A Cloud based Optimal VM Techniques for Road Traffic Data with Performance Evaluation

(1)

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 8, Issue 3, March 2018)

81

A Cloud based Optimal VM Techniques for Road Traffic Data

with Performance Evaluation

Md. Rafeeq M.Tech(PhD)

1

, Dr.C. Sunil Kumar

2

, Dr. N. Subhash Chandra

3 1

Associate Professor in CSE, CMRTC, Kandlakoya,Medchal , Hyderabad, TS, India.

2_{Professor in IT, SNIST}_{, Yamnampet, Ghatkesa}_{, Hyderabad, TS, India.} 3_{Professor in CSE, CVRCE, Manglapalli, Ibrahimpatnam R.R(D), TS, India}

Abstract- With the exponential growth of traffic and road links, it is the need of the research to explore new directions of managing and predicting the traffic situations in order to gain road safety, better traffic managements and finally gain higher productivity during pick hours by reducing the traffic burdens. Across the world ranging from city to urban the tremendous growth of road traffic is leading towards a major problem. The highly populated cities around the world are facing the problem of better traffic management. The technologically advanced cities deploy agent based multiple multimedia censor based networks to collect and analyze the traffic data to provide better solutions for management and prediction of the road conditions.

I. INTRODUCTION

[image:1.612.335.553.257.502.2]

The data collections methods include automatic and manual collections of high amount of data and then the analysis are often done in legacy system or manually. However the collected data reaches a high volume and became highly difficult to manage. Moreover the predictive analysis also demands a high computational power to run the predictive analysis algorithms. Hence considering the situations we identify the following problems to be addressed in the current traffic management situations: Management of Agent based Censor network to implement a low cost infrastructure and normalization of the data under preprocessing. Comparison and identification of most suitable cloud storage architecture for replication of data considering the low cost Erasure models.

Table 1:

Image Sensor Node Configuration

Parameter Type Proposed Optimal Value

Exposure Responsiveness

1/120 to 1/160 Seconds

Focal Ratio F/5 to F/4

ISO Film Speed Rating 100

Lance Focal Length 18 to 20 mm

Resolution 300 to 350 DPI (Horizontal

and Vertical)

Rotation and Movement Positioning

Co – Sited

On Board Compression Mode

3

APEX Brightness 6 to 7.5

Colour Space sRGB (Recommended)

Digital Zoom 5 to 8

Image Compression Mode

JPEG

(2)

International Journal of Emerging Technology and Advanced Engineering

82

II. PROPESED WORK

The programmable data accumulator agents are configured to capture the image and video data. The optimizer components into these agents are configured to capture the video or image data while the motion in the object is detected. This makes the captured data optimized in size.

The algorithm used for motion detection for multiple objects consider each object as collective vector object as

t

O_{at any given time instance}t_.

2.1 Data Accumulator Agent

The collection of the detectable features based on the time instance are noted as

1, 2.... t

F F F_{. The algorithm is deploying a probabilistic}

function to estimate the next state of the specified object

denoted asP O F F( t| 1, 2,.... )Ft _{. The outcome of this} probabilistic function is the next possible coordinates for the object under motion tacking. Further the probabilistic function at any given time is considered to function on a weighted and time depended sample set of video data.

The sample set is defined as

( )

{Stv,v1,.... }V

with the considered weight for corrected predictive

location is ( )v t



_{. The V parameter denotes the number of}

video data in the selected data set.

Henceforth we understand the relation between the learning features and the probabilistic prediction as,

Firstly we consider the sample data set with V number of video data as

( ) 0

{S v,v1,.... }V _{…Eq. 1}

The weight component for enhanced learning is considered as

( ) 0

{



v,v1,.... }V _{…Eq. 2}

Hence, the final probabilistic function for next location determination is

1 2 ( _t| , ,.... )_t

P O F F F

…Eq. 3

Henceforth we consider the change in enhanced learning factor as

1 ( t| t )

P O O _{…Eq. 4}

The next step is to apply the Eq. 4 on Eq. 1and obtain a

new dataset for prediction and obtain ( ) {Stv}

Hence the final step is to consider the updated feature tracking for iterative calculation of the motion as

( ) ( ) 1 ( | ) ( | ) v v t t V i t i

P F S P F S







…Eq. 5

Local Storage: The local storage container is designed to store data locally from the capture agents in case of loss of connectivity between the network layer and other layers. The local storage containers are desired to perform the on board compression and store small amount of data.

Table 2:

Image Sensor Node Configuration

Parameter Type Proposed Optimal Value

Technology ZIP or JAZ or USB

Preferred Mode of Operation

No or Less Mechanical Parts

Read Speed 20 to 200 MBPS

Write Speed 30 to 280 MBPS

Form Factor 2.5 to 3.5

Storage Capacity 5 to 12 GB

Supported Storage Format _JPEG

However the storage containers are expected to deliver the low latency. Here we propose the factors to be considered for a local storage to balance the performance and cost implications [Table - 2].

3.2 Applied machine learning techniques for improved navigational systems

(3)

International Journal of Emerging Technology and Advanced Engineering

83

The most effective and popular framework under Erasure Coding is Reed-Solomon framework. The

framework can be applied in case of

2

z

n



_{, where n} denotes number of disks and z denotes number of customer data. To understand the framework for 256 storage containers or disks are considered. For a 256 disks, a Reed – Solomon code can be defined and implemented using Galois Field Arithmetic or GF (28). The coefficient “a” can be defined in various ways. The basic implementation of Reed – Solomon is Couchy construction. To understand Couchy construction, we select any n unique numbers in the space of GF(2Z).

Hence the selected n number are distributed in two sets called X and Y, where X contains m

Elements and Y contains k elements. Hence:

,

1 ( )

i j

i j

a

x

y





with the help of GF (2z)..Eq 6

The most important factor that makes Reed-Solomon framework to implement is the simplicity. In this framework selecting k and m is random and does not depend on any factors and can be selected independently. The performance can be questioned as the time complexity for performing an XOR operating is less compared to GF. However the modern processors rely on vector instruction sets for performing array based multiplication operation. Hence the reduction in time for computation can be and the probabilistic prediction as, achieved. Moreover with the improvement of latency time for the I/O devices and cache memory is also been improving to match with the highly complex Erasure Codes.

The implementation of Reed – Solomon is simple as many open source solutions are readily available for storage solutions.

III. MACHINE LEARNING TECHNIQUES FOR ROAD TRAFFIC DATA

The most effective functionalities of a road traffic management system also may include the predictive analysis system as an alternative to the guided navigation system. The prediction system must be enabled to predict the road traffic conditions specially the congestion control. Hence in this work we understand the most popular machine learning predictive approaches for road traffic data analysis:

3.1.1. Gaussian Naïve Bayes

In case of continuous parameter evaluation of number of parameters, the Gaussian Naïve Bayes probabilistic solution plays the major role. The Gaussian Naïve Bayes is formulated and understood here:

2 2 ( ) 2 2

1 (

| )

2

c c v c

P x

v c

e

 



 





_{... Eq 7}

Where,

(

| )

P x



v c

_is _defined _as _any _class _for distribution

x denotes the continuous attributes used for the classification

c



is defined as mean value of the attributes

2

c



is defined as variance of the attributes

3.1.2. Multinomial Naïve Bayes

In case of the event generation probabilistic models, the Multinomial Naïve Bayes probabilistic model is widely used. Here the Multinomial Naïve Bayes is formulated and understood:

(

)!

( |

)

!

i i i x i K k i i i

x

P x C

P

x







_{... Eq 8}

Where,

P denotes the probability of any event to occur or K number of multinomial cases in multiple event classes. X denotes the number of times any event I took place in the space.

3.1.3. Bernoulli Naïve Bayes

In case of the document or generic text classification the use of Bernoulli Naïve Bayes is highly adopted. Here we formulate and understand the Bernoulli Naïve Bayes probabilistic model:

(1 )

1

( |

)

i

(1

)

i

i i

n

x x

K k k

i

P x C

P









(4)

International Journal of Emerging Technology and Advanced Engineering

84

Where,

i

k

P

denotes the probability for any class

C

K_{producing any}

term.

i

x

denotes the probability of producing any term from the possible set.

The above mentioned mining techniques are widely employed by multiple researchers in the past and recent researches. Henceforth this understanding will help us to extend the work of proposed framework by proposing the traffic prediction system.

IV. PERFORMANCE EVALUATION MATRIX

A novel matrix to evaluate the performance of the proposed migration algorithm is been coined in this work. The parameters names, details of the parameter with the optimality expectation are been proposed here [Table – 3]: The analysis of the cost matrix is demonstrated in the results and discussion section of the work

4.1 Results And Discussions

The results of simulation testing for the proposed framework under the NetSIM tool are satisfactory.

The framework is been tested on the parameters proposed in the performance evaluation matrix [Table – 4]. The results demonstrates the most effective and sustainable nature of the framework. However the due to the network congestion during the data transmission, it is been observed that the availability of the parameter values are seems not to be available for longer runs. Hence this issue will be addressed in the further research for this work.

This work has performed extensive testing to demonstrate the improvement over the existing migration techniques [6] [7] [8] [9] [10]. The various considered migration techniques are listed with the used acronyms here [Table – 5]:

(5)

International Journal of Emerging Technology and Advanced Engineering

[image:5.612.50.563.160.695.2]

85

Table 3.

Performacne Evaluation Matrix and Parameters

Parameter Details Optimality Expectation

Number of hosts _{Number of Host Machines during the simulation or testing} _{Same throughout all simulations}

Number of VMs

Number of Virtual Machines during the simulation or testing Same throughout all simulations Total simulation time

Duration of the Simulation Same throughout all simulations Energy consumption

The amount of Energy difference during migration Expected to be Minimum Number of VM migrations

Total number of Virtual machine migrations

Expected to be Mean of all the techniques

SLA performance degradation

SLA performance degradation due to migration

SLA time

SLA time per active host

SLA violation

Overall SLA violation Expected to be Minimum Average SLA violation

Average SLA violation

Host shutdowns

Number of host shutdowns Expected to be Maximum Host shutdown – Mean

Mean time before a host shutdown

Host shutdown – Standard Deviation

Standard Deviation time before a host shutdown Expected to be Minimum VM migration time - Mean

Mean time before a VM migration Expected to be Minimum VM migration time – Standard Deviation

Standard deviation time before a VM migration Expected to be Minimum VM selection mean

Execution time for VM selection in mean Expected to be Minimum VM selection time - Standard Deviation

Execution time - VM selection standard deviation Expected to be Minimum Host selection time - mean

Execution time for host selection in mean Expected to be Minimum Host selection time - Standard Deviation Execution time for host selection in standard deviation Expected to be Minimum

VM reallocation time - Mean Execution time for VM reallocation in mean Expected to be Minimum

VM reallocation time - Standard Deviation Execution time for VM reallocation in standard deviation Expected to be Minimum

Total execution time – Mean Total Execution time for VM reallocation in mean Expected to be Minimum

(6)

International Journal of Emerging Technology and Advanced Engineering

[image:6.612.89.528.160.595.2]

86

TABLE 4: Property Availability Matrix

Parameter Observation

Availability of the Properties

Test Process – 1 (Duration – 60 Mins)

Name of Node Available Available Available

Type Available Available Available

Overall Health Indicator Available Available Available

Last Accessed Available No Continuous

Availability Available

Total Availability Available Available No Continuous Availability

Memory Utilization Available Available Available

Disk Utilization Available Available Available

Network Utilization Available No Continuous

Availability No Continuous Availability

Active Memory Available No Continuous

Over heading Memory Available Available No Continuous Availability

Swappable Memory Available Available Available

Total Shared Memory Available No Continuous

Memory Temperature Available Available Available

Container Name Available Available Available

Container Size Available Available Available

Container Utilization Available Available Available

Network Device ID Available Available Available

Up Time Available Available No Continuous Availability

Down Time Available Available No Continuous Availability

IP Address Available Available Available

MAC Address Available Available Available

Data Transfer Rate Available Available Available

Device ID Available Available Available

Type Available Available Available

Read Count Available Available No Continuous Availability

Write Count Available Available No Continuous Availability

These works considers the parallel research

outcomes of recent time and find multiple approaches

to build a scalable and sustainable framework for

traffic monitoring system. Based on the parallel

research outcomes we understand the focus area of

this research. We alsounderstand the generic

components of any road traffic management and

monitoring system consisting of Sensor Nodes, Data

Collector, and Monitoring agents, Storage Controller,

Information Server and Application Server[9-12].

(7)

International Journal of Emerging Technology and Advanced Engineering

[image:7.612.67.550.156.674.2]

87

Table 5.

List of Techniques used for Performance Comparison

Used Name

in this Work

Selection Policy

Allocation Policy

IQR MC

Maximum Correlation

Inter Quartile Range

IQR MMT

Minimum Migration Time

Inter Quartile Range

LR MC

Random Selection

Local Regression

LR MMT

Minimum Migration Time

Local Regression

LR MU

Minimum Utilization

Local Regression

LR RS

Rom Selection

Local Regression

LRR MC

Maximum Correlation

Robust Local Regression

LRR MMT

Minimum Migration Time

Robust Local Regression

LRR MU

Minimum Utilization

Robust Local Regression

LRR RS

Rom Selection

Robust Local Regression

MAD MC

Maximum Correlation

Median Absolute Deviation

MAD MMT

Minimum Migration Time

Median Absolute Deviation

MAD MU

Minimum Utilization

Median Absolute Deviation

MAD RS

Rom Selection

Median Absolute Deviation

THR MC

Maximum Correlation

Static Threshold

THR MMT

Minimum Migration Time

Static Threshold

THR MU

Minimum Utilization

Static Threshold

THR RS

Rom Selection

Static Threshold

(8)

International Journal of Emerging Technology and Advanced Engineering

88

V. CONCLUSION

This work also proposes the performance evaluation matrix for evaluating the performance of any generic road traffic management and monitoring system. Load Balancing can be achieved through virtual machine migration. However the existing migration techniques constraints to improve the SLA and often compromise to a higher scale on the other performance evaluation factors. This work also proposes the performance evaluation matrix for evaluating the performance of any generic road traffic management and monitoring system. Load Balancing can be achieved through virtual machine migration. However the existing migration techniques constraints to improve the SLA and often compromise to a higher scale on the other performance evaluation factors. This work, demonstrates the optimal three phase virtual machine migration technique with up to 70% improvement to retain SLA compared to the other virtual machine migration technique. However the proposed technique is independent of the virtual machine image format and demonstrates the same improvement. The comparative analysis is been done with the proposed technique with the existing techniques like IQR MC, IQR MMT, LR MC, LR MMT, LR MU, LRR MC, LRR MMT, LRR MU, LRR RS, LR RS, MAD MC, MAD MMT, MAD MU, MAD RS, THR MC, THR MMT, THR MU and THR RS. The work also furnishes the practical evaluation results from the simulation to retain the improvement of the other parameters at least to the mean of other techniques during SLA improvement. Also this proposed technique for virtual machine migration demonstrates no loss in existing CPU utilization during load balancing[8-11].

REFERENCES

[1] R. Yu, M. Abdel-Aty, and M. Ahmed, “Bayesian random effect models incorporating real-time weather and traffic data to investigate mountainous freeway hazardous factors,” Accid. Anal. Prev., vol. 50, pp. 371–376, Jan. 2013.

[2] R. Yu and M. Abdel-Aty, “Utilizing support vector machine in real time crash risk evaluation,” Accid. Anal. Prev., vol. 51, pp. 252–259, Mar. 2013.

[3] C. Xu, A. Tarko, W. Wang, and P. Liu, “Predicting crash likelihood and severity on freeways with real-time loop detector data,” Accid. Anal. Prev., vol. 57, pp. 30–39, Aug. 2013.

[4] J. Lee and F. Mannering, “Impact of roadside features on the frequency and severity of run-off-roadway accidents: An empirical analysis,” Accid. Anal. Prev., vol. 34, no. 2, pp. 149–161, Mar. 2012.

[5] Ammar Alazab, SitalakshmiVenkatraman, JemalAbawajy and anmounAlazab, “An Optimal Transportation Routing Approachusing GIS – based Dynamic Traffic Flows”, International Conferenceon Management Technology and Applications,10-12 Sep,2010. [6] M. Ahmed, M. Abdel-Aty, and R. Yu, “A Bayesian updating

approach for real-time safety evaluation using AVI data,” Transport. Res. Rec.,vol. 2280, pp. 51–59, 2012.

[7] W. Zhao and X. Tang, “Scheduling Data Collection with Dynamic Traffic Patterns in Wireless Sensor Networks,” Proc. IEEE INFOCOM ’11, pp. 286-290, Apr. 2011.

[8] The Mobile Millenium Project, http://traffic.berkeley.edu, 2013. [9] R.K. Balan, K.X. Nguyen, and L. Jiang, “Real-Time Trip

Information Service for a Large Taxi Fleet,” Proc. ACM MobiSys, 2011.

[10] Bejan and R.J. Gibbens, “Evaluation of Velocity Fields via Sparse Bus Probe Data in Urban Areas,” Proc. IEEE 14th Int’l Conf. Intelligent Transportation Systems (ITSC), 2011.

[11] Z. Li, Y. Zhu, H. Zhu, and M. Li, “Compressive Sensing Approach to Urban Traffic Sensing,” Proc. 31st Int’l Conf. DistributedComputing Systems (ICDCS), 2011.