阅读说明：本技术 多节点装置及其备援通信方法 (Multinode device and its redundant communication means ) 是由 谢孟桓 张毓融 刘得恺 于 2018-12-13 设计创作，主要内容包括：本发明公开一种多节点装置及其备援通信方法,用于确保多节点服务器的网络通信可靠度。多节点服务器包括第一节点、第一内建式交换器以及第二内建式交换器。第一节点具发送数据分组的端口。第一内建式交换器具有耦合到第一节点的端口的下游端口、互连端口及耦合到网络的上行链路端口。上行链路端口将数据分组从第一节点路由到网络。第二内建式交换器具有下游端口、耦合到第一内建式交换器的互连端口的互连端口、以及耦合到网络的上行链路端口。(The present invention discloses a kind of multinode device and its redundant communication means, for ensuring the network communication reliability of multi node server.Multi node server includes first node, the first in-building type exchanger and the second in-building type exchanger.First node tool sends the port of data grouping.First in-building type exchanger has downstream port, interconnection port and the uplink port for being coupled to network for the port for being coupled to first node.Data grouping is routed to network from first node by uplink port.Second in-building type exchanger have downstream port, the interconnection port for being coupled to the first in-building type exchanger interconnection port and be coupled to the uplink port of network.)

1. a kind of multinode device, includes:

First node, having can operate to send or receive the first port of data grouping；

First in-building type exchanger has the first downstream port, the first interconnection of the first port for being coupled to the first node Port and at least one first uplink port for being coupled to network, the first in-building type exchanger are configured to detect these Network communication failure on first uplink port；And

Second in-building type exchanger, the first interconnection end that there is the second downstream port, be coupled to the first in-building type exchanger Second interconnection port of mouth and the second uplink port for being coupled to network；

Wherein when the first in-building type exchanger detects these the first uplink port simultaneous faults, the first node is come from Data grouping is via first downstream port of the first in-building type exchanger, first interconnection port and this is second built-in Second uplink port of formula exchanger is routed to network.

2. multinode device as described in claim 1, further includes:

Second node, have be coupled to the second in-building type exchanger second downstream port second port, and this second Port can be operated to send or receive data grouping.

3. multinode device as described in claim 1, wherein the first in-building type exchanger includes packets forwarding table, the grouping Forwarding table includes:

The first information, for data grouping to be routed to network via these first uplink ports from the first node；With And

Second information, for data grouping to be routed to the first node via first downstream port, wherein this in first When the formula exchanger of building detects these the first uplink port simultaneous faults, remove the first information in the packets forwarding table with And second information.

4. multinode device as described in claim 1, wherein the first in-building type exchanger includes filter, this crosses filter fly Sequence only allow to indicate the message of the operating state of these the first uplink ports of the first in-building type exchanger this first It is transmitted on interconnection port, and in the first in-building type exchanger to detect these first uplink ports same for the filter When failure when, allow the first in-building type exchanger to receive data grouping on first interconnection port.

5. providing the method that redundant communicates between a kind of multinode device and network, include:

Multinode device as described in Claims 1 to 4 item is any；

These the first uplink port simultaneous faults are detected in the first in-building type exchanger；And

It executes route processing and detects these the first uplink port simultaneous faults, the road to respond the first in-building type exchanger Include by processing:

The data grouping of the first port from the first node is first built-in via this from the first in-building type exchanger It is second built-in that first interconnection port of formula exchanger and second interconnection port of the second in-building type exchanger route to this Formula exchanger；And

The data grouping is routed to network via second uplink port of the second in-building type exchanger.

6. method as claimed in claim 5, wherein the route processing further includes being somebody's turn to do the second in-building type exchanger The received data grouping of second uplink port is via first interconnection port, second interconnection port and this is first built-in First downstream port of formula exchanger is routed to the first node.

7. method as claimed in claim 5, wherein the route processing is further included incites somebody to action via the first in-building type exchanger Indicate that the broadcast message of these the first uplink port simultaneous faults is sent to the second in-building type exchanger, this is second built-in For the configuration of formula exchanger to run filter, permission receives number on second interconnection port that first interconnection port links According to wherein the broadcast message includes the MAC Address of the first node.

8. the method for claim 7, further including and being updated in the second in-building type exchanger based on the broadcast message Packets forwarding table in information, wherein the information includes the routing data to the first node.

9. the method for claim 7, further including and sending the broadcast to network via the second in-building type exchanger Message makes to be achieved from network via the data transmission of the second in-building type exchanger to the first node.

10. method as claimed in claim 5, if further include detect the first in-building type exchanger it is any these first When uplink port fault restoration, the data grouping road of the first uplink port from the first node to the reparation is restored By.

Technical field

This announcement relates generally to the network communication reliability ensured for server interior joint, is to close more specifically Redundant communication party is provided when interrupting between exchanger using interconnection port with the uplink between exchanger and network Formula.

Background technique

The appearance that cloud calculates application increases the demand of non-at-scene (off-site) or strange land installation, such as known data center (data center), to storing data and run by the computer installation user distally connected access application.It is common Data center have with power and communicate to connect function entity floor structure.Each rack can hold multiple network dresses It sets, such as server for calculating, storing or control and network and management exchanger.

Calculation server generally includes several processors or node.Each processor must be handled with other on server Device is communicated to efficiently accomplish calculating task.Therefore, between the kernel (cores) of calculating task often on the server at division Reason, and data can be between core and a larger network via the uplink port (uplink in in-building type exchanger Ports it) exchanges.These in-building type exchangers allow other devices in network to communicate with the node in server.Network with The communication between node in server is that the different uplinks of each node on server are connected to by in-building type exchanger Road port is completed.A problem brought by such configuration be when the uplink communication failure between node and network, Communication just can not reach, thus hinder server operation.

Each server node in microserver via each in-building type exchanger fixation uplink port with it is micro- Device communication outside server.It is by in-building type switch management by the uplink port that server node accesses.Therefore, When uplink port breaks down, associated server node cannot be with network communication.

In order to increase microserver meshed network communication robustness, current method be by starting in-building type exchanger, To provide alternative route in uplink port failure or interiorly or exteriorly network connecting fault.Fig. 1 be include micro services The block diagram of the prior art network system 10 of device 12.Server 12 includes multiple server nodes, such as node 20,22.Clothes Device 12 be engaged in also including multiple in-building type exchangers, such as in-building type exchanger 30,32.Each in-building type exchanger 30,32 includes Corresponding uplink port 42,44,46,48, these uplink ports with network 50 to be communicated.Each in-building type Exchanger 30,32 includes corresponding downstream port (downstream ports) 52,54,56,58, these downstream ports 52,54, 56,58 port 62,64,66,68 to be connected on node 20,22.

Herein in the prior art, node 20 includes being linked to the port 62 of the downstream port 52 of exchanger 30 and being linked to The port 64 of the downstream port 56 of exchanger 32.Node 22 include be linked to exchanger 30 downstream port 54 port 66 and It is linked to the port 68 of the downstream port 58 of exchanger 32.Node 20 can be via exchanger 30 and corresponding port 52 and network 50 communications.Similarly, node 22 can be communicated via exchanger 32 and corresponding port 58 with network 50.

In the case where uplink port 42,44 or uplink port 46,48 failure, associated server node 20 or 22 detect the loss of the input and output grouping that carry out automatic network 50.Associated server node 20,22 then by It is logical come the redundant for starting uplink port that grouping is sent and received from the substitution uplink port of another in-building type exchanger Letter mechanism.For example, server node 20 will test packet loss and lead to grouping if uplink port 42,44 failures Letter is switched to port 64.Therefore grouping is sent to downstream port 56 via link, and then arrive exchanger 32, and via uplink Road port 46 is sent to network 50.

Occur two in the current embodiment of in-building type exchanger node however, changing in uplink port failure Problem.First problem, which is attributed to the uplink redundant too long starting time, leads to significant packet loss.Due to server Node is not directly connected to corresponding uplink port, will expend the time to detect the failure of uplink port, therefore The problem of leading to potential a large amount of packet loss before the redundant communication mechanism for starting uplink port.Node port switching More additional times are also needed to send packets to other redundant exchangers.Therefore, in the preceding paragraphs, due to node 20 It is not directly connected to the uplink port 42,44 of failure, therefore in uplink port 42,44 failures to detecting grouping Many groupings are lost in the time of loss.In the case where packet communication is switched to connection exchanger 32 newly by redundant communication mechanism Also many groupings are lost between trip port 56 and corresponding new uplink port 46 in the spent time.

Second Problem is the deterioration that the computing capability of the node involved in uplink port redundant mechanism utilizes (degradation).The software-based routing program of current uplink road port redundant communication mechanism, this routing program includes must The software task of packet loss must frequently be detected.Such mechanism must be run on Servers-all node, so as to cause service The reduction that device node computing capability utilizes, because computing resource must be put into software task by node.

Therefore, industry has to for ensuring what network was communicated to node in the case where a network connection is damaged now The demand of one mechanism, to for ensuring the mechanism of minimal packet loss when the connection between node and network uplink is damaged Another demand and do not need using server computational power redundant mechanism another demand.

Summary of the invention

The example of one announcement is a kind of system that the reliable network for ensuring multinode device communicates.Multinode device Including first node, the first in-building type exchanger and the second in-building type exchanger.First node have can operate with send or Receive the port of data grouping.First in-building type exchanger has the downstream port for the port for being coupled to first node, interconnection end Mouth and the uplink port for being coupled to network.Data grouping is routed to network from first node by uplink port.Second In-building type exchanger has downstream port, is coupled to the interconnection port of the first in-building type exchanger and is coupled to the uplink of network Link port.One route processing is activated the net detected at line link port on it to respond the first in-building type exchanger Network communication failure, make the data grouping from first node via the downstream port of the first in-building type exchanger, interconnection port and The uplink port of second in-building type exchanger is routed to network.

The above content is not intended to indicate each aspect of each embodiment or this announcement.Above content is provided solely for herein Some examples in the novel aspect and feature of elaboration.When being considered in light of actual conditions in conjunction with attached drawing and appended claim, this announcement Features above and advantage and other features and advantage by from for realizing representative embodiment and mode of the invention with It is apparent in lower detailed description.

Detailed description of the invention

For above and other purpose, feature, advantage and the embodiment of this announcement can be clearer and more comprehensible, the explanation of attached drawing is such as Under:

Fig. 1 is in the event of uplink failure using the existing skill of the microserver of known redundant mechanism routing packets Art multi-node configuration schematic diagram.

Fig. 2 is the multi-node server system of the redundant mechanism for uplink failure of one embodiment of this announcement Block diagram.

Fig. 3 is the block diagram of the multi-node system under routine operation in Fig. 2.

Fig. 4 is the square of the multi-node system when the running of the redundant communication mechanism of one embodiment of this announcement, in Fig. 2 Figure.

Fig. 5 is the program for the redundant of the uplink failure between the node and network in the system in Fig. 4 Flow chart.

Fig. 6 is when restoring uplink port for providing the flow chart of the recovery program initially routed.

[symbol description]

10,100: network system

12,102: server

20,22,120,122: node

30,32,130,132: exchanger

42、44、46、48、52、54、56、58、62、64、66、68、142、144、146、

148,152,154,156,158,162,164,166,168,172,174,176,178: port

50,150: network

182: host

500,502,504,506,508,510,512,600,602,604,606,608,610: step

Specific embodiment

The present invention can be embodied with numerous different forms.Representative embodiment illustrates in the drawings, and will be herein Detailed description.This announcement is the embodiment or explanation of the principle of this announcement, and is not intended to for the extensive aspect of this announcement being limited to Embodiment described.In this regard, it for example making a summary, disclosed in summary of the invention and embodiment part but not in claim The element being expressly recited in book and limitation should not individually or jointly be incorporated to claim by implicit, reasoning or other modes In book.For the purpose of present embodiment, unless be not claimed specifically, otherwise singular includes plural form and anti- ；And word " comprising " means " including but not limited to ".In addition, for example, word is approximate, such as " about ", " almost ", " substantial ", " about " and fellow can be used to mean herein " being in ", " close " or " being nearly at " or " ... 3- In 5% " or " in acceptable fabrication tolerance " or its any logical combination.

This example is included between in-building type exchanger about a kind of server uplink port redundant communication mechanism Connection and its forward the packet table synchronization.When uplink failure occurs, the in-building type exchanger for detecting failure will divide Group is forwarded to another in-building type exchanger from server node.On the other hand, due to the packets forwarding between in-building type exchanger Table is to synchronize, and is received in the uplink of in-building type exchanger, and the server for being attached to other in-building type exchangers The grouping of node may be forwarded to the interconnection port of in-building type exchanger and then by being attached to other in-building type exchangers Server node receives.Therefore, the mechanism of this announcement provides the seamless communication for server node to network, because of service Device node is not related to the processing of redundant mechanism.Due to eliminating the time lag of port switching, the mechanism of this announcement can increase The network robustness of node server is added, and then prevents packet loss and eliminates to based on the server of redundant communication mechanism The demand of calculation ability.

The redundant mechanism of this announcement is to be based on that in-building type exchanger being connected to each other to allow in-building type via interconnection port There is the communication of substitution between exchanger and network.As shown in Fig. 2, Fig. 2 is the square of the network system 100 of one embodiment of this announcement Figure.Network system 100 includes the multinode device (such as server 102) with multiple nodes (such as node 120,122).Clothes Device 102 be engaged in also including multiple in-building type exchangers, such as in-building type exchanger 130,132.Each in-building type exchanger 130,132 Including corresponding uplink port 142,144,146,148, these uplink ports 142,144,146,148 allow built-in The transmission and reception of the data communication of formula exchanger 130,132 and network 150.Each in-building type exchanger 130,132 includes phase The downstream port 152,154,156,158 answered, these downstream ports 152,154,156,158 are connected respectively to node 120,122 Port 162,164,166,168.

In this embodiment, node 120 includes port 162 and the link for being linked to the downstream port 152 of exchanger 130 To the port 164 of the downstream port 156 of exchanger 132.Node 122 includes the end for being linked to the downstream port 154 of exchanger 130 Mouthfuls 166 and be linked to exchanger 132 downstream port 158 port 168.Node 120 can be via exchanger 130 and corresponding Port 152 or communicated via exchanger 132 and corresponding port 156 with network 150.Similarly, node 122 can be via friendship Parallel operation 132 and corresponding port 158 are communicated via exchanger 130 and corresponding port 154 with network 150.

Exchanger 130 includes two interconnection ports 172,174.Exchanger 132 also includes two interconnection ports 176,178. In this embodiment, the interconnection port 172 of exchanger 130 is linked to the interconnection port 176 of exchanger 132.Exchanger 130 it is mutual Even port 174 is linked to the interconnection port 178 of exchanger 132.Each exchanger 130,132 includes synchronous packets forwarding table (packet forwarding tables).Exchanger 130,132 is using interconnection port 172,174,176,178 with by uplink The operating state (health status) of road port 142,144,146,148 is periodically reported to other exchangers.Interconnect end Mouth 172,174,176,178 is also (also that is, same at uplink port for all uplink port failures when an exchanger When failure (concurrent failure)) in the case where execute packets forwarding table synchronization.It is grouped being received from interconnecting nodes When, which port exchanger (such as exchanger 130) determines to forward the packet to by with reference to its packets forwarding table.Grouping turns Deliver provide network address and by corresponding relationship between the received port for forwarding a packet to associated network address. In the case where the destination network address for receiving grouping is not present in packets forwarding table, grouping will be broadcast in addition to connecing Receive all of the port except the port of its grouping.

In this embodiment, when all uplink port (such as uplinks of an exchanger (such as exchanger 130) Port 142,144) failure when, exchanger is built-in to other by interconnection port (such as interconnection port 172,174,176,178) Formula exchanger (such as exchanger 132) notifies its uplink failure.Notice be according to the redundant program run by exchanger come It carries out.Detect the in-building type exchanger of failure also its packets forwarding table synchronous with other in-building type exchangers.Specifically, tool There is server section of the exchanger of uplink failure to other exchangers notice via its uplink port communication to network The network address of point.When receiving this notice, other exchangers modify its packets forwarding table, so that destination is above-mentioned network The grouping of address is forwarded to interconnection port.For example, node 120 via exchanger 130 uplink port 142 and net Network 150 communicates.If uplink port 142,144 equal failures, and exchanger 130 detects failure.Exchanger 130 will notify Exchanger 132 its uplink port 142,144 failures, and synchronize the packets forwarding table of its packets forwarding table Yu exchanger 132. After the packets forwarding table of the two is synchronous, the grouping for being sent to node 120 can be forwarded to interconnection port 176 by exchanger 132 Or 178.

The in-building type exchanger that network communication failure occurs for uplink port notifies thereon to other in-building type exchangers Line link failure, so that associated server node is via interconnection port and the uplink port and net of other exchangers Network 150 communicates.In addition, synchronous packets forwarding table, which reduces the grouping because of caused by uplink failure, spreads unchecked (packet Flooding exception).In the case where destination network address is not present in packets forwarding table, grouping will be broadcast to and remove Receive all of the port except the port of the grouping, including uplink port, downstream port and interconnection port.Packets forwarding The synchronization of table enable uplink port can the in-building type exchanger of normal operation perceive the section for attaching to other exchangers The network address of point.Therefore, destination can be that the grouping for the network address of node for being attached to other exchangers turns by exchanger It is dealt into its interconnection port, and is not necessary to broadcast this grouping.

By method as described above, each server node can be via being connected to corresponding in-building type exchanger Identical port sends and receives grouping, even if it is also such that wherein the uplink port of an exchanger, which breaks down,.Therefore, it takes Business device node does not need to be related in the processing of uplink port failure.In addition, in-building type exchanger 130,132 alternative services Device node 120,122 handles the problem of uplink failure.Therefore, the mechanism of this announcement more flexiblely responds uplink Failure, thus reduce the packet loss as caused by uplink failure.

Each in-building type exchanger of uplink port with normal operation periodically issues activity message (keep-alive message), and the state of its uplink port is reported via one in interconnection port to other In-building type exchanger.Packet filtering rule configures on interconnection port, only allows activity message between the port of normal operating It is sent and is received.When receiving activity message, the failure of uplink port will be solved.Due to only activity message period Between grouping be lost, therefore packet loss is positively retained at minimum value.

Fig. 3 to Fig. 4 depicts the network system 100 in Fig. 2 to ensure the redundant mechanism with the communication of network 150.Fig. 3 Element into Fig. 4 is used and the identical numeral mark of its counterpart in Fig. 2.Fig. 3 depicts the node in server 102 Normal flow between network.For example, the host 182 in network 150 is communicated with the node 120 in server 102. Fig. 3 is painted when all uplink ports 142,144,146,148 normal operation between host 182 and server node 120 Communication lines by.In the case, server node 120 selects the host 182 in in-building type exchanger 132 and network 150 logical Letter.Therefore, data flow to the downstream port 156 of exchanger 132 via the port 164 of node 120.Data are by exchanger 132 It is routed to network 150 via uplink port 148, and then arrives host 182.Packet filtering rule on in-building type exchanger is anti- Only data are communicated via interconnection port 172,174,176,178 and are transmitted.Similarly, in-building type exchanger 132 is via its uplink Road port 148 is forwarded towards host 182 and is grouped.

The case where uplink port 146,148 that Fig. 4 depicts in-building type exchanger 132 breaks down.In this feelings Under shape, in-building type exchanger 132 will notify its uplink port 146,148 failures to in-building type exchanger 130.Then setting The interconnection port 176,178 of in-building type exchanger 132 packet filtering rule and the mutual of in-building type exchanger 130 be set Even the packet filtering rule of port 172,174 can be closed.Grouping is allowed to via interconnection port 172,174,176,178 later Transmitting.In this example, the interconnection port 174 of exchanger 130 is arrived in grouping by interconnection port 178.As shown in figure 4, from clothes The grouping of business device node 120 is sent to the downstream port 156 of exchanger 132 via port 164.Grouping by exchanger 132 via Interconnection port 178 is routed to the interconnection port 174 of exchanger 130.Exchanger 130 will be grouped road via uplink port 144 By to network 150, and then arrive host 182.

In the case of any uplink port 146 or 148 of exchanger 132 restores function, exchanger 132 is to exchange Device 130 notifies its uplink port 146 or 148 to restore, point being then arranged in in-building type exchanger 130,132 the two Group filtering rule will reactivate.Therefore the data grouping transmitting between interconnection port 172,174,176,178 will again by It stops.Then, as shown in figure 3, in-building type exchanger 132 can be again through having restored the uplink port 146 or 148 of function In one to host 182 forward be grouped.

When all uplink port failures of an exchanger, the redundant communication mechanism of one embodiment of this announcement has as follows Operation.Fig. 5 is the flow chart of redundant communication mechanism.In step 500, when the uplink port of an exchanger breaks down, Initially built-in formula exchanger indicates the activity message of its uplink port failure to its adjacent in-building type exchanger publication.Then Execute the route processing of step 502.Step 502, initially built-in formula exchanger closes the interconnection end being configured at adjacent exchanger Packet filtering rule on the interconnection port of mouth link.Thus allow to receive and send out on the interconnection port between two exchangers Send all types of groupings.

Step 504, all packets forwarding table information relevant to its uplink port are removed in initial exchange device.It connects Step 506, initially built-in formula exchanger be grouped the packets forwarding table that forwarding table is synchronized to adjacent exchanger.It is specific next It says, in-building type exchanger issues broadcast message to adjacent exchanger.Each of broadcast message carrys out source MAC and is arranged to be attached To the MAC Address of the server node of in-building type exchanger.

Step 508, when receiving the activity message for indicating the uplink port failure of initially built-in formula exchanger, phase Adjacent in-building type exchanger closes the packet filtering rule being arranged on its interconnection port.Thus allow adjacent in-building type exchanger Interconnection port and its transmission for being linked to all types of groupings between the interconnection port for sending the initial exchange device of activity message And it receives.Then step 510, starting packets forwarding table synchronization mechanism are executed.In step 510, table synchronization mechanism is forwarded the packet According to the packets forwarding table for updating adjacent exchanger from the received broadcast message of initial exchange device.This to come to be connected to have The grouping of the server node of the initially built-in formula exchanger of failure uplink port, is able to be redirected into interconnection port, And network is sent to via the uplink port of adjacent in-building type exchanger.

Step 512, the received broadcast message of institute is also sent to network 150 via uplink port, so that in server 102 external network equipments (as shown in Figure 2) perceive server node will be via the uplink of adjacent in-building type exchanger Port is communicated with external network equipment.

In network system 100 as shown in Figure 2, during the operation for being transferred to another exchanger, when initial exchanger Any of failure uplink port again normal operation when, network system 100 is allowed to restore its routinely to operate.Figure Flow chart in 6 depicts recovery routine, and wherein one in initial exchange device discovery failure uplink port is recovered And it can normal operation.In this embodiment, such as step 600, initially built-in formula exchanger is sent out to adjacent in-building type exchanger Cloth indicates the activity message that its uplink port restores.Then step 602 is executed, the interconnection port with initial exchange device is removed Relevant all packets forwarding table information.

In step 604, the packet filtering rule of initially built-in formula exchanger is enabled, therefore initially built-in formula exchanger is mutual It is even only permitted to send and receive activity message between port and the interconnection port of adjacent exchanger.Then step 606, it is initial interior The formula exchanger of building then issues broadcast message, these broadcast messages have to be arranged via one or more uplink ports of recovery For the source MAC of the server node of attachment.This make the external network equipment (as shown in Figure 2) of server 102 be able to through It is communicated by one or more uplink ports newly restored with server node.

Step 608, when the activity news for receiving one or more uplink ports recovery for indicating initially built-in formula exchanger When breath, the packet filtering rule for being configured at the interconnection port of adjacent in-building type exchanger is enabled, so that only activity message is allowed to It sends and receives on interconnection port.Then step 610 is executed, all packets forwarding table letters relevant to interconnection port are removed Breath.Therefore communication of the node with network switches back into the uplink port that is restored by node via initial exchange device to network.

Flow chart of the Fig. 5 into Fig. 6 is used to indicate the readable instruction fetch of example machine of the network system 100 in Fig. 2.In In this embodiment, machine readable instruction fetch include by following devices execute algorithm: (a) processor, (b) controller, and/or (c) one or more other suitable treatment devices.It is embodied in the software that algorithm can store on tangible medium, this tangible medium is such as Flash memory, CD-ROM, floppy discs, hard drive, digital number video (multi-functional) disk (DVD) or other storages Device device.However, those skilled in the art will easily recognize, entire algorithm and/or its part are alternately by addition to processing Device other than device is executed and/or is embodied in firmware or specialized hardware (for example, it can be integrated by special applications in known ways Circuit (application specific integrated circuit, ASIC), Programmable Design logic device (programmable logic device, PLD), field-programmable design logic device (field programmable Logic device, FPLD), field-programmable design gate array (field programmable gate array, FPGA), Individual logical etc. is realized).For example, any or all component of interface can be by software, hardware and/or firmware Lai real It is existing.Equally, some or all of machine readable instruction fetch indicated by flow chart can also be realized by artificial.In addition, although by The algorithm of one state sample implementation of this announcement is described with reference to Fig. 5 flow chart depicted into Fig. 6, those skilled in the art will easily It recognizes, alternatively using numerous other methods that the machine readable instruction fetch of the present embodiment can be achieved.For example, changeable square Execute sequence, and/or it is changeable, eliminate or combination the present embodiment described in some step squares.

In conclusion this announcement is about a kind of multinode microserver device and its redundant mechanism, redundant mechanism can The widely application deployment in any system, the terminal installation in above system have to the dual-network for being connected to core network The network connection of access device.For example, wireless access point can play the role of in-building type exchanger explained hereinafter.If in uplink Link is the similar approach that another access point of normal operation has and is applicable at the interconnection of initial Access Point, then is connected to access The mobile device of access point need not solve the problems, such as uplink failure in access point.

If used in this application, it is related real that term " component ", " module ", " system " or fellow refer generally to computer Body, be perhaps the combination of software (for example, circuit), hardware and software, software or for there are one or more specific functions Operate the relevant entity of machine.For example, component can be but not limited to transport on processor (for example, digital signal processor) Capable processing, processor, object, executable thread, execution thread, program and/or computer.By the mode of explanation, controlling The application run on device and controller may each be component.One or more components can reside in processing and/or execution thread, And component local terminal and/or can be distributed between two or more computers on a computer.In addition, " device " can be with In lower form: the hardware of special designing；By software is executed on it, specially manufactured typical hardware, this software enable hardware and hold Row specific function；The software being stored on computer-readable medium；Or combinations thereof.

Technology used herein is not intended to limit the present invention merely for for the purpose of describing particular embodiments.As herein It uses, unless the context clearly indicates otherwise, otherwise singular " one " and "the" are intended to also include plural form.In addition, just For this, term " includes ", " having ", " " or its modification are used in embodiment and/or in being, these terms be intended to with The similar mode of term "comprising" is included.

Unless otherwise defined, otherwise term (including technology and scientific term) used used herein have with by this field The identical meaning of the technical staff person of being generally understood.In addition, term (such as term defined in common dictionary) should be explained To have and its consistent meaning of meaning in the context of the relevant technologies, and will be not interpreted as ideal or excessively formal Meaning, unless being clearly defined herein.

Although each embodiment of the invention has been described above, it should be understood that each embodiment of the invention is only with example Mode present, and with no restriction.Although the present invention illustrates and describes about one or more implementations, reading And when understanding this specification and attached drawing, equivalence changes and modification by generation or are known to those skilled in the art.In addition, although originally The special characteristic of invention may be disclosed about the only one embodiment in several implementations, such as any given Or specific application can for expectation and it is advantageous, this feature can be in conjunction with one or more other features of other implementations.Therefore, Range and range of the invention should not be limited by any embodiment in examples described above.But it is of the invention Range should be defined according to the claim of this application book and its equivalent.