CN102664784B - A kind of mapping method of virtual network of node link pressure weight adaptive equalization - Google Patents
A kind of mapping method of virtual network of node link pressure weight adaptive equalization Download PDFInfo
- Publication number
- CN102664784B CN102664784B CN201210116239.6A CN201210116239A CN102664784B CN 102664784 B CN102664784 B CN 102664784B CN 201210116239 A CN201210116239 A CN 201210116239A CN 102664784 B CN102664784 B CN 102664784B
- Authority
- CN
- China
- Prior art keywords
- node
- physical network
- link
- net
- empty
- 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.)
- Active
Links
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides the mapping method of virtual network of a kind of node link pressure weight adaptive equalization.The method can according to the state of present physical network node link stress, the optimization aim that empty net node is mapped adjusts in real time, the weight making this optimization aim interior joint pressure and link stress obtains adaptive equalization, and employ testing mechanism, adaptive equalization to ensure weight will not be dispersed, so that the result that empty net maps has the characteristic that node link pressure is comprehensively optimum.The method can strengthen the stability of virtual network operation, and network virtualization technology practical application in network operation is significant.
Description
Technical field
Network virtualization technology is one of important method promoting oriented Internet Architecture development, its essence is by abstract, distribution, isolation mech isolation test runs multiple virtual subnet independently on a public physical network, each virtual subnet can use separate protocol architecture, and according to the demand that user dynamically changes, whole nodes and link circuit resource can be carried out reasonable disposition, thus strengthening motility and the multiformity of network, realize the measurable and controllable property of network, the distribution of peak optimizating network resource and scheduling, improve safety and service quality, reduce operation maintenance cost, in the hope of essence solve the Internet existing ossify, with patch and be updated to main current situation.
Network virtualization technology may be used for the research of new network architecture and provides the basis sharing Physical Experiment network, bottom physical facility provider and network service operators can also be separated by it simultaneously, the network allowing multiple operator shares same public bottom physical network architecture (link, switching node etc.), each network has the Network resource share that neither can be adjusted flexibly again by other web influences wherein, heterogeneous networks operator can adopt different procotols, the end-to-end service of innovation is provided, therefore network virtualization also gets a good chance of becoming the main flow operation mode of a kind of future network.
Background technology
Virtual network mapping problems is then requisite link in network virtualization technology, its major function is that the virtual network requests (VirtualRequest) of user reasonably maps to bottom physical network facility (SubstrateNetwork) that operator provides, mapping process not only to realize separation between virtual network and be independent of each other, thus ensureing the service quality (QoS) of each virtual network user, also to try one's best simultaneously and reasonably distribute bottom physical network resource, improve resource utilization.
Virtual network mapping problems can be further subdivided into node and map and link maps two parts, and owing to node cpu capacity and link bandwidth being had all Multiple Constraints in mapping process, therefore virtual network mapping problems is a NP-hard problem.For ensureing that this way to solve the problem has engineering practice, domestic and international research worker mainly employs the heuritic approach of suboptimum, and proposes the mapping model based on time window.Time window model is by mapping problems online treatment void net request in units of time window, adds up current all of void net mapping request, and use mapping algorithm to map in each time window, to mapping successfully request, and corresponding renewal bottom physical network state;The lost request lost of mapping, puts into waiting list by request;Or after meeting certain condition, directly refuse this request.This process is as shown in Figure 1.
In empty net mapping problems, node maps and link maps carries out respectively generally according to precedence, and complexity so can be made to reduce, but owing to coupling not between two steps, also brings along the reduction of performance.The main stream approach tackling this problem is by adjusting and optimizing target in node mapping step, looks after the feature that following link maps, thus coupling with being effectively realized two mapping step.This research also will adopt this starting point, map optimization aim with the node link pressure realizing virtualization network this specific objective balanced by adjusting node flexibly.Load balancing in network virtualization mapping is primarily referred to as the maximum pressure as far as possible reducing void net mapping posterior nodal point and link, thus reducing the variance of network pressure, makes the pressure of network various piece be more nearly average pressure.In the research to this problem, first existing work consider the method for equalizing section point pressure and link stress respectively: in only considering the research that node pressure is balanced, owing to the node that prioritizing selection pressure is little frequently can lead to euclidean distance between node pair farther out, therefore a virtual link is accomplished by a plurality of physical link and realizes, therefore link stress increases relatively big, and mapping efficiency and success rate are also relatively low;In the research only considering link stress, also due to not accounting for node pressure, it is possible to cause that node pressure is excessive or it is low to be mapped to power.Therefore, in order to solve this contradiction, node link pressure two should be considered and map optimization aim, result of study based on this thinking includes: one is the formula that direct summation or the mode of product combine two optimization aim, so in optimization process, two factors are all taken into account, but both equilibrium relations are also indefinite, it is possible to cause that one of them optimization aim can occupy leading position all the time, and the effect of another optimization aim is difficult to manifest;Two is be dynamically selected use node or link stress as optimization aim according to current network pressure state, and the problem of do so is the state maintaining single optimization aim all the time, and therefore complex optimum effect is difficult to ensure.
In sum, realizing between node link pressure optimization aim in the research of good coupling, still have a lot of problem values to inquire into, especially need a kind of adjustable optimization aim equilibrium strategy of efficient stable, to realize the equilibrium of whole network node link stress.
Summary of the invention
The complex optimum situation that it is optimization aim with node and link stress in empty net mapping process that the present invention analyzes simultaneously, it has been found that there is trade-off relation between the two, be namely reduced to optimization aim with node pressure and will cause that link stress raises, vice versa.Therefore, in order to realize the equilibrium relation between two optimization aim, it is necessary to a kind of adjustable complex optimum object function of weight of design;Meanwhile, the undulatory property of multiformity and request arriving rate in order to adapt to request topology, arranging of this parameter needs dynamic scalable;Finally, in order to ensure the stability of system, it is necessary to certain measure ensures the convergence of this dynamic customized parameter.
The present invention is according to this starting point, devise the mapping method of virtual network that a kind of node link pressure adaptive is balanced, the method use the adjustable complex optimum target of balance parameters, it is possible not only to realize the complex optimum for node pressure, link stress, weight relationship between the two can also be regulated according to mapping result, two optimization aim are made fully to couple, thus realizing the aggregative equilibrium of the whole network pressure;It addition, we devise a judgment mechanism that adjustable parameter is restrained direction, to prevent adjustable parameter from dispersing, it is achieved Fast Convergent, thus ensureing the stability of system.
The definition that the present invention relates to:
1) node link pressure
Refer to the pressure of bottom physical network nodes i, the total CPU capacity of this node determine with residue CPU capacity;Refer to the pressure of bottom physical network links j, this link total bandwidth capacity determine with remaining bandwidth capacity:
Wherein,WithRefer to the CPU capacity of this physical network nodes i current residual and total CPU capacity respectively,WithRefer to bandwidth capacity and the total bandwidth capacity of this Physical Network link j current residual respectively.
2) average pressure and maximum pressure
With reference to the definition of upper node link stress, obtain average pressure and be defined as with maximum pressure:
Wherein, N, L refers to physical network nodes number of links respectively.
3) empty network planning mould
Empty network planning mould essentially describes the size of virtual network (VN), its node cpu capacity and link bandwidth determine:
WhereinRefer to the CPU capacity of node i,Refer to the bandwidth of link j.
3) node scale (Hn(i))
Node scale essentially describes node significance level in a network, the CPU capacity of this node and connection bandwidth determine:
WhereinRefer to the CPU capacity of node i,Referring to the bandwidth of link j, L (i) represents the link set being joined directly together with node i.
According to above-mentioned definition, the present invention one proposes the virtual network mapping objects that node link pressure adaptive is balanced, and this target can realize making overall plans and coordinate of node optimization and link optimizing, thus reaching the effect that overall situation pressure optimizes;Two is that the adjustable parameter in object function is designed, it is proposed that a kind of adaptive parameter prediction pattern, makes weight parameter optimize and revise gradually according to network state change and self two factor of evolution;Three stability being to ensure that this self adaptation adjustable parameter, propose a kind of mechanism that adjustable parameter convergence direction is judged, this mechanism can be passed through statistical weight parameter and judge whether system enters divergent state in span border residence time, and adopts corresponding strategy to make system again recover balanced.
(1) the empty net mapping objects of node link pressure weight adaptive equalization:
As described above, existing balanced with node link pressure be that the empty net mapping algorithm of target employs summation or seeks the simple combination optimization aim of product, or the method that two optimization aim of use are called in turn by demand, these methods are all relatively simple, cannot really embody the relation between node link pressure, so that the Optimization Work of the whole network pressure equilibrium is thorough not.Therefore present invention firstly provides the empty net of a node link pressure weight adaptive equalization and mapped optimization aim, namely minimized:
In above formula, L (i) represents the link set being joined directly together with node i, and so, this optimization aim has just considered the pressure of node i and the link stress being connected with node i, and α is that weight regulates parameter, α ∈ (0,1).NARepresent and the request of deficiency in origin net has mapped successfully empty physical network nodes set corresponding to net node, d (i, u) distance (jumping figure) between i, u is represented at 2, so, one of distance factor also major influence factors becoming this optimization aim, distance is near will reduce the difficulty that following link maps.
(2) adaptive forecasting method of weight parameter α:
In order to make the weight of node pressure and link stress can reach equilibrium, and along with the quantity of void net request, scale and the continuous adjusting and optimizing of arrival rate, the present invention proposes a kind of to the weight parameter α method carrying out adaptive prediction, namely
In above formula, α-Representing the α value of last time window, β represents the convergence rate of α, when empty net solicited status change is very fast, it is possible to increase β to accelerate the convergence rate of α.δ is greater than the real number of 0, is used for ensureing that denominator is not zero,WithThen representing the difference of maximum node link stress and average nodal link stress respectively, meansigma methods is for ensureing that system pressure benchmark constantly can adjust along with the change of void net request.If maximum node pressure is bigger with the gap of average link pressure than maximum link pressure with the gap of average nodal pressure, illustrate that node equilibrium is poor, then α increases, and causes (7) formula interior joint pressureWeight relatively big, the node that such node pressure is little can preferentially be selected, and therefore reduces maximum node pressure, completes the adaptive equalization of system;Vice versa.The initial value of α is set to [1-(1/D)], and wherein D represents the average Connected degree of physical network nodes, and β is set to 0.1, to ensure the stability of system.
(3) system mode divergence restraint mechanism:
In last point, the adjustment process of weight parameter α is an adaptive regression process, and when system runs well, the convergence of this parameter can be protected.But mapping problems is a complication system, no matter it is physical network topology, state, or virtual network topology, state, capital is on mapping it to affecting, in order to avoid occurring under special circumstances failing to the adjustment of α to bring network pressure in time it is anticipated that regulate, namely system is in the situation of divergent state, the present invention devises a system mode divergence restraint mechanism, namely when situation α < 0 or α > 1 being detected occurs, α is reset as initial value, thus can ensure that the properly functioning of system.
Accompanying drawing explanation
Virtual network under Fig. 1 time window pattern maps flow process
The substep mapping method of Fig. 2 node link pressure weight adaptive equalization
Embodiment
The concrete operations flow process of the present invention is before every minor node maps, and its optimization aim is made adaptive equalization, thus realizing taking into full account that following link maps in node mapping process, reaches the mutual balanced of physical network nodes link stress.Idiographic flow is as shown in Figure 2:
A. add up all empty net requests in this time window, be designated as set Rv。
If B. RvFor sky, entrance step F.If RvIt is not empty, chooses the empty net request VN that current empty net is largestk, add up VNkIn all empty net nodes, be designated as set Nv。
If C. NvFor sky, then node maps and terminates, and enters step E;If NvIt is not empty, then chooses the empty net node that node is largestResidue CPU is selected more than node from bottom physical networkThe physical network nodes set of CPU, be designated as Ns。
If D. NsFor sky, then the request of this void net maps unsuccessfully, enters step F;If NsNot being empty, the currency regulating parameter alpha according to weight calculates NsIn the H of each physical network nodes istressI (), chooses HstressI physical network nodes that () is minimumAnd the empty net node will chosen in step CMap to this physical network nodesOn.WillFrom NvMiddle deletion, returns step C.
E. use shortest path first to complete link maps, if mapping unsuccessfully, then next time window sent in this request or directly refuse;If mapping successfully, then update bottom physical network state, by VNkFrom RvMiddle deletion, returns step B.
F. the α value according to the physical network state computation next one time window after updating, and apply judgment mechanism examination and correction α value, this time window void net mapping terminates.
Claims (2)
1. a mapping method of virtual network for node link pressure weight adaptive equalization, the step carrying out virtual network mapping in a time window includes:
A. add up all empty net requests in this time window, be designated as set Rv;
If B. RvFor sky, entrance step F;If RvIt is not empty, chooses the empty net request VN that current empty net is largestk, add up VNkIn all empty net nodes, be designated as set Nv;
If C. NvFor sky, then node maps and terminates, and enters step E;If NvIt is not empty, then chooses the empty net node that node is largestThe residue CPU capacity of node is selected more than node from bottom physical networkThe physical network nodes set of CPU capacity, be designated as Ns;
If D. NsFor sky, then the request of this void net maps unsuccessfully, enters step F;If NsNot being empty, the currency regulating parameter alpha according to weight calculates NsIn the H of each physical network nodes istressI (), chooses HstressI physical network nodes that () is minimumAnd the empty net node will chosen in step CMap to this physical network nodesOn;WillFrom NvMiddle deletion, returns step C;Wherein,
In above formula, L (i) represents the link set being joined directly together with node i, and α is that weight regulates parameter, α ∈ (0,1);(i u) represents the distance (jumping figure) between i, u at 2 to d;NARepresent and the request of deficiency in origin net has mapped successfully empty physical network nodes set corresponding to net node;Refer to the pressure of bottom physical network nodes i,Refer to the pressure of bottom physical network links j, define as follows:
Wherein,WithRefer to the CPU capacity of this physical network nodes i current residual and total CPU capacity respectively,WithRefer to bandwidth capacity and the total bandwidth capacity of this Physical Network link j current residual respectively;
Wherein weight parameter α is defined as:
In above formula, α-Representing the α value of last time window, β represents the convergence rate of α;
N is physical network nodes quantity;
L is physical network links quantity;
WithThen represent the difference of maximum node link stress and average nodal link stress respectively, change that meansigma methods is asked along with void net for ensureing system pressure benchmark and constantly adjust;The initial value of α is set to [1-(1/D)], and wherein D represents the average Connected degree of physical network nodes, and β is set to 0.1, to ensure the stability of system;
E. use shortest path first to complete link maps, if mapping unsuccessfully, then next time window sent in this request or directly refuse;If mapping successfully, then update bottom physical network state, by VNkFrom RvMiddle deletion, returns step B;
F. the α value according to the physical network state computation next one time window after updating, and apply judgment mechanism examination and correction α value, this time window void net mapping terminates.
2. the method for claim 1, wherein the judgment mechanism of step F refers to: when situation α<0 or α>1 being detected occurs, α is reset as initial value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210116239.6A CN102664784B (en) | 2012-04-19 | 2012-04-19 | A kind of mapping method of virtual network of node link pressure weight adaptive equalization |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210116239.6A CN102664784B (en) | 2012-04-19 | 2012-04-19 | A kind of mapping method of virtual network of node link pressure weight adaptive equalization |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102664784A CN102664784A (en) | 2012-09-12 |
CN102664784B true CN102664784B (en) | 2016-07-06 |
Family
ID=46774203
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210116239.6A Active CN102664784B (en) | 2012-04-19 | 2012-04-19 | A kind of mapping method of virtual network of node link pressure weight adaptive equalization |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102664784B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103457852B (en) * | 2013-09-13 | 2016-04-20 | 电子科技大学 | A kind of survivability mapping method of multicast virtual network |
CN104038400B (en) * | 2013-11-18 | 2017-04-05 | 电子科技大学 | A kind of mapping method of virtual network across data center |
CN103595610B (en) * | 2013-11-25 | 2016-06-22 | 电子科技大学 | A kind of the anti-of non-deterministic source demand multicast virtual network ruins mapping method |
CN104753751B (en) * | 2013-12-27 | 2019-10-29 | 南京中兴新软件有限责任公司 | A kind of method and system being dynamically determined virtual network |
CN103812748B (en) * | 2014-01-20 | 2017-05-10 | 北京邮电大学 | Mapping method of survivable virtual network |
CN104320276A (en) * | 2014-10-28 | 2015-01-28 | 北京邮电大学 | Virtual network mapping method and system based on cut set |
CN104506337B (en) * | 2014-11-20 | 2018-02-13 | 北京邮电大学 | Mapping method of virtual network and device based on regional faults prediction |
CN105337834B (en) * | 2015-12-04 | 2019-03-08 | 重庆邮电大学 | A kind of mapping method under wireless network virtualized environment |
CN106100964B (en) * | 2016-08-24 | 2019-06-07 | 北京邮电大学 | A kind of method and apparatus of virtual network mapping |
CN107276664B (en) * | 2017-08-02 | 2019-07-02 | 大连大学 | Mixing void net mapping method based on the load of thresholding formula |
CN109150627B (en) * | 2018-10-09 | 2021-11-23 | 南京邮电大学 | Virtual network mapping construction method based on dynamic resource demand and topology perception |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102075429A (en) * | 2011-01-21 | 2011-05-25 | 北京邮电大学 | Virtual network mapping method based on principle of proximity |
CN102075402A (en) * | 2011-02-12 | 2011-05-25 | 华为技术有限公司 | Virtual network mapping processing method and system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7565495B2 (en) * | 2002-04-03 | 2009-07-21 | Symantec Corporation | Using disassociated images for computer and storage resource management |
-
2012
- 2012-04-19 CN CN201210116239.6A patent/CN102664784B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102075429A (en) * | 2011-01-21 | 2011-05-25 | 北京邮电大学 | Virtual network mapping method based on principle of proximity |
CN102075402A (en) * | 2011-02-12 | 2011-05-25 | 华为技术有限公司 | Virtual network mapping processing method and system |
Non-Patent Citations (3)
Title |
---|
A new algorithm based on the proximity principle for the virtual network embedding problem;Jiang Liu,Tao HUANG,Jian-ya CHEN,Yun-jie LIU;《Science Direct》;20111104;正文第910-918页 * |
New algorithm for hub-and-spoke topological virtual networks embedding problem;LIU Jiang,HUANG Tao,CHEN Jian-ya,LIU Yun-jie,LU Bo;《Science Direct》;20120228;正文第55-61页 * |
网络虚拟化资源管理架构与映射算法研究;吕博;《中国优秀博士学位论文数据库》;20110619;正文第1-108页 * |
Also Published As
Publication number | Publication date |
---|---|
CN102664784A (en) | 2012-09-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102664784B (en) | A kind of mapping method of virtual network of node link pressure weight adaptive equalization | |
WO2020181761A1 (en) | Sdn enhanced path allocation device and method employing bin-packing technique | |
Praveen et al. | An Adaptive Load Balancing Technique for Multi SDN Controllers | |
CN105471764B (en) | A kind of method of end-to-end QoS guarantee in SDN network | |
CN103650435B (en) | Routing traffic method of adjustment, device and controller | |
CN106100990B (en) | A kind of communication network various dimensions resources regulation method and system based on SDN framework | |
CN106464593B (en) | A kind of system and method for optimization routing data flow | |
CN105242956A (en) | Virtual function service chain deployment system and deployment method therefor | |
CN102664814B (en) | Grey-prediction-based adaptive dynamic resource allocation method for virtual network | |
CN103457852B (en) | A kind of survivability mapping method of multicast virtual network | |
CN105897575A (en) | Path computing method based on multi-constrained path computing strategy under SDN | |
WO2023040022A1 (en) | Computing and network collaboration-based distributed computation offloading method in random network | |
CN113572686A (en) | Heaven and earth integrated self-adaptive dynamic QoS routing method based on SDN | |
EP3793206B1 (en) | Physical optical network virtualization mapping method and apparatus, and controller and storage medium | |
CN104468688A (en) | Method and apparatus for network virtualization | |
CN106817306B (en) | Method and device for determining target route | |
CN106416157A (en) | A method to provide elasticity in transport network virtualization | |
CN103297341A (en) | Method for configuring flow on nodes of internal router | |
US20170310395A1 (en) | System and method for communication network service connectivity | |
CN110198280A (en) | A kind of SDN link allocation method based on BP neural network | |
CN105472484A (en) | Wave channel balancing route wavelength allocation method of power backbone optical transport network | |
CN111556173A (en) | Service chain mapping method based on reinforcement learning | |
CN107733796A (en) | A kind of preferentially path calculation method and system | |
CN105847146B (en) | A method of it improving level distribution SDN and controls plane router efficiency | |
CN103051546A (en) | Delay scheduling-based network traffic conflict prevention method and delay scheduling-based network traffic conflict prevention system |
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 |