CN1882166A - Method for selecting service node according to load status - Google Patents
Method for selecting service node according to load status Download PDFInfo
- Publication number
- CN1882166A CN1882166A CN200510076802.1A CN200510076802A CN1882166A CN 1882166 A CN1882166 A CN 1882166A CN 200510076802 A CN200510076802 A CN 200510076802A CN 1882166 A CN1882166 A CN 1882166A
- Authority
- CN
- China
- Prior art keywords
- node
- load
- lcdn
- load level
- selection probability
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/08—Load balancing or load distribution
- H04W28/088—Load balancing or load distribution among core entities
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention relates to a method for selecting the service node according to the load condition, which uniformly divides the load of each node into at least one load levels, while the present load condition of each node is relative to one load level, wherein the LCDN obtains the load level of each node, based on which, to confirm the select probability of each node; then based on the probability, selecting the service node for the user terminal; therefore, the node with low load is relative to the higher select probability, to make the load of them uniform; in addition, the invention uses the load level as the index for balancing the node load, to make the balance process more easily. And the invention detail describes the transmission process of load information, to make the LCDN obtain the present load condition of node on time.
Description
Technical field
The present invention relates to load balancing techniques, be meant a kind of method of selecting service node according to load state especially.
Background technology
In the communications field, can have when a plurality of, need select the service node of a node therein as user terminal for user terminal provides the node of the same race of service, as shown in Figure 1, Node B
1, Node B
2Node B
nBe node of the same race, promptly can realize the node of identical function, n Node B formed the node resource pond, and user terminal is set up professional the connection with Node B, to obtain respective service.Network side can further comprise node A, and like this, user terminal can be set up professional the connection with Node B by node A, to obtain respective service.Any one Node B in the node resource pond all can be a certain user terminal service is provided, but the business that user terminal carries out at every turn can only provide respective service by a Node B.
Usually, when the user terminal requests service access, can in the node resource pond service node of Node B of user terminal selecting by node A as user terminal; If signaling interface is arranged between the Node B, can in the node resource pond service node of Node B of user terminal selecting also by a certain Node B as user terminal.
Mainly contain following several for the method for user terminal selecting Node B in the node resource pond: one, selecting at random, promptly is Node B of user terminal selecting randomly in the node resource pond; Two, the static selection, promptly the corresponding relation between each Node B in configure user terminal and the node resource pond is the user terminal selecting Node B according to configured corresponding relation; Three, repeating query is selected, and is each the Node B layout sequence number in the node resource pond, when being the user terminal selecting service node, successively according to Node B of selective sequential of sequence number at every turn; Four, be the user terminal selecting service node according to user ID (User-ID), be each the Node B layout sequence number in the node resource pond, be the individual Node B of user terminal selecting (User-ID Mode n).
Typical case's application in the above node resource pond is the core network (CN in Wideband Code Division Multiple Access (WCDMA) (WCDMA) network, Core Network) node pool, foregoing node A can be considered Radio Access Network (RAN, Radio Access Network) node, and Node B can be considered the CN node.
According to third generation collaborative project (3GPP, 3rd Generation Partnership Project) standard, RAN, Radio Access Network) node can link to each other with a plurality of CN nodes, forms the corresponding relation of multi-to-multi.The RAN node can be radio network controller (RNC, Radio Network Controller) or base station controller (BSC, Basic Station Controller).The CN node of circuit domain is mobile switching centre (MSC, Mobile Switching Centre), and the CN node of packet domain is service universal grouping wireless business supporting node (SGSN, Serving GPRS Support Node).
The notion of pool area has been proposed in the 3GPP standard, pool area is meant the zone that a plurality of RAN nodes cover, when user terminal is roamed in pool area, for providing the CN node of service, this user terminal can not change, can reduce like this owing to the CN node changes the more execution number of times of new technological process of the position causes, the Signalling exchange between the minimizing CN node.
A pool area provides service concurrently by a plurality of CN nodes usually, also can provide service by a CN node certainly.When a pool area provided service concurrently by a plurality of CN nodes, these CN nodes had just been formed a CN node pool, i.e. MSC pond or SGSN pond.A CN node pool can be simultaneously for a plurality of pool areas provide service, and promptly a CN node pool can be corresponding with a plurality of pool areas.
For CN node of unique identification in a CN node pool, each CN node in the CN node pool is assigned one or more network resource identifier (NRI, Network ResourceIdentifier), and by Operation and Maintenance the NRI that distributes is configured in the CN node, in order to identify the different CN nodes in the CN node pool.When the CN node is that user terminal distributes Temporary Mobile Subscriber Identity (TMSI, Temporary Mobile Subscriber Identity) or packet TMSI (P-TMSI, Packet Temporary Mobile Subscriber Identity) time, in TMSI/P-TMSI, inserts the NRI of this CN node.Variable-length is the 23rd bit to the 14 bits that the NRI of 0 to 10 bit (bits) is positioned at TMSI/P-TMSI.TMSI is by the CN node of circuit domain, is that MSC is that user terminal distributes; P-TMSI is by the CN node of packet domain, is that SGSN is that user terminal distributes.The actual significant bit length of NRI depends on the network planning among the TMSI/P-TMSI, and can pass through operation maintenance center (OMC, Operation and Maintenance Centre) and be configured.
Dispose the corresponding relation between the NRI and each CN node in the CN node pool in each RAN node of a pool area.In the CN node pool, dispose the corresponding relation between NRI and other CN nodes in each CN node.
When user terminal enters pool area, for it provides the CN node of service, promptly serves the CN node carries this service CN node in the TMSI/P-TMSI that distributes for user terminal NRI.In this pool area, when the Non-Access Stratum (NAS, Non-Access Stratum) that user terminal is established to the CN node at every turn connected, the RAN node routed to CN node corresponding to NRI according to NRI with initial NAS message.Like this, can make user terminal in a pool area, use same CN node.This shows that the RAN node is according to user ID, is that the NRI that comprises among the TMSI/P-TMSI comes for user terminal selecting CN node.
In the networking, can promote the operation ability of network greatly, be embodied in: if certain node failure can use other nodes in the node resource pond to continue to provide service by the mode that adopts the node resource pond; If two certain nodes are about to overload, the certain customers' terminal in this node can be transferred to other nodes in the node resource pond; If three because attended operation (O﹠amp; M) need stop the service that certain node externally provides, can select other nodes in the node resource pond to continue externally to provide service; If increased new node in the four node resource ponds, can apace some user terminals be transferred to this newly-increased node.
By above description as seen, the node resource pond has the advantage of load balancing, and realizes that the key of load balancing is and can comes to be the user terminal selecting service node according to the load of each node in the node resource pond.As shown in fig. 1, node A and Node B can both be according to the loads of each Node B in the node resource pond, come to be the user terminal selecting serving node B, therefore, these nodes of carrying out selection function can be called load control decision node (LCDN, Load Controlling and Deciding Node).If the selection of Node B is to be carried out by node A, then LCDN is positioned at node A; If the selection of Node B is to be carried out by Node B, then LCDN is positioned at Node B.
Node in the node resource pond can provide load information to LCDN, when LCDN is the user terminal selecting service node in the node resource pond, with the service node of preferentially selecting the minimum node of present load as user terminal.Like this, when LCDN is the user terminal selecting service node in the node resource pond, only select the minimum node of present load simply, but, if a large number of users accessing terminal to network was arranged in the short time, then can be owing to the service node of only selecting the minimum node of present load as user terminal, and the load of respective nodes is sharply raise, even overload; And select the minimum node of load simply, the load of each node in the node resource pool is tended to balance.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of method, the load of each node in the node resource pool is tended to balance more according to load state selection service node; Further, also provide detailed description among the present invention, made LCDN can in time get access to the present load situation of each node in the node resource pond the standard of evaluation node load and the transport process of load information.
In order to achieve the above object, the invention provides a kind of method of selecting service node according to load state, the load of each node in the node resource pond is divided into more than one load level equably, the present load situation of each node is corresponding to a load level, and the method includes the steps of:
A, load control decision node LCDN obtain the load level of each node in the node resource pond;
B, LCDN determine the selection probability corresponding to each node according to the load level of each node;
C, LCDN are the user terminal selecting service node according to the selection probability of determining.
Further comprise steps A 0 before the described steps A: the node in the node resource pond sends to LCDN and carries other load information message of this node present load level.
Comprise further before the described steps A 0 that the node of steps A 01:LCDN in the node resource pond sends the load information request.
When LCDN carries out described steps A 01, further comprise: LCDN starting load request timer, and the load information message that node sends described in the wait receiving step A01, if before the load requests timer expiry, LCDN receives the load information message that node described in the steps A 01 sends, and then stops the load requests timer; If load requests timer expiry, LCDN are not received the load information message that node described in the steps A 01 sends, then LCDN sends the load information request to node described in the steps A 01 once more.
Carry the LCDN sign in the described load information request.
Further comprise after the described steps A 0: LCDN returns the load information response to node described in the steps A 0.
When node is carried out described steps A, further comprise: node starting load response timer, and wait for the load information response that reception LCDN returns, if before the load response timer expiry, node is received the load information response that LCDN returns, and then stops the load response timer; If load response timer expiry, node are not received the load information response that LCDN returns, then node sends to LCDN once more and carries other load information message of this node present load level.
Further carry in the described load information message: node identification, or be used for the new node indication of the node whether identification nodes increase newly for the node resource pond, or be used for the O﹠amp whether identification nodes needs to carry out Operation and Maintenance; M indication, or above combination arbitrarily.
Carry O﹠amp in the described load information message; During the M indication, further carry in the described load information message: the time interval.
Further comprise steps A 02 before the described steps A 0: the node detection in the node resource pond changes to the load level of this node.
When node is carried out described steps A 02, further comprise: the starting load timer, behind the load timer expiry, further comprise steps A 021 before the described steps A 0: node judges whether and need send load information message to LCDN, if, execution in step A0 then; Otherwise, do not send load information message to LCDN.
Described steps A 021 is: whether the load level when node is at first judged the present load rank with the starting load timer is identical, if the load level of present load rank during with the starting load timer is identical, execution in step A0 then, if the load level when present load rank and starting load timer is inequality, continue then to judge whether the preceding load level of present load rank and starting load timer is identical, if identical, then this node does not send load information message to LCDN; If inequality, execution in step A0 then; Or be: whether the load level before node is judged the present load rank and started timer is identical, if identical, then this node does not send load information message to LCDN; If inequality, execution in step A0 then.
Described step B may further comprise the steps:
B1, LCDN number each node sequencing in the node resource pond, and the load level of each node are added up according to the load level corresponding to node;
The accumulation result that B2, LCDN obtain according to step B1 and the load level of each node, and, determine selection probability corresponding to each node according to the low more principle of the high more selection probability of load.
Comprise m node in the node resource pond, the described load level of node is high more, the present load of this node is big more, described step B1 is: LCDN is according to the load level corresponding to node, the order that each node in the node resource pond is raise according to load level or reduce sorts and writes sequence number, and the load level of each node is added up; Described step B2 is: LCDN obtains the load level of each node divided by step B1 accumulation result, the result that to be divided by is as the middle probability of each node, the middle probability that with all node ID sums is two nodes of m+1 then exchanges, and the middle probability after LCDN will exchange is as the final selection probability of each node.
Overload threshold is set among the LCDN, and the LCDN load level is set to 0 more than or equal to the selection probability of the node of overload threshold, and a described m node does not comprise the node of load level more than or equal to overload threshold.
If more than one load level is a k load level, then 1 load level is represented the 1/k load of node full load, and the raise load of this node of expression of the described load level of node increases.
Comprise m node in the node resource pond, the described load level of node is high more, the present load of this node is more little, described step B1 is: LCDN is according to the load level corresponding to node, the order that each node in the node resource pond is raise according to load level or reduce sorts and writes sequence number, and the load level of each node is added up; Described step B2 is: LCDN obtains the load level of each node divided by step B1 accumulation result, the result that will be divided by is as the selection probability of each node.
Overload threshold is set among the LCDN, and the selection probability that the LCDN load level is not more than the node of overload threshold is set to 0, and a described m node does not comprise that load level is not more than the node of overload threshold.
If more than one load level is a k load level, then 1 load level is represented the 1/k load of node full load, and the raise load of this node of expression of the described load level of node reduces.
The node that LCDN is identical with load level is randomly ordered, or sorts according to node processing ability and capacity.
The LCDN directly needs selection probability that carries out the node of Operation and Maintenance is set to 0, and a described m node does not comprise the node that needs to carry out Operation and Maintenance in the node resource pond.
When increasing node newly in the resource node pond, this method further comprises: the selection probability that LCDN increases node newly is set to the MAXIMUM SELECTION probability in original node, the selection probability that maybe increases node newly is set to the value greater than MAXIMUM SELECTION probability in original node, and will remain according to described step B1 and step B2 and select probability assignments to give each original node, it is 1 poor with the selection probability that increases node newly that described residue is selected probability.
The described selection probability that increases node newly is set to greater than the value of the MAXIMUM SELECTION probability in original node: LCDN is provided with weight coefficient, if this weight coefficient is a decimal, the selection probability that then increases node newly be 1 with this decimal and with original node in the product of MAXIMUM SELECTION probability, if this weight coefficient is an integer, then 1 with this integer and divided by the result of this integer, with the product of MAXIMUM SELECTION probability in original node; Or dispose fixation probability among the LCDN, the selection probability that increases node newly be in original node MAXIMUM SELECTION probability and this fixation probability and.
Comprise m node in the node resource pond, further comprise before the described step C: LCDN is according to the selection probability of determining, the scope division with 0 to 1 is m interval, and is interval corresponding with node, the selection probability of each length of an interval degree corresponding node; Described step C is: LCDN generates decimal at random, and decimal is affiliated interval at random to determine this, selects and the service node of the corresponding node in this interval as user terminal.
Transition thresholds is set among the LCDN, this method further comprises: the difference of high capacity rank and minimum load rank is during more than or equal to the transition thresholds that is provided with, certain customers' terminal in LCDN the node that load is the highest is transferred to the lower node of load, carries out described step C.
According to the proposed method, the load of each node in the node resource pond is divided into more than one load level equably, the present load situation of each node is corresponding to a load level.LCDN determines the selection probability of each node in the node resource pond according to the load level of each node in the node resource pond that gets access to, and is the user terminal selecting service node according to this selection probability.Like this, LCDN is that the lower node of load is determined higher selection probability in the node resource pond, makes that the load of each node tends to balance more in the node resource pond.
In addition, adopt load level among the present invention, make more simple to the measurement of node load, effective as the index of weighing node load.Among the present invention also the transport process to load information detailed description is provided, make LCDN can in time get access to the present load situation of each node in the node resource pond, and make the transmission of load information more reasonable, effectively.When the load level of node changed, respective nodes provided the present load rank to LCDN, made LCDN can ground dynamically determine the selection probability of each node according to the load level of each node in the node resource pond.
Description of drawings
Fig. 1 shows node resource pond schematic diagram;
Fig. 2 shows load information transport process schematic diagram among the present invention;
Fig. 3 shows among the present invention and selects the service node flow chart according to load state.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, the present invention is described in further detail below in conjunction with accompanying drawing.
Among the present invention, the load of each node in the node resource pond is divided into more than one load level equably, the present load situation of each node is corresponding to a load level.LCDN determines the selection probability of each node in the node resource pond according to the load level of each node in the node resource pond that gets access to, and is the user terminal selecting service node according to this selection probability.
The factor of definite load comprises one or more in the following stated in existing: central processing unit (CPU, Central Processing Unit) occupancy, the user terminal quantity that comprises excited users terminal and online user's terminal, signaling and service traffics, taking of memory source, or the like, there are very big-difference in the method and the index of the node device assessment load of different manufacturers production, therefore, among the present invention with the unique index of load level as the load of overall merit node, the node device of different manufacturers production can adopt diverse ways computational load rank, what this load level was represented is the relative load of node, i.e. load for full load can be with the full load of the open ended maximum number of user amount of node as node.
Load with node among the present invention is divided into k load level equably, and 1 load level is represented the 1/k of node full load, and the present load situation of node is represented that by load level the present load situation of each node is corresponding to a load level.For example, the load of node is divided into 8 load level equably, 1 load level is represented 1/8 of node full load, its load level is 4 during the node half load, and the load level when can node unloaded is set to 0, and load level raises along with the rising of this node load, load level during the node full load is set to 8, load level when also can node unloaded is set to 8, and load level reduces along with the rising of this node load, and the load level during the node full load is set to 0.Above-described zero load is not have user terminal in this node of expression, and is meant that the load of this node is minimum; Fully loaded is not to represent that the load in this node has reached 100% yet, and is meant that the load of this node is the highest.
Under the node of a plurality of manufacturer production and the situation of depositing, in different nodes, can represent identical or close load for making the identical load rank, as 8 load level are set, the load level of node 1 and node 2 is 2, load level is 2 should represent present load near 1/4 of full load in these two nodes, can by attended operation the load level computational methods be revised according to the practical experience of each node operation, for example, revise the weighted value of each load-factor by attended operation.
Each node in the node resource pond can be actively or passively reports to LCDN with self load information, the specific implementation process as shown in Figure 2:
The node of step 201:LCDN in the node resource pond sends load information request (LoadingInformation REQ), and the request respective nodes provides present load information to it.LCDN is after node sends the load information request, but starting load request timer is waited for the present load information that respective nodes is returned then; If until the load requests timer expiry, LCDN does not receive the present load information that respective nodes is returned, and then LCDN can send the load information request to respective nodes once more.The concrete structure of load information request is referring to the description of table one.
| Parameter name | Optional/essential | Meaning of parameters |
| Global LCDN-ID | M | The global identity of LCDN, if the RAN node as LCDN, then this parameter is Global RAN-ID; If the CN node is as LCDN, then this parameter is Global CN-ID |
Table one load information appealing structure
Wherein: the mandatory parameter that carries in the request of M-load information
Below several situations, LCDN can send the load information request to the node in the node resource pond, is described in detail below.
One, be provided with the load requests cycle among the LCDN, each load requests cycle, the node of LCDN in the node resource pond sent the load information request when expiring.This load requests cycle can be at the whole node resource pool, also can be at the individual node in the node resource pond.If the load requests cycle is at the whole node resource pool, then when a load requests cycle expires, LCDN each node in the node resource pond sends the load information request, and LCDN can send the load information request by broadcast mode each node in the node resource pond.If the load requests cycle is at the individual node in the node resource pond, then when a load requests cycle at certain node expires, LCDN sends the load information request to respective nodes, at this moment, can be set the load requests cycle at each node in the node resource pond among the LCDN.
Two, certain node in the node resource pond does not provide its load information to LCDN in the setting-up time scope, and the difference in load requests cycle is that this time range can be bigger value in the time range of this setting and the situation one.
Three, LCDN just comes into operation, and thinks that it is example that pool area provides the CN node pool of service, if the RAN node as LCDN, then new RAN node is increased to after the pool area, each CN node that can be in pool area sends load information request; If the CN node is as LCDN, then new CN node is increased to after the CN node pool, can send the load information request by each the CN node in pool area.
Step 202: each node in the node resource pond sends load information message (LoadingInformation) to LCDN, carries load information and state information in this load information message, and load information wherein is the present load rank of node.Node is after LCDN sends load information message, but starting load response timer is waited for the load information response that LCDN returns then; If until the load response timer expiry, node is not all received the load information response that LCDN returns, and then node can send load information message to LCDN once more.The concrete structure of load information message is referring to the description of table two.
| Parameter name | Optional/essential | Meaning of parameters |
| Global ID | M | The global identity of node, if node is the CN node, then this parameter is Global CN-ID |
| Load level | M | The present load rank of node can be set to 0 for the load level of the node that newly comes into operation |
| The new node indication | M | Be used for identifying the node whether respective nodes increases newly for the node resource pond, 1 is newly-increased node, and 0 is original node |
| The O﹠M indication | O | Be used for identification nodes and whether will carry out attended operation, 1 for needing to be serviced operation, and 0 for not needing to be serviced operation or cancellation attended operation |
| The time interval | O | Be used to indicate from sending load information message to the time interval of carrying out the attended operation, unit can be hour, when O﹠M is designated as 1, this parameter of portability in the load information message |
Table two load information message structure
Wherein: the mandatory parameter that carries in the M-load information message
The optional parameters that carries in the O-load information message
Below several situations, the node in the node resource pond can send load information message to LCDN, is described in detail below.
One, after the node in the node resource pond is received the load information request that LCDN sends, responds this load information request and send load information message, carry the present load rank of respective nodes in this load information message to LCDN.
Two, increased new node in the node resource pond, this node will send load information message to LCDN after starting, owing to be the node that increases newly, the present load rank of this node that carries in this load information message is 0; If also carry the new node indication in this load information message, then will to identify respective nodes be the node that increases newly in the node resource pond for new node indication, for example, new node is designated as at 1 o'clock represents that respective nodes is the node that increases newly in the node resource pond, and the new node indication that the original node in the node resource pond can carry in the load information message that LCDN sends is set to 0.
Three, the node in the node resource pond periodically sends load information message to LCDN, carries the present load rank of respective nodes in this load information message.
When four, the load level of node changed in the node resource pond, respective nodes sent load information message to LCDN, carried the present load rank of respective nodes in this load information message.
Five, because O﹠amp; Reasons such as M need stop certain node externally provides service, and respective nodes sends load information message to LCDN, carries identification nodes in this load information message and will carry out the O﹠amp of attended operation; The M indication, for example, O﹠amp; M is designated as at 1 o'clock and represents respective nodes will carry out attended operation, also carry the present load rank of respective nodes in this load information message.
Six, need be to carrying out the node cancellation O﹠amp of Operation and Maintenance; The M plan, respective nodes sends load information message to LCDN, carries the O﹠amp that identification nodes will be cancelled attended operation in this load information message; The M indication, for example, O﹠amp; M is designated as at 0 o'clock and represents respective nodes will cancel attended operation, also carry the present load rank of respective nodes in this load information message.
After step 203:LCDN receives load information message, return load information response (Loading Information ACK) to respective nodes, the notice respective nodes has been received the load information and the state information of its transmission.The concrete structure of load information response is referring to the description of table three.
| Parameter name | Optional/essential | Meaning of parameters |
| Global LCDN-ID | M | The global identity of LCDN, if the RAN node as LCDN, then this parameter is Global RAN-ID; If the CN node is as LCDN, then this parameter is Global CN-ID |
Table three load information response structure
Wherein: the mandatory parameter that carries in the response of M-load information
If LCDN has started the load requests timer after sending the load information request to respective nodes, and LCDN received the load information message that respective nodes sends before the load requests timer expiry after, then LCDN stops the load requests timer.
If node has started the load response timer after sending load information message to LCDN, and the load information response that LCDN returns, then this node stop load response timer before the load response timer expiry, have been received.
After LCDN receives the load information message that the node in the node resource pond sends, can obtain in the following stated load information and the state information one or more by this load information message: the unique identification of node; The present load rank of node; Whether the sign respective nodes is the new node indication of the node that increases newly in the node resource pond; Whether identification nodes will carry out the O﹠amp of attended operation; The M indication, for example, need to be serviced operation by 1 identification nodes, do not need to be serviced operation or cancellation attended operation by 0 identification nodes, LCDN can no longer select to carry out the service node of the node of Operation and Maintenance as user terminal, and can be further be transferred to other nodes in the node resource pond with needing to be serviced user terminal in the node of operation; Be used to indicate from sending load information message, work as O﹠amp to the time interval of carrying out the attended operation; M is designated as at 1 o'clock, this parameter of portability in the load information message.
In a complete load information transport process, step 201 shown in Fig. 2, step 202 and step 203 are not all to carry out, but according to various situations described above, can omit corresponding steps, for example, execution in step 202～step 203, or execution in step 201～step 202 only.
The node that proposes among the present invention in the node resource pond can be when detecting the present load rank and change, carry other load information message of present load level to the LCDN transmission, for avoiding frequently sending load information message to LCDN owing to the fluctuation of load causes the node in the node resource pond, node can be when detecting the load level change, the starting load timer, behind this load timer expiry, whether the load level when node can at first be judged the present load rank with the starting load timer is identical, if the load level of present load rank during with the starting load timer is identical, determine that then load level changes, this node sends load information message to LCDN.If the load level when present load rank and starting load timer is inequality, before then continuing to judge present load rank and starting load timer, whether promptly detect the load level preceding load level that changes identical, if it is identical, then be defined as the fluctuation of load, this node does not send load information message to LCDN; If inequality, determine that then load level changes, this node sends load information message to LCDN.In addition, behind the load timer expiry, whether the load level before node also can directly be judged the present load rank and start timer is identical, if identical, then is defined as the fluctuation of load, and this node does not send load information message to LCDN; If inequality, determine that then load level changes, this node sends load information message to LCDN.
For example, the load level of certain node in the node resource pond is changed to 5 by 4, then this node is changed at 5 o'clock detecting load level, the starting load timer, behind the load timer expiry, the present load rank of this node is j, if j still is 4, load level when then determining the starting load timer is changed into the fluctuation of load, and this node does not send load information message to LCDN; If j is not 4, determine that then change has taken place load level really, this node sends load information message to LCDN, at this moment, even j is not 5, represents that also change has taken place load level.
LCDN determines the selection probability of each node in the node resource pond according to the present load rank of each node in the node resource pond that gets access to, and is the user terminal selecting service node according to this selection probability, the specific implementation process as shown in Figure 3:
Step 301:LCDN obtains the load level of each node in the node resource pond, detailed process can realize by load information transport process described above, each node can initiatively provide its present load rank to LCDN in the node resource pond, for example, when the present load rank of certain node changes, like this, can reduce in LCDN and the node resource pond mutual between each node; Also can provide its present load rank to LCDN according to the request of LCDN.
Step 302～step 303:LCDN then according to the load level of accumulation result and each node, determines the selection probability of each node with the summation that adds up of the load level of each node in the node resource pond.
Step 304:LCDN is the user terminal selecting service node according to the selection probability of determining.
In addition, if O﹠amp; M points out that node need carry out Operation and Maintenance, and then LCDN no longer selects the service node of this node as user terminal, and the selection probability that is about to this node is defined as 0, that this node sends until receiving, as to carry cancellation Operation and Maintenance O﹠amp; The new load information message of M indication.If it is the node that increases newly in the node resource pond that new node is pointed out respective nodes, then LCDN can select the service node of respective nodes as all user terminals in a period of time, and redefines the selection probability of each node in the node resource pond after a time; Or LCDN is provided with the selection probability of the node that increases newly than higher, and redefines the selection probability of original each node in the node resource pond.
Below by the mode of giving an example, to being that the process of user terminal selecting service node is further described according to the selection probability of determining among the present invention.
A kind of implementation of the present invention is: the load of each node in the node resource pond is divided into k load level equably, k is the integer greater than 1,1 load level is represented the 1/k load of node full load, the present load situation of node is represented by load level, the present load situation of each node is corresponding to a load level, and other load that raises the expression respective nodes of node load level increases.
Comprise m node in the node resource pond, m is the integer greater than 1, and LCDN is according to the present load rank of each node of getting access to, to this m node sequencing, and according to the layout sequence number that puts in order of each node.LCDN can sort according to the rising of load level to this m node, also can sort according to the reduction of load level.Be that example is described with the rising ordering according to load level only below, reduction ordering and its realization basically identical according to load level do not repeat them here.
LCDN can sort according to the rising of load level to the node of the m in the node resource pond, and for example, the sequence number of the node that load level is minimum is 1, and the sequence number of the node that load level is time low is 2, and the rest may be inferred, and the sequence number of the node that load level is the highest is m.If there is the load level of more than one node identical in the node resource pond, then LCDN can be any or randomly ordered to these nodes, that is to say, which node preceding be the less position of sequence number, which node after be that the bigger position of sequence number can; LCDN also can be according to factors such as the disposal ability of each identical node of load level, capacity to these node sequencings, that is to say, with bigger nodes such as disposal ability, capacity come forward, be the less position of sequence number, with less nodes such as disposal ability, capacity come by the back, be the bigger position of sequence number.
I in m the node after the process LCDN ordering (i=1,2 ..., m) load level of individual node is Load
i, like this, Load
1≤ Load
2≤ ... ≤ Load
(i-1)≤ Load
i≤ Load
(i+1)≤ Load
m, the concrete load L in each node is identical with the order of Load, also is L
1≤ L
2≤ ... ≤ L
(i-1)≤ L
i≤ L
(i+1)≤ L
mLCDN adds up the load level of m node, i.e. Sum=Load
1+ Load
2+ ... + Load
i+ ... + Load
m
LCDN divided by Sum, determines the middle probability P of each node with the load level of each node
i, i.e. P
i=(Load
i/ Sum), can be corresponding to the middle probability of each node in the node resource pond referring to table four.
| Node ID | 1 | 2 | … | i | … | m-1 | m |
| Middle probability | P 1 | P 2 | … | P i | … | P m-1 | P m |
The middle probability of each node in the table four node resource pond
After LCDN determines the middle probability of m node in the node resource pond, be that the middle probability of 1 node and the middle probability of the node that sequence number is m exchange with sequence number, be that the middle probability of 2 node and the middle probability of the node that sequence number is m-1 exchange with sequence number, the rest may be inferred, be that the middle probability of node of i and the middle probability of the node that sequence number is m+1-i exchange with sequence number, middle probability after LCDN will exchange is as the final selection probability of respective nodes, this shows, load level is minimum, be the selection probability maximum of the minimum node of concrete load 1, can be corresponding to the selection probability of each node in the node resource pond referring to table five.
| Node ID | 1 | 2 | … | i | … | m-1 | m |
| Select probability | P m | P m-1 | … | P m+1-i | … | P 2 | P 1 |
The selection probability of each node in the table five node resource pond
If m is an even number, then LCDN according to the middle probability of principle exchange that two node ID sums are m+1, just in time can finish the exchange in twos of middle probability with each node in the node resource pond; If m is an odd number, then except that sequence number is the node of (m+1)/2, the sequence number sum that LCDN all can search with all the other nodes is the node ID of m+1, it is probability in the middle of LCDN all can exchange the principle that all the other nodes are m+1 according to two node ID sums, but for node ID (m+1)/2, with its sum be the node ID of m+1 still for (m+1)/2, therefore, can in the middle of the exchange of probability can be considered the exchange of probability in the middle of himself.
For the node that in the node resource pond, newly comes into operation, in order to select the service node of this node more as user terminal, the selection probability of the node that can newly come into operation is set to the selection probability that node ID is 1 node, or to be provided with than node ID be the selection probability P of 1 node
mBig, for example, configuration weight coefficient K among the LCDN, for example weight coefficient K is a decimal, then the node definite selection probability P of LCDN for newly coming into operation
NewBe P
m* (1+K),
Again for exampleWeight coefficient K is an integer, then the node definite selection probability P of LCDN for newly coming into operation
NewBe P
m* (1+K)/K; Again for example, configuration fixation probability P among the LCDN
Fixed, the selection probability P that LCDN determines for the node that newly comes into operation
HewBe P
m+ P
FixedDetermined the selection probability P of the node that newly comes into operation
NewAfter, LCDN redefines the selection probability of original each node in the node resource pond, is about to residue and selects probability 1-P
NewDistribute to original each node in the node resource pond.If in the node resource pond, newly come into operation a plurality of nodes simultaneously, then can suitably determine to select probability, as long as guarantee that the selection probability sum of existing each node in the node resource pond is 1 according to above principle for the node that newly comes into operation.
Another kind of implementation of the present invention is: the load of each node in the node resource pond is divided into k load level equably, 1 load level is represented the 1/k of node full load, the present load situation of node is represented by load level, the present load situation of each node is corresponding to a load level, and other load that raises the expression respective nodes of node load level reduces.
Comprise m node in the node resource pond, m is the integer greater than 1, and LCDN is according to the present load rank of each node of getting access to, to this m node sequencing, and according to the layout sequence number that puts in order of each node.LCDN can sort according to the rising of load level to this m node, also can sort according to the reduction of load level.Be that example is described with the rising ordering according to load level only below, reduction ordering and its realization basically identical according to load level do not repeat them here.
LCDN can sort according to the rising of load level to the node of the m in the node resource pond, and for example, the sequence number of the node that load level is minimum is 1, and the sequence number of the node that load level is time low is 2, and the rest may be inferred, and the sequence number of the node that load level is the highest is m.If there is the load level of more than one node identical in the node resource pond, then LCDN can be any or randomly ordered to these nodes, that is to say, which node preceding be the less position of sequence number, which node after be that the bigger position of sequence number can; LCDN also can be according to factors such as the whole capability of each identical node of load level, capacity to these node sequencings, that is to say, with less nodes such as whole capability, capacity come forward, be the bigger position of sequence number, with less nodes such as whole capability, capacity come by the back, be the bigger position of sequence number.
I in m the node after the process LCDN ordering (i=1,2 ..., m) load level of individual node is Load
i, like this, Load
1〉=Load
2〉=... 〉=Load
(i-1)〉=Load
i〉=Load
(i+1)〉=Load
m, concrete load L in each node and the order of Load are on the contrary, are L
1≤ L
2≤ ... ≤ L
(i-1)≤ L
i≤ L
(i+1)≤ L
mLCDN adds up the load level of m node, i.e. Sum=Load
1+ Load
2+ ... + Load
i+ ... + Load
m
LCDN divided by Sum, determines the selection probability P of each node with the load level of each node
i, i.e. P
i=(Load
i/ Sum), this shows that load level is the highest, be the selection probability maximum of the minimum node of concrete load 1, can be corresponding to the selection probability of each node in the node resource pond referring to table six.
| Node ID | 1 | 2 | … | i | … | m-1 | m |
| Select probability | P 1 | P 2 | … | P i | … | P m-1 | P m |
The selection probability of each node in the table six node resource pond
For the node that in the node resource pond, newly comes into operation, in order to select the service node of this node more as user terminal, the selection probability of the node that can newly come into operation is set to the selection probability that node ID is the node of m, or to be provided with than node ID be the selection probability P of the node of m
mBig, for example, configuration weight coefficient K among the LCDN, if weight coefficient K is a decimal, then LCDN is the definite selection probability P of node that newly comes into operation
NewBe P
m* (1+K), if weight coefficient K is an integer, then LCDN is the definite selection probability P of node that newly comes into operation
NewBe P
m* (1+K)/K; Again for example, configuration fixation probability P among the LCDN
Fixed, the selection probability P that LCDN determines for the node that newly comes into operation
NewBe P
m+ P
FixedDetermined the selection probability P of the node that newly comes into operation
NewAfter, LCDN redefines the selection probability of original each node in the node resource pond, is about to residue and selects probability 1-P
NewDistribute to original each node in the node resource pond.If in the node resource pond, newly come into operation a plurality of nodes simultaneously, then can suitably determine to select probability, as long as guarantee that the selection probability sum of existing each node in the node resource pond is 1 according to above principle for the node that newly comes into operation.
Do not have tangible execution sequence between the step that each node in the node resource pond is sorted according to load level and the load level of each node is added up described above, can carry out simultaneously; Also can carry out earlier node is sorted, carry out again load level is added up; Also can carry out earlier load level is added up, carry out again node is sorted.
In addition, overload threshold can further be set in LCDN, LCDN directly load level is set to 0 more than or equal to the selection probability of the node of overload threshold, it is the no longer selected service node of respective nodes as user terminal, like this, above-described m node do not comprise that those load level are more than or equal to the node of overload threshold in the node resource pond
If have node to need to be serviced operation in the node resource pond, then the LCDN selection probability that will need to be serviced the node of operation directly is set to 0, it no longer is the user terminal selecting respective nodes, like this, above-described m node do not comprise that those need carry out the node of Operation and Maintenance in the node resource pond.
If existing load level is more than or equal to the node of overload threshold in the node resource pond, the node that the operation of needing to be serviced is arranged again, and LCDN all is set to 0 with the selection probability of this node under situation in two, and then the above m node does not comprise that those load level in the node resource pond are more than or equal to the node of overload threshold and need carry out the node of Operation and Maintenance.
LCDN is after each node in the node resource pond has been determined the selection probability, to select probability according to this is the user terminal selecting service node, the specific implementation process can be: LCDN is after each node in the node resource pond has been determined the selection probability, scope division with 0 to 1 is m interval, i interval and i node are to corresponding, each length of an interval degree is the selection probability of respective nodes, each node in the node resource pond is corresponding with one of them interval, corresponding to table five, can be with m interval division (0, P
m], (P
m, P
m+ P
M-1], (P
m+ P
M-1, P
m+ P
M-1+ P
M-2] ... (P
m+ P
M-1+ ... + P
2, 1), and sequence number be 1 node with interval (0, P
m] corresponding, sequence number is 2 node and interval (P
m, P
m+ P
M-1] corresponding, the rest may be inferred, and sequence number is the node and interval (P of m
m+ P
M-1+ ... + P
2, 1) and corresponding, corresponding to table six, can be with m interval division (0, P
1), [P
1, P
1+ P
2), [P
1+ P
2, P
1+ P
2+ P
3) ... [P
1+ P
2+ ... + P
M-1, 1), and sequence number be 1 node with interval (0, P
1) corresponding, sequence number is 2 node and interval [P
1, P
1+ P
2) corresponding, the rest may be inferred, and sequence number is the node and interval [P of m
1+ P
2+ ... + P
M-1, 1) and corresponding; When being the user terminal selecting service node, LCDN at first generates a decimal at random, and decimal is affiliated interval at random to determine this, selects then and the service node of the corresponding node in this interval as user terminal.More than each interval all can be between closed area, if the decimal at random that LCDN generates is the critical point in two intervals, then LCDN can select the service node of one of them node as user terminal at random, also can select the service node of that less relatively node of load as user terminal.
As seen from the above description, LCDN is that the lower node of load is determined higher selection probability in the node resource pond, the selected selection probability as the subscriber terminal service node of each node in the node resource pond is corresponding with its corresponding siding-to-siding block length, siding-to-siding block length is big more, then the probability that drops in this interval range of the decimal at random that generates of LCDN is just big more, and the selected probability as the subscriber terminal service node of respective nodes is also just big more; But because LCDN can generate decimal at random, this at random decimal to drop on which interval all possible, therefore, make LCDN can in a period of time, not concentrate certain node of selection, also just avoided the load of certain node sharply to rise.
For example, 5 nodes are arranged in the node resource pond, other load that raises the expression respective nodes of node load level increases, LCDN according to process noted earlier be the selection probability determined of 5 nodes as shown in Table 7.
| Node ID | 1 | 2 | 3 | 4 |
| Select probability | 0.4 | 0.3 | 0.2 | 0.1 |
The selection probability of 5 nodes in the table seven node resource pond
LCDN is 4 intervals according to the selection probability of determining with 0 to 1 scope division, (0,0.4], (0.4,0.7], (0.7,0.9], (0.9,1), and node 1 is corresponding to interval (0,0.4], node 2 corresponding to interval (0.4,0.7], node 3 is corresponding to interval (0.7,0.9], node 4 is corresponding to interval (0.9,1).When LCDN is the user terminal selecting service node, generate decimal v at random, if decimal v is 0.3 at random, then LCDN determine at random decimal 0.3 drop on the interval (0,0.4] in, therefore, select the service node of node 1 as user terminal.If the v of decimal at random that LCDN generates is 0.7, then LCDN determine at random decimal 0.7 drop on the interval (0.4,0.7] in, therefore, select the service node of node 2 as user terminal; If above interval is between closed area, then LCDN can select the service node of node 2 as user terminal, also can select the service node of node 3 as user terminal, because the load of node 2 will be lower than the load of node 3, therefore, LCDN can select the service node of the lower node 2 of load as user terminal in node 2 and node 3.
In addition, the division in each interval also can can't help 0 as original position, is example to select probability shown in the table seven still, LCDN can be such 4 intervals with 0 to 1 scope division according to the selection probability of determining, [0,0.1], [0.1,0.5], [0.5,0.7], [0.7,1], and node 1 is corresponding to interval [0.1,0.5], and node 2 is corresponding to interval [0.7,1], node 3 is corresponding to interval [0.5,0.7], node 4 corresponding to interval [0,0.1).
LCDN is that the situation of user terminal selecting service node includes but not limited to: user terminal access node resource pool; User terminal requests node resource pond provides business service; LCDN is according to the load state of each node in the node resource pond, and the user terminal in certain or some nodes is transferred to other nodes in the node resource pond.
For the load that makes each node in the node resource pool reaches balanced, user terminal in the node that can load is higher is transferred to the lower node of load, but, if the transfer of user terminal is too frequent, can make the workload of LCDN excessive, and differ in load that transferring load is not again very necessary between the less node, therefore, transition thresholds can be set in LCDN, in the difference of the load level of the minimum node of the load level of the highest node of load and load during more than or equal to the transition thresholds that is provided with, the difference of promptly high capacity rank and minimum load rank is during more than or equal to the transition thresholds that is provided with, LCDN begins that user terminal is transferred to the lower node of load in the node that load is the highest, is the user terminal selecting service node according to the selection probability of determining.
Various implementation procedures described in the present invention can be applied in the various node resources pond, for example the CN node pool.The above user terminal not only comprises subscriber terminal equipment, also can comprise any network node that service can be provided by the node in the node resource pond.
In a word, the above is preferred embodiment of the present invention only, is not to be used to limit protection scope of the present invention.
Claims (25)
1, a kind of method of selecting service node according to load state, it is characterized in that, the load of each node in the node resource pond is divided into more than one load level equably, and the present load situation of each node is corresponding to a load level, and the method includes the steps of:
A, load control decision node LCDN obtain the load level of each node in the node resource pond;
B, LCDN determine the selection probability corresponding to each node according to the load level of each node;
C, LCDN are the user terminal selecting service node according to the selection probability of determining.
2, method according to claim 1 is characterized in that, further comprises before the described steps A:
Node in A0, the node resource pond sends to LCDN and carries other load information message of this node present load level.
3, method according to claim 2 is characterized in that, further comprises before the described steps A 0:
A01, the LCDN node in the node resource pond sends the load information request.
4, method according to claim 3 is characterized in that,
When LCDN carries out described steps A 01, further comprise: LCDN starting load request timer, and wait for the load information message that node sends described in the receiving step A01,
If before the load requests timer expiry, LCDN receives the load information message that node described in the steps A 01 sends, and then stops the load requests timer;
If load requests timer expiry, LCDN are not received the load information message that node described in the steps A 01 sends, then LCDN sends the load information request to node described in the steps A 01 once more.
5, according to claim 3 or 4 described methods, it is characterized in that, carry the LCDN sign in the described load information request.
6, according to claim 2 or 3 described methods, it is characterized in that, further comprise after the described steps A 0: LCDN returns the load information response to node described in the steps A 0.
7, method according to claim 6 is characterized in that,
When node is carried out described steps A, further comprise: node starting load response timer, and wait for and receive the load information response that LCDN returns,
If before the load response timer expiry, node is received the load information response that LCDN returns, and then stops the load response timer;
If load response timer expiry, node are not received the load information response that LCDN returns, then node sends to LCDN once more and carries other load information message of this node present load level.
8, method according to claim 2, it is characterized in that, further carry in the described load information message: node identification, or be used for the new node indication of the node whether identification nodes increase newly for the node resource pond, or be used for the O﹠amp whether identification nodes needs to carry out Operation and Maintenance; M indication, or above combination arbitrarily.
9, method according to claim 8 is characterized in that, carries O﹠amp in the described load information message; During the M indication, further carry in the described load information message: the time interval.
10, method according to claim 2 is characterized in that, further comprises before the described steps A 0:
Node detection in A02, the node resource pond changes to the load level of this node.
11, method according to claim 10 is characterized in that,
When node is carried out described steps A 02, further comprise: the starting load timer,
Behind the load timer expiry, further comprise before the described steps A 0:
A021, node judge whether and need send load information message to LCDN, if, execution in step A0 then; Otherwise, do not send load information message to LCDN.
12, method according to claim 11 is characterized in that, described steps A 021 is:
Whether the load level when node is at first judged the present load rank with the starting load timer is identical, if the load level of present load rank during with the starting load timer is identical, execution in step A0 then, if the load level when present load rank and starting load timer is inequality, continue then to judge whether the preceding load level of present load rank and starting load timer is identical, if identical, then this node does not send load information message to LCDN; If inequality, execution in step A0 then;
Or be: whether the load level before node is judged the present load rank and started timer is identical, if identical, then this node does not send load information message to LCDN; If inequality, execution in step A0 then.
13, method according to claim 1 is characterized in that, described step B may further comprise the steps:
B1, LCDN number each node sequencing in the node resource pond, and the load level of each node are added up according to the load level corresponding to node;
The accumulation result that B2, LCDN obtain according to step B1 and the load level of each node, and, determine selection probability corresponding to each node according to the low more principle of the high more selection probability of load.
14, method according to claim 13 is characterized in that,
Comprise m node in the node resource pond, the described load level of node is high more, and the present load of this node is big more,
Described step B1 is: LCDN is according to corresponding to the load level of node, each node in the node resource pond raise according to load level or the order that reduces sorts and writes sequence number, and the load level of each node is added up;
Described step B2 is: LCDN obtains the load level of each node divided by step B1 accumulation result, the result that to be divided by is as the middle probability of each node, the middle probability that with all node ID sums is two nodes of m+1 then exchanges, and the middle probability after LCDN will exchange is as the final selection probability of each node.
15, method according to claim 14, it is characterized in that, overload threshold is set among the LCDN, and the LCDN load level is set to 0 more than or equal to the selection probability of the node of overload threshold, and a described m node does not comprise the node of load level more than or equal to overload threshold.
16, method according to claim 14 is characterized in that, if more than one load level is a k load level, then 1 load level is represented the 1/k load of node full load, and the raise load of this node of expression of the described load level of node increases.
17, method according to claim 13 is characterized in that,
Comprise m node in the node resource pond, the described load level of node is high more, and the present load of this node is more little,
Described step B1 is: LCDN is according to corresponding to the load level of node, each node in the node resource pond raise according to load level or the order that reduces sorts and writes sequence number, and the load level of each node is added up;
Described step B2 is: LCDN obtains the load level of each node divided by step B1 accumulation result, the result that will be divided by is as the selection probability of each node.
18, method according to claim 17 is characterized in that, overload threshold is set among the LCDN, and the selection probability that the LCDN load level is not more than the node of overload threshold is set to 0, and a described m node does not comprise that load level is not more than the node of overload threshold.
19, method according to claim 17 is characterized in that, if more than one load level is a k load level, then 1 load level is represented the 1/k load of node full load, and the raise load of this node of expression of the described load level of node reduces.
According to claim 14 or 17 described methods, it is characterized in that 20, the node that LCDN is identical with load level is randomly ordered, or sort according to node processing ability and capacity.
According to claim 14 or 17 described methods, it is characterized in that 21, the LCDN directly needs selection probability that carries out the node of Operation and Maintenance is set to 0, a described m node does not comprise the node that needs to carry out Operation and Maintenance in the node resource pond.
22, according to claim 14 or 17 arbitrary described methods, it is characterized in that, when increasing node newly in the resource node pond, this method further comprises: the selection probability that LCDN increases node newly is set to the MAXIMUM SELECTION probability in original node, the selection probability that maybe increases node newly is set to the value greater than MAXIMUM SELECTION probability in original node, and will remain according to described step B1 and step B2 and select probability assignments to give each original node, it is 1 poor with the selection probability that increases node newly that described residue is selected probability.
23, method according to claim 22, it is characterized in that, the described selection probability that increases node newly is set to greater than the value of the MAXIMUM SELECTION probability in original node: LCDN is provided with weight coefficient, if this weight coefficient is a decimal, the selection probability that then increases node newly be 1 with this decimal and with original node in the product of MAXIMUM SELECTION probability, if this weight coefficient is an integer, then 1 with this integer and divided by the result of this integer, with the product of MAXIMUM SELECTION probability in original node; Or dispose fixation probability among the LCDN, the selection probability that increases node newly be in original node MAXIMUM SELECTION probability and this fixation probability and.
24, according to claim 1,13,14 or 17 described methods, it is characterized in that,
Comprise m node in the node resource pond, further comprise before the described step C: LCDN is according to the selection probability of determining, the scope division with 0 to 1 is m interval, and is interval corresponding with node, the selection probability of each length of an interval degree corresponding node;
Described step C is: LCDN generates decimal at random, and decimal is affiliated interval at random to determine this, selects and the service node of the corresponding node in this interval as user terminal.
25, method according to claim 1, it is characterized in that, transition thresholds is set among the LCDN, this method further comprises: the difference of high capacity rank and minimum load rank is during more than or equal to the transition thresholds that is provided with, certain customers' terminal in LCDN the node that load is the highest is transferred to the lower node of load, carries out described step C.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100768021A CN100459795C (en) | 2005-06-15 | 2005-06-15 | Method for selecting service node according to load status |
| PCT/CN2006/001229 WO2006133630A1 (en) | 2005-06-15 | 2006-06-06 | A method for selecting a servicing node according to load status |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100768021A CN100459795C (en) | 2005-06-15 | 2005-06-15 | Method for selecting service node according to load status |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1882166A true CN1882166A (en) | 2006-12-20 |
| CN100459795C CN100459795C (en) | 2009-02-04 |
Family
ID=37520085
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005100768021A Expired - Fee Related CN100459795C (en) | 2005-06-15 | 2005-06-15 | Method for selecting service node according to load status |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN100459795C (en) |
| WO (1) | WO2006133630A1 (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008113300A1 (en) * | 2007-03-20 | 2008-09-25 | Huawei Technologies Co., Ltd. | Method, system and apparatus for selecting network devices |
| CN101272614B (en) * | 2007-03-20 | 2010-12-08 | 华为技术有限公司 | Method, system and device for selecting network equipment |
| CN102170422A (en) * | 2010-02-25 | 2011-08-31 | 腾讯科技（深圳）有限公司 | Method, device and system for inquiring node |
| CN101730201B (en) * | 2008-10-10 | 2012-06-06 | 大唐移动通信设备有限公司 | Destination node selection method and device based on loading information |
| CN102487509A (en) * | 2010-12-02 | 2012-06-06 | 中兴通讯股份有限公司 | Method and system for dynamically allocating resources of packet control unit |
| CN101925123B (en) * | 2009-06-12 | 2012-12-12 | 中国移动通信集团设计院有限公司 | Method, system and related functional entities for realizing load balancing in pool area |
| CN102970243A (en) * | 2012-11-19 | 2013-03-13 | 北京思特奇信息技术股份有限公司 | Load-coefficient-based message forwarding link equalizing method and device |
| CN103188164A (en) * | 2011-12-31 | 2013-07-03 | 中国移动通信集团云南有限公司 | Method, device and system for shunting peer node load based on upgraded client node |
| CN105207818A (en) * | 2015-09-23 | 2015-12-30 | 杭州数梦工场科技有限公司 | Node control method and device |
| CN107977270A (en) * | 2017-11-22 | 2018-05-01 | 用友金融信息技术股份有限公司 | Peers distribution method, peers distribution system and computer installation |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8457133B2 (en) | 2007-10-29 | 2013-06-04 | Telefonaktiebolaget L M Ericsson (Publ) | Load balancing in core network comprising a pool of core network elements |
| CN101895932B (en) * | 2009-05-19 | 2013-01-23 | 中国移动通信集团公司 | Load transferring method, system and device |
| CN103152774B (en) * | 2013-03-26 | 2015-12-09 | 广东电网公司电力调度控制中心 | Support the LTE Interface Expanding method and system of distributed freedom load balancing |
| CN110933136A (en) * | 2019-10-31 | 2020-03-27 | 北京浪潮数据技术有限公司 | Service node selection method, device, equipment and readable storage medium |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6002676A (en) * | 1997-06-30 | 1999-12-14 | Motorola, Inc. | Method and apparatus for selecting a resource in a communication system with resources having unbalanced load capacity |
| KR100300348B1 (en) * | 1998-09-23 | 2001-09-06 | 윤종용 | Method for call setup in cellular system |
| CN1112813C (en) * | 1999-03-19 | 2003-06-25 | 上海贝尔电话设备制造有限公司 | Intelligent routing method in communication network |
| US6522881B1 (en) * | 2000-03-08 | 2003-02-18 | Lucent Technologies Inc. | Method and apparatus for selecting an access point in a wireless network |
| SE0004591D0 (en) * | 2000-12-12 | 2000-12-12 | Ericsson Telefon Ab L M | Method and means in a telecommunication system |
| EP1368982A1 (en) * | 2001-03-16 | 2003-12-10 | BRITISH TELECOMMUNICATIONS public limited company | Selection of an appropriate network resource node in a cellular telecommunication system |
| US7142860B2 (en) * | 2001-03-30 | 2006-11-28 | Telefonaktiebolaget Lm Ericsson (Publ) | Network/cell/interface selection in mixed networks |
| KR100421892B1 (en) * | 2001-12-29 | 2004-03-11 | 엘지전자 주식회사 | Mobile Communacation System and Method of Selecting a Server at the Mobile Communication System |
| CN100377554C (en) * | 2004-05-25 | 2008-03-26 | 华中科技大学 | Load balancing method for cluster server |
-
2005
- 2005-06-15 CN CNB2005100768021A patent/CN100459795C/en not_active Expired - Fee Related
-
2006
- 2006-06-06 WO PCT/CN2006/001229 patent/WO2006133630A1/en active Application Filing
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101272614B (en) * | 2007-03-20 | 2010-12-08 | 华为技术有限公司 | Method, system and device for selecting network equipment |
| WO2008113300A1 (en) * | 2007-03-20 | 2008-09-25 | Huawei Technologies Co., Ltd. | Method, system and apparatus for selecting network devices |
| US8509163B2 (en) | 2007-03-20 | 2013-08-13 | Huawei Technologies Co., Ltd. | Method and system for selecting network equipment |
| CN101730201B (en) * | 2008-10-10 | 2012-06-06 | 大唐移动通信设备有限公司 | Destination node selection method and device based on loading information |
| CN101925123B (en) * | 2009-06-12 | 2012-12-12 | 中国移动通信集团设计院有限公司 | Method, system and related functional entities for realizing load balancing in pool area |
| CN102170422A (en) * | 2010-02-25 | 2011-08-31 | 腾讯科技（深圳）有限公司 | Method, device and system for inquiring node |
| CN102487509A (en) * | 2010-12-02 | 2012-06-06 | 中兴通讯股份有限公司 | Method and system for dynamically allocating resources of packet control unit |
| CN103188164B (en) * | 2011-12-31 | 2016-01-27 | 中国移动通信集团云南有限公司 | Based on method, the Apparatus and system of upgrade client node shunting peer node load |
| CN103188164A (en) * | 2011-12-31 | 2013-07-03 | 中国移动通信集团云南有限公司 | Method, device and system for shunting peer node load based on upgraded client node |
| CN102970243A (en) * | 2012-11-19 | 2013-03-13 | 北京思特奇信息技术股份有限公司 | Load-coefficient-based message forwarding link equalizing method and device |
| CN102970243B (en) * | 2012-11-19 | 2015-09-09 | 北京思特奇信息技术股份有限公司 | A kind of message conversion link equalization methods based on load factor and device |
| CN105207818A (en) * | 2015-09-23 | 2015-12-30 | 杭州数梦工场科技有限公司 | Node control method and device |
| CN107977270A (en) * | 2017-11-22 | 2018-05-01 | 用友金融信息技术股份有限公司 | Peers distribution method, peers distribution system and computer installation |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2006133630A1 (en) | 2006-12-21 |
| CN100459795C (en) | 2009-02-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1882166A (en) | Method for selecting service node according to load status | |
| CN101068370A (en) | User equipment calling method and system | |
| CN1720759A (en) | Apparatus and method for controlling access to network in wireless communication system | |
| CN1756429A (en) | Cell information change informing method in multimedia broadcast/multicast service | |
| CN1706123A (en) | Method and system for providing status information for broadcast/multicast service in a mobile communication system | |
| CN1717880A (en) | RRC connection request method and apparatus for mobile communications | |
| CN1943146A (en) | Transmission of control information in wireless communication system | |
| CN1720677A (en) | RRC group reject method and apparatus for mobile communications | |
| CN1717069A (en) | The method of effective communicating control information in the multimedia broadcast/multicast service | |
| CN1303799C (en) | Method for controlling multimedia broadcast/multicast service conversation | |
| CN1518255A (en) | Method for moving UE in RRC connection mole | |
| CN1930904A (en) | Method and apparatus for sending signaling for multicast and broadcast services | |
| CN1411671A (en) | Method and arrangement for optimizing re-establishment of connection in cellular radio system supporting real time and non-real time communications | |
| CN1524393A (en) | Method to control the packet switching traffic in digital cellular communication systems | |
| CN101911779A (en) | Methods and apparatus for cell reselection | |
| CN1934805A (en) | A method of transmitting information related to a multicast service in a mobile telecommunications network | |
| CN1875559A (en) | Calling method for a multimedis broadcast/multicast service in a mobile communication system | |
| CN1794827A (en) | Receiving method of multimedia brondiasting/group broadcast service control information | |
| CN1839655A (en) | Mbms method and apparatus for umts network | |
| CN1934894A (en) | Mobile communication system | |
| CN1500360A (en) | Method for assessing characteristics of network for mobile telecom appts. | |
| CN1711787A (en) | Collective notification of node reset to subset of connections in radio access network | |
| CN1503474A (en) | Wireless insertion network system, method, and control equipment | |
| CN1859766A (en) | Method for processing service bearing while communication system sponsor service request | |
| CN101039523A (en) | Joint user transferring method and communication system and service GPRS support node |
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 | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090204 Termination date: 20130615 |