follows:
Management server B, C, D all monitor application server 1,2, and 3,4 break down;
Continue to be responsible for the overall work scheduling according to pre-defined rule management server B;
According to the pre-defined rule switch network traffics of user capture fault application server 1 are forwarded to management server D;
Switch is forwarded to management server C with the network traffics of user capture fault application server 2;
Switch is forwarded to management server B with the network traffics of user capture fault application server 3;
Management server B broadcasts corresponding ethernet address modification information on network, switch is transmitted to management server A with the network traffics that are about to user capture fault application server 4; But because management server A is in malfunction, and it handles this part work of application server fault and is taken over by management server D already, so finally switch is transmitted to management server D with the network traffics of user capture fault application server 4;
Switch directly is transmitted to user capture to the flow of normal use server the application server of operate as normal;
The user's that management server D transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch
The user's that management server C transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch
The other work of management server B is user's that switch is transferred network access traffic is transmitted to other operate as normal once more via switch application server
The application server of operate as normal returns the network traffics of returning to the user via switch;
The management server of non-fault all is in operating duty, and wherein the load of management server D is higher relatively;
Seven, malfunction 6
If initial condition is: certain four application server 1,2,3,4 break down; Management server A, B breaks down.The communication system course of work of the embodiment of the invention is as follows:
Management server C, D all monitor application server 1,2, and 3,4 break down;
Replace management server B catcher overall work scheduling according to pre-defined rule management server C;
Work according to the handling failure application server on the pre-defined rule management server D catcher management server B comprises work unborn and that may be assigned with from now on;
According to the pre-defined rule switch network traffics of user capture fault application server 1,4 are forwarded to management server D;
Switch is forwarded to management server C with the network traffics of user capture fault application server 2;
Management server D broadcasts corresponding ethernet address modification information on network, switch is changed into and is transmitted to management server D by original management server B that is transmitted to the network traffics that are about to user capture fault application server 3;
Switch directly is transmitted to user capture to the source network traffics of normal use server the application server of operate as normal;
The user's that management server D transfers switch network access traffic is transmitted to the application server of other operate as normal once more via switch;
The other work of management server C is user's that switch is transferred network access traffic is transmitted to other operate as normal once more via switch application server;
The application server of operate as normal returns the network traffics of returning to the user via switch;
The management server of non-fault all is in operating duty, and the load of management server D is higher relatively;
It more than is the description for example of the scheduling of each malfunction.It is pointed out that the situation that fault takes place occurs at random, might not be according to occurring in sequence of " malfunction 1,2,3,4,5,6 " above.For the fault order of occurrence of other situation, whole management system is still according to default rule-based scheduling work.For the equipment after the fault recovery, management system by health examination mechanism and working resource election mechanism confirm its can operate as normal after, the operating load " return " that can will originally belong to this equipment, finally after all faulty equipments all recover operate as normal, make total system also revert to the initialization state.
Embodiment two
As shown in Figure 8, present embodiment provides a kind of communication means, comprises step:
Step 801, switch receive user's request, and under the control of management server user's request are forwarded to destination address corresponding application server or management server.
The state of step 802, management server monitoring application server, and according to the data forwarding of the state of application server and pre-defined rule control switch, so that when the application server fault, make user's request that switch will mail to this fault application server be transmitted to management server.
User's request that step 803, application server desampler and/or management server are transmitted, and handle this request is issued switch with result, and by switch the result of application server is transmitted to the user.
Because the effect of the active/standby election negotiation module of management server self, no matter be health examination, forwarded or ARP proxy, these work all can be along with dynamic migration on many management servers is consulted in the election to working resource, break down until last management server, any work can not received influence.And when management server was recovered by malfunction, active/standby election negotiation module can be redistributed working resource automatically, to reach the load balancing between the management server.
Communication system of the present invention has realized load balancing, server failure shielding, faults itself shielding, user transparent mode access, five functions of fault recovery self-healing of server; Because in communication system of the present invention, in normal course of communications, management server is communication process between participating user terminal and the application server not, only when the application server fault, just participate in the management work of communication system, this working method is called " out of band supervision, fault get involved " mode of operation, therefore, the load of communication system of the present invention self, have nothing to do with number of users, visit capacity, irrelevant with the quantity of application server, only relevant with the failure rate and the failure recovery time of server; So thoroughly solve the lifting of switching performance and the contradiction between the growing network traffics, opened up a novel application model.
Because communication system of the present invention has adopted the working method of " out of band supervision, fault get involved ", management server need not process user and the normal use server between network communication data, the network communication data that application server during for fault, management server also only need process user to send over.So the hardware performance requirement to management server self is very low, management server can use and the hardware of application server with class, and the hardware cost of himself is well below the hardware cost of the switching system of " work in the band " form.
Because communication system of the present invention adopts the working method of " out of band supervision, fault get involved ", the network communication data between user and the normal use server only needs by switch processes once.The Internet resources of 50% switch have been saved with respect to the inline patterns of four layers of exchange scheme " commercial "; The Internet resources of 25% switch have been saved with respect to the single armed patterns of four layers of exchange scheme " commercial ".
Because communication system of the present invention adopts the working method of " out of band supervision, fault get involved ", so it is the performance of management server self need not increase with the increase of user or application server, only relevant with the fault amount and the failure recovery time of application server to the performance requirement of management server.This makes Exchange Service of the present invention peel away from application service on performance.So in theory, communication system of the present invention can be carried the network traffics of any size and do not increased the hardware cost input of four layers of switching part.
Though described the present invention by embodiment, those of ordinary skills know, without departing from the spirit and substance in the present invention, just can make the present invention that many distortion and variation are arranged, and scope of the present invention is limited to the appended claims.
Claims (10)
1. a communication system is characterized in that, comprises switch, application server and management server,
Described application server is used for user's request that desampler and/or management server are transmitted, and handles this request, and result is issued switch;
Described switch is used for the result of application server is transmitted to the user, and under the control of management server user's request is forwarded to destination address corresponding application server or management server;
Described management server, be used to monitor the state of application server, and,, make user's request that switch will mail to this fault application server be transmitted to management server so that when the application server fault according to the data forwarding that the state and the pre-defined rule of application server are controlled switch.
2. system according to claim 1 is characterized in that, described management server is 2 to 254.
3. system according to claim 2 is characterized in that, described management server is 4.
4. according to claim 2 or 3 described systems, it is characterized in that, described management server comprises active/standby relation negotiation election module, it is used for the priority to the working resource of other this management server of management server multicast, and obtain the priority of the working resource of other management server, determine the operating state of this management server working resource according to the priority of each management server.
5. system according to claim 4 is characterized in that, described active/standby election negotiation module is operated in the kernel state of operating system.
6. according to claim 1 to 3 each described system wherein, it is characterized in that described management service implement body comprises:
The Frame forwarding module is used for user's request that desampler is transmitted, and according to transmitting described user is asked to carry out Frame forwarding operation, so that should be transmitted to normal application server by user's request that the fault application server is born.
7. according to claim 1 to 3 each described system wherein, it is characterized in that described management service implement body comprises:
The address analysis protocol proxy module is used to receive the ARP inquiry of the network equipment, and this ARP inquiry is used to obtain the MAC Address corresponding with SVCIP, and this MAC Address is returned this network equipment.
8. system according to claim 7 is characterized in that the described network equipment comprises switch.
9. according to claim 1 to 3 each described system wherein, it is characterized in that described management service implement body comprises:
Health examination module is used to check the state of application server, finding the application server of fault application server and operate as normal, the IP of fault application server is replaced with the VSIP of management server.
10. a communication means is characterized in that, comprises step:
Switch receives user's request, and under the control of management server user's request is forwarded to destination address corresponding application server or management server;
The state of management server monitoring application server, and according to the data forwarding of the state of application server and pre-defined rule control switch, so that when the application server fault, make user's request that switch will mail to this fault application server be transmitted to management server;
User's request that application server desampler and/or management server are transmitted, and handle this request is issued switch with result, and by switch the result of application server is transmitted to the user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010100025994A CN102130776B (en) | 2010-01-19 | 2010-01-19 | Communication method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010100025994A CN102130776B (en) | 2010-01-19 | 2010-01-19 | Communication method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102130776A true CN102130776A (en) | 2011-07-20 |
| CN102130776B CN102130776B (en) | 2013-05-01 |
Family
ID=44268676
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010100025994A Active CN102130776B (en) | 2010-01-19 | 2010-01-19 | Communication method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102130776B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103108019A (en) * | 2011-10-17 | 2013-05-15 | 株式会社日立制作所 | Service providing system |
| CN103581107A (en) * | 2012-07-19 | 2014-02-12 | 国基电子（上海）有限公司 | Server and working mode setting method thereof |
| CN103731294A (en) * | 2013-03-28 | 2014-04-16 | 深圳市金溢科技有限公司 | Method, device and system for two-machine hot-standby communication |
| CN104811483A (en) * | 2015-04-02 | 2015-07-29 | 广东睿江科技有限公司 | Cloud system allocation method and cloud system |
| CN104935479A (en) * | 2015-06-19 | 2015-09-23 | 郑州悉知信息技术有限公司 | Method and system for monitoring service module |
| CN105450775A (en) * | 2015-12-28 | 2016-03-30 | 曙光信息产业(北京)有限公司 | Load balancing method and load balancing device for DNS |
| CN105897827A (en) * | 2015-11-27 | 2016-08-24 | 乐视云计算有限公司 | Server node, local area network server cluster and realizing method thereof |
| CN107888405A (en) * | 2016-09-29 | 2018-04-06 | 富士通株式会社 | Management equipment and information processing system |
| CN108600381A (en) * | 2018-04-28 | 2018-09-28 | 广州联欣信息科技有限公司 | A kind of load-balanced server, user terminal and data transmission system |
| CN109565460A (en) * | 2017-03-29 | 2019-04-02 | 松下知识产权经营株式会社 | Communication device and communication system |
| CN110505176A (en) * | 2018-05-16 | 2019-11-26 | 中兴通讯股份有限公司 | Determination, sending method and device, the route system of message priority |
| CN111813625A (en) * | 2020-06-30 | 2020-10-23 | 中国工商银行股份有限公司 | Health check method and device for distributed server cluster |
| CN111903097A (en) * | 2018-03-30 | 2020-11-06 | 三菱电机株式会社 | Transmission device, transmission method, and transmission program |
| CN113220421A (en) * | 2021-05-31 | 2021-08-06 | 深圳市恒扬数据股份有限公司 | Management method, management server and management system of server cluster |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101320339A (en) * | 2007-06-06 | 2008-12-10 | 日立通讯技术株式会社 | Information-processing equipment and system therefor |
| CN101431648A (en) * | 2008-12-11 | 2009-05-13 | 北京东方广视科技有限责任公司 | Cable television VOD backup method, system and equipment |
-
2010
- 2010-01-19 CN CN2010100025994A patent/CN102130776B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101320339A (en) * | 2007-06-06 | 2008-12-10 | 日立通讯技术株式会社 | Information-processing equipment and system therefor |
| CN101431648A (en) * | 2008-12-11 | 2009-05-13 | 北京东方广视科技有限责任公司 | Cable television VOD backup method, system and equipment |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103108019A (en) * | 2011-10-17 | 2013-05-15 | 株式会社日立制作所 | Service providing system |
| CN103581107B (en) * | 2012-07-19 | 2017-05-10 | 国基电子（上海）有限公司 | Server and working mode setting method thereof |
| US9584365B2 (en) | 2012-07-19 | 2017-02-28 | Ambit Microsystems (Shanghai) Ltd. | Server and method for setting a mode of the server |
| TWI492577B (en) * | 2012-07-19 | 2015-07-11 | Hon Hai Prec Ind Co Ltd | Server and service mode setting method |
| CN103581107A (en) * | 2012-07-19 | 2014-02-12 | 国基电子（上海）有限公司 | Server and working mode setting method thereof |
| CN103731294A (en) * | 2013-03-28 | 2014-04-16 | 深圳市金溢科技有限公司 | Method, device and system for two-machine hot-standby communication |
| CN104811483B (en) * | 2015-04-02 | 2018-07-06 | 广东睿江云计算股份有限公司 | Cloud system dispositions method and cloud system |
| CN104811483A (en) * | 2015-04-02 | 2015-07-29 | 广东睿江科技有限公司 | Cloud system allocation method and cloud system |
| CN104935479A (en) * | 2015-06-19 | 2015-09-23 | 郑州悉知信息技术有限公司 | Method and system for monitoring service module |
| CN105897827A (en) * | 2015-11-27 | 2016-08-24 | 乐视云计算有限公司 | Server node, local area network server cluster and realizing method thereof |
| CN105450775A (en) * | 2015-12-28 | 2016-03-30 | 曙光信息产业(北京)有限公司 | Load balancing method and load balancing device for DNS |
| CN107888405A (en) * | 2016-09-29 | 2018-04-06 | 富士通株式会社 | Management equipment and information processing system |
| CN107888405B (en) * | 2016-09-29 | 2021-04-13 | 富士通株式会社 | Management apparatus and information processing system |
| CN109565460A (en) * | 2017-03-29 | 2019-04-02 | 松下知识产权经营株式会社 | Communication device and communication system |
| CN111903097A (en) * | 2018-03-30 | 2020-11-06 | 三菱电机株式会社 | Transmission device, transmission method, and transmission program |
| CN108600381A (en) * | 2018-04-28 | 2018-09-28 | 广州联欣信息科技有限公司 | A kind of load-balanced server, user terminal and data transmission system |
| CN110505176A (en) * | 2018-05-16 | 2019-11-26 | 中兴通讯股份有限公司 | Determination, sending method and device, the route system of message priority |
| US11303567B2 (en) | 2018-05-16 | 2022-04-12 | Xi'an Zhongxing New Software Co., Ltd. | Method and device for determining and sending priority of packet, and routing system |
| CN110505176B (en) * | 2018-05-16 | 2023-02-21 | 中兴通讯股份有限公司 | Method and device for determining and sending message priority, and routing system |
| CN110505176B9 (en) * | 2018-05-16 | 2023-04-11 | 中兴通讯股份有限公司 | Method and device for determining and sending message priority, and routing system |
| CN111813625A (en) * | 2020-06-30 | 2020-10-23 | 中国工商银行股份有限公司 | Health check method and device for distributed server cluster |
| CN111813625B (en) * | 2020-06-30 | 2024-03-08 | 中国工商银行股份有限公司 | Health checking method and device for distributed server cluster |
| CN113220421A (en) * | 2021-05-31 | 2021-08-06 | 深圳市恒扬数据股份有限公司 | Management method, management server and management system of server cluster |
| CN113220421B (en) * | 2021-05-31 | 2023-01-31 | 深圳市恒扬数据股份有限公司 | Management method, management server and management system of server cluster |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102130776B (en) | 2013-05-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102130776B (en) | Communication method and system | |
| EP3681110B1 (en) | A region interconnect control using vrf tables across heterogeneous networks | |
| US10142160B1 (en) | System and methods for managing network hardware address requests with a controller | |
| CN100334866C (en) | Method for realizing dynamic gateway load sharing and backup | |
| CN101217448B (en) | Method and system to realize gateway dynamic load sharing | |
| CN111638957B (en) | Method for realizing cluster sharing type public cloud load balance | |
| CN104104570A (en) | Aggregation processing method in IRF (Intelligent Resilient Framework) system and device | |
| CN101702689B (en) | Transmission control method with balanced multicast service data load and access device thereof | |
| WO2001039438A9 (en) | Network architecture and call processing system | |
| WO2012013050A1 (en) | Method and server used for information synchronization between master equipment and standby equipment | |
| CN102201999A (en) | Multicast service load sharing method and system | |
| CN113259235B (en) | IPv 6-based dual-active route redundancy method and system | |
| CN111083061B (en) | Method, equipment and system for determining DF (distribution function) of multicast flow | |
| CN102801632A (en) | Unified electoral method of PIM-SM (Protocol Independent Multicast-Sparse Mode) designated router and IGMP (Internet Group Management Protocol) querier | |
| CN101127755A (en) | Real time video image transmission method and system | |
| CN102571527A (en) | Backup and load balancing system based on virtual router redundancy | |
| CN100488201C (en) | Link backup method based on route | |
| CN106375211B (en) | Exchange data processing method and interchanger in VRRP load-balancing scenario | |
| WO2023092778A1 (en) | Method for realizing uninterrupted multicast traffic during migration of virtual machine | |
| CN103200117A (en) | Method and device of load balancing | |
| CN100359875C (en) | Method for realizing backup and load shared equally based on proxy of address resolution protocol | |
| CN103401954A (en) | Implementation method of virtual DHCP (dynamic host configuration protocol) | |
| CN100411387C (en) | Dualhoming network holding method based on RPR network | |
| CN114268581B (en) | Method for realizing high availability and load sharing of network equipment | |
| CN101164299B (en) | IP network resource configuring method and IP network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230414 Address after: Room 501-502, 5/F, Sina Headquarters Scientific Research Building, Block N-1 and N-2, Zhongguancun Software Park, Dongbei Wangxi Road, Haidian District, Beijing, 100193 Patentee after: Sina Technology (China) Co.,Ltd. Address before: 100080, International Building, No. 58 West Fourth Ring Road, Haidian District, Beijing, 1510 floor Patentee before: Sina.com Technology (China) Co.,Ltd. |
|
| TR01 | Transfer of patent right |