Topology Robust Study in Multiple Cloud Computing Environment

2017 2nd _{International Conference on Artificial Intelligence and Engineering Applications (AIEA 2017)}

ISBN: 978-1-60595-485-1

Topology Robust Study in Multiple Cloud

Computing Environment

M. YUAN and J.H. DUAN

ABSTRACT

In order to realize the optimization of multi database environment index of cloud, it needs to construct index topology, and robustness of the proposed control, depth first traversal mechanism in cloud computing environment index design method based on robust topology. Constructing adaptive robust multilayer routing detection protocol, build cloud computing database the index under the load balance topology control node distribution model, through the transmission cost function of intra cluster and inter cluster index, quantitative equation to build cloud computing node topology multi database environment, quantitative analysis of the quantitative recursive equation, realize the accurate positioning of topological node deployment, and uses the depth first traversal mechanism and the minimum hop count routing method to design robust topology structure, improve the ability of cloud computing database multilayer index environment. Experimental results show, this method is used to design index topology, which improves the efficiency of database multi index, improves the timeliness of database access and the accuracy of data retrieval.

KEYWORDS

Cloud computing environment; multi index; topology; robustness

INTRODUCTIONS

With the development and application of stochastic network information technology, the main form of cloud computing gradually developed into modern network information computing and processing. Cloud computing is a computing concept in recent years, the basic idea is to use the idle network, distributed in other local resources, such as computer hardware resources, software resources, centralized intelligent scheduling and information processing of network information resources. In the cloud computing grid space, cloud computing grid space resources needed for automatic deployment management, dynamic expansion, according to need, the effective integration of cloud data center resources, cloud computing environment database multi index control, improve the utilization rate of cyber source cloud computing improve access and scheduling capability, cloud storage database, research in cloud computing environment index topology robust control problem, improve access to database and data Scheduling capability[1].

_________________________________________

Cloud computing topological structure of multilayer database index topology environment refers to the physical layout of the transmission medium interconnection of various equipment’s, with the network nodes and the distributed edge vector and reasonable arrangement and distribution of array logic network, adjusting the physical connection and node, improve the database access control the survivability and robustness in the traditional method [2]. The management method and the linear programming model of grid resources on cloud computing grid spatial index topology design algorithm of prickle a, building a storage virtualization and computing resource pool, the pool of computing resources, the mathematical model of linear programming, the existing computing resource discovery mechanism based on multiple cloud computing environments topology design. Lead by assuming that the sender receives the information sent to the host command, a virtual network topology, the realization of P2P network topology based on HTTP protocol Architecture, topology design model of multiple modulation using multivariate statistical method and time series model for modeling, using Elman neural network tracking model to analyze structure modeling of the network survivability and robustness design, network variable structure model has good stability, but the complex system, the robustness index topology design and control not good[3].

In view of the above questions, this depth first traversal mechanism in cloud computing environment index design method based on robust topology. Constructing adaptive multilevel routing robustness detection protocol, build cloud computing environment database index load balancing topology control node distribution model, transmission cost function construction of intra cluster and inter cluster index, depth first traversal mechanism and the minimum hop count routing method to design robust topology using cloud computing, improve the ability of multi-tier database environment. Finally, the index simulation experiment is taken and valid conclusions are obtained.

TOPOLOGY ROUTING DETECTION PROTOCOL AND NODE DISTRIBUTION DESIGN

Topology routing detection protocol

(a) (b)

(c) (d)

Figure 1. Cloud computing environment of multi data topology element mode.

In Figure 1, a cloud computing communication between network topology model of multi-layer database index node under the environment of mutual equivalence, a model called the structure of P2P network. But the combination of complex network, it is vulnerable to malicious attacks and input error, disk failure, network overload behavior leads to the collapse of the system[4], cause the index performance is not good, need to optimize the topology structure design, construction of multi topology routing index detection protocol. If cloud computing environment of multi-tier database index topology for node set is V { , , ... }v v v1 2 3 vN ,

N is the number of nodes, including server nodes and wireless nodes; the set of

edges is E{ , , ... }e e e1 2 3 eM , M is the number of edges, edge vector is the path

between nodes, generate m isolated nodes, the connection probability under the

cloud computing environment isolated node multilayer database indexing topology and edge vector is:

, 1

1 n

a k a

k

Pe y

n 





In the formula, yk a, represents a node divergence, a multi-layer database index concept lattice topology can be expressed as P(A,B).T. , B(M) called node

extension, A,Balso meet the following two conditions:

1)_AB'_{_gG|_mB,_gIm}

2) ' { | , Im}

g A g M m A

B    

That is, A is an object set that has all the attributes inB, andBis a collection of

common attributes of all objects that satisfy the conditional P inA, and the concept

lattice with such a structure is called the constraint concept lattice [5].

Database index load balancing topology control node distribution model

Build the database index load balancing topology control node distribution model in the cloud computing environment, the topological structure of the index model is a graph model, the initial routing number is1, 2,3, 4,5_{, and the routing}

1,3,5, 2, 41, 4, 2,5,32, 4,1,3,52, 4,1,5,32,5,3,1, 4

3,1, 4, 2,53,1,5, 2, 44, 2,5,3,13,5, 2, 4,15,3,1, 4, 2

3,5,1, 4, 24,1,3,5, 25, 2, 4,1,35,3,1, 4, 2

According to the routing decision model, the node deployment of the multi-layer database is carried out in the cloud computing environment [7]. The topological structure of the index model is a graph model, and the initial routing

number is1, 2,3, 4,5_{, the routing decisions are arranged in the following order:}

1,3,5, 2, 41, 4, 2,5,32, 4,1,3,52, 4,1,5,32,5,3,1, 4

3,1, 4, 2,53,1,5, 2, 44, 2,5,3,13,5, 2, 4,15,3,1, 4, 2

3,5,1, 4, 24,1,3,5, 25, 2, 4,1,35,3,1, 4, 2

According to the routing decision model, the deployment of nodes in a multi-tier database under cloud computing environment is carried out.

ROBUSTNESS DESIGN OF TOPOLOGY

Node topological quantization equation

According to the transmission cost function of intra cluster and inter cluster index, quantitative equation to build cloud computing node topology multi database environment, under sampling method in cloud computing the optimal time window design of multi-tier database send data packets to the environment, establish the model of optimal deployment of nodes, considering the coverage factor clustering radius, cloud computing the cluster head election control multi database environment, assuming cloud projection vector I send data packets to the multi database environment, using the least squares method to solve the multi path fading channel regularization, get:

T ₁

(  )

 

a X X I X y (2)

T T ₁

(  )

 

a X X X I y (3)

With the discrete Fourier transform T

i i

z A x _{the scheduling time window vector}

X projection data packets to a low dimensional space d

R , ni calculated the distance

to the base station, by using the nonlinear mapping :( )X _{to cluster head}

candidate vector X is mapped to a high dimensional space, ( )X ( ),..., (x1  xm),

assuming the time window distribution of nuclear matrix _{[ ( , )] ( ) ( )}T

i j

k  

 

K x x X X _has

the center. By solving the cloud computing solutions of two adjacent nodes in multi database environment covering contribution, calculate the energy coefficient between two nodes:



 

Wy y

Kβ y (4)

Wherein, K is the Euclidean distance vector quantization matrix of multi-layer

database index node ( , )x ys s and relay node s in cloud computing environment, β is

the combination coefficient corresponding to y_{, and} β_{can be solved by the lower}

1

(  )

 

β K I y (5)

Forβ_{, the first energy threshold} KI_{on each node is decomposed by Chelsey,}



 

K I _LLT_{, solving method of W and y is same as above, through the calculation}

of SN nodes and sink time window distribution, quantitative analysis of recursive node multi-tier database extension.

Topology depth first traversal mechanism

Resume time slot matrix s t

R



C of Topology Routing hypothesis is a real

matrix, the edges of the network overload assignment matrix is rank( )C t ,

intermediate nodes of the initial routing information D _{C C}T _Rt t is symmetric

positive definite matrix. Set e1 f1 ,

T 1 T 1 , 2,3,..., k i k

k k i

i i i

k t





 



e f e

e De orthogonal column

vectors, the adaptive robust multi route detection target function:

N E n E r V N i r a i r res

res



  1 2 } ) ( { )

( (6)

Wherein, ( i)

r res n

E represents the exact transfer probability of the most transmitted

packets in the round r and the node ni, and adopts the depth first traversal mechanism to obtain the topological structure of the measurement vectors of the multi index of the database:

Along with the adaptive optimization of position, the residual energy as the constraint vector optimal model parameters for stable topology design. The stability of the network topology design mode, node position distribution of Euclidean distance is:

T

T T

( , ) ( )

( ) ( )

i j i j i j

i j i j

dist    

  

z z z z A x x

x x AA x x (7)

Wherein, e e1, 2 on the matrix orthogonal with D , and satisfy the

T _{0; , 1,2,...,}

i j i j l i,  j

e De _{, when} k l 1 , along with the change of position

measurement error of routing 1 TT

1

k

i k

k k i

i i i





 



e f e

e De , routing list corresponding to the

main diagonal, the equivalent vector group of A derivedIorthogonal, the shortest

path multi node topology database index cloud environment connectivity, design and Realization of robust topology.

SIMULATION EXPERIMENT

average degree of the generated network model with Gnutella and network model in the network, the performance of the three aspects of the average efficiency of the evolution model of degree distribution and node, node number N=500, simulation database index number is 200000 times, the topological structure of the coverage area is set to the size of 500*500, the connectivity probability of three layer

network nodes for p1 =80%, p2 =30%, p3 =15%. The host node with [1-100]

distribution, the server node with [150, 250] distribution, distribution diagram of 500 nodes is shown in Figure 2.

100 ₁₀1 ₁₀2 ₁₀3

10-5 10-4 10-3

10-2

10-1 100

k

p(

k)

=-2.8076±0.2153 R2_=0.9012

(a)

100 ₁₀1 ₁₀2

10-5 10-4 10-3 10-2 10-1 100

k

p(

k)

=-2.6551±0.2529 R2_=0.8712

(b)

100 101 102 103

10-5 10-4 10-3 10-2 10-1 100

k

p(

k)

=-2.5341±0.1782 R2=0.9199

(c)

From Figure 3 the fitting result shows that the network topology model constructed in this paper for the scale-free network, said the power index is between 2 to 2.8, the database index distribution topology with self-similar characteristics, improve the load and capacity data index database. According to the topology model of cloud computing, network topology design and analysis the test of multilayer database environment, by using this method and the traditional method, the packet forwarding number test node forwarding number under the same data, get the results shown in Figure 3.

The capacity contrast of the network index channel under different node deployment is shown in Figure 4.

10 20 30 40 50 60 70 80 90 100 0

50 100 150 200 250 300 350 400 450

Dat packet size/Gbit

the

num

ber

o

f

for

w

ards

Traditional method New method

Figure 3. Comparison of forwarding times.

10 20 30 40 50 60 70 80 90 100 2000

4000 6000 8000 10000 12000 14000 16000

Nodes number

C

apac

ity

inde

x/

G

bi

t

Analysis of Figure 3 and Figure 4 shows that using this method to index topology design, it has high forwarding number of nodes, and improve the efficiency of database index, the accuracy of database access and timeliness of data retrieval is improved.

CONCLUSION

This paper presents a computing environment of multiple depth first traversal mechanism based on cloud topology robust design method. Adaptive multilevel routing robustness detection protocol is constructed, build cloud computing environment database index load balancing topology control node distribution model, through the transmission cost function of intra cluster and inter cluster index, quantitative equation to build cloud computing multi node topology database environment, quantitative analysis of the quantitative recursive equation, realize the accurate positioning of topological node deployment, and uses the depth first traversal mechanism and the minimum hop count routing method to design robust topology structure, improve the ability of cloud computing database multilayer index environment. Research shows that using this method to index topology structure design, it can improve the efficiency of multi index database, the performance of database access and timeliness of data retrieval is better, data index accuracy has been improved.

ACKNOWLEDGEMENTS

Project of Hainan Natural Science Foundation of China in 2017: “Research on Multi Index Strategy and Intelligent Fusion in Cloud Computing Environment (Project Number: 617178)”

REFERENCES

1. Liao Jun, Jiang Chaohui, Guo Chun, et al. Classification Anonymity Algorithm Based on Weight

Attributes Entropy [J]. Computer Science, 2017, 44(7): 42-46.

2. Lee Y.Y., Wang C.H., and Huang Y.H. A hybrid RF/ baseband preceding processor based on parallel-index- selection matrix-inversion-bypass simultaneous orthogonal matching pursuit for millimeter wave MIMO systems [J]. IEEE Transactions on Signal Processing, 2015, 63(2): 305-317.

3. Czibula G., Marian Z., Czibula I.G. Detecting software design defects using relational association rule mining [J]. Knowledge and Information Systems, 2015, 42(3): 545-577.

4. MORADI M., KEYVANPOUR M.R. An analytical review of XML association rules mining [J].

Artificial Intelligence Review, 2015, 43(2): 277-300.

5. Chen Y.C., Lin Z.H., Zhao X., et al. Deep learning-based classification of hyperspectral data [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(6): 2094-2107.

6. Wan L.J., Tang K., Li M.Z., et al. Collaborative active and semi supervised learning for hyperspectral remote sensing image classification [J]. IEEE Transactions on Geoscience and Remote Sensing, 2015. 53(5): 2384-2396.