Building the Cloud Platform for the Next Generation Public Security

Application

Xin Wang, Jie Dai and Zheng Xu

The Third Research Institute of the Ministry of Public Security, Shanghai, China

Keywords: Big Data, Public Security Information Systems, Cloud Platform, Distributed Storage, Distributed

Computing, Data Retrieval, Virtualization.

Abstract: A great variety of public security information systems have been built for the traffic accidents governance,

crimes events and terrorist incidents prediction. However, the large-scale redundant construction of systems

leads to “great waste of IT resource” and “information overload. Technologies such as big data, cloud

computing and virtualization have been applied in the public security industry to solve the above problems.

This paper concludes a novel architecture for next generation public security system, and the “front + back”

pattern is adopted. Under the architecture, cloud computing technologies such as distributed storage and

computing, data retrieval of huge and heterogeneous data are introduced, and multiple optimized strategies

to enhance the utilization of resources and efficiency of tasks.

1 INTRODUCTION

In recent years, a great variety of public security

information systems have been built, which have

played important roles in the traffic accidents

governance, crimes events and terrorist incidents

prediction. Series of problems appear, on the one

hand, redundant construction of systems leads to

great waste of resource, such as the video

surveillance systems throughout the country, which

are built with their independent software and

hardware in each place. Furthermore, it is difficult to

organize, manage and store the large-scale

heterogeneous data including video, audio, text and

structured data collected efficiently. And the most

important is how to find valuable clues or

knowledge quickly from great amount of

information.

Technologies such as internet of things (Hu et al.,

2014; Luo et al., 2011), big data (Xu et al., 2014; Xu

et al., 2015) and cloud computing (Liu et al., 2010;

Liu et al., 2011) have been applied in the public

security industry to solve the above problems by

governments all over the world. Utah Data Center,

which was built for the American police and

government, has the large-scale storage capacity of

yotta bytes, and it has been collecting kinds of

information, including personal e-mails, phone calls,

parking receipts, travel schedules, shopping records

* The corresponding author: Jie Dai

and other records (www.nsa.gov). Boundless

Informant project has developed a cloud platform

which analyses data such as telephone, financial

information and other intelligence transferred by

wired and wireless network, satellite and other

channels, a to achieve the global target of any real-

time monitoring and network monitoring

(www.nsa.gov1.info/dni/boundless-informant.html).

ACCUMULO was a data storage software

developed by the US National Security Agency, and

submitted to Apache as an open source project in

2011 (www.accumulo.apache.org). Based on the

Google's BigTable data model, structured and

unstructured data are stored as distributed KV

format, and the properties of database security,

scalability and speed are enhanced. In Shandong

province of China, the cloud platform for police was

constructed in 2014 (www.prnasia.com), which

provides applications such as “cloud search”, “cloud

video surveillance” and so on. There are 146 kinds

of data acquired from kind of public security and

other social information systems every day, and the

number of structured data is up to 6 billion, and the

platform has strong computing ability and store

capacity for large-scale data.

Based on the research above this paper reviews

the common architectures of the public security

cloud platforms and data centers, and introduces the

techniques for big data storage, organization, and

analysis. The rest of the paper is arranged as follows.

Section 2 introduces the problems. The architectures

142

Wang X., Dai J. and Xu Z.

Building the Cloud Platform for the Next Generation Public Security Application.

DOI: 10.5220/0006020701420145

In Proceedings of the Information Science and Management Engineering III (ISME 2015), pages 142-145

ISBN: 978-989-758-163-2

are detailed in section 3. In section 4, key techniques

and optimized are described for the current problems

faced by the public security of big data. Section 5

introduces the recent work and research. In the end,

some conclusions are drawn and future work is

given in section 6.

2 PROBLEMS DESCRIPTION

Redundant construction of information systems for

public security governance, the traffic accidents,

crimes events and terrorist incidents prediction leads

to series of problems:

(1)Redundant construction of systems results in

great waste of IT resource. There exist kinds of

detective systems for police such as the video

surveillance systems, which are built all over the

country with independent software and hardware in

each place. Redundant construction of systems leads

to great waste of resource, it is still hard to discover

deep information and complex content to support

crime predictions without unified platform to

integrate the distributed data.

(2)Data collected include video, audio, text and

other unstructured, structured data, and it is

necessary to store, manage, and process them under

unified architecture and optimized strategies based

on the cloud computing technologies, so that data

computing and storage could be effectively.

3 THE ARCHITECTURE OF

CLOUD PLATFORM FOR THE

NEXT GENERATION PUBLIC

SECURITY SYSTEM

We proposed the architecture of cloud platform for

the next generation public security system, to show

how to process, organize, manage and store large-

scale heterogeneous data. As shown in Fig.1.

Due to the limited bandwidth, the “front + back”

pattern is adopted, that is: data such as video, audio

and other unstructured and structured data are

collected by sensors such as cameras or from existed

information systems and preprocessed in the “front”

part, then the results are packaged with unified

standard format and transferred to the “back” data

center with strong storing and computing ability to

support more complex computing and applications.

The pattern could avoid network congestion caused

by distributed heterogeneous data after data

preprocessing in the “front”; in addition, based on

the cloud computing and virtualization technologies,

the cloud center realizes the resource consolidation

of multiple IT resources, and provides unified

computing and storage environment for more data

analysis and applications such as data mining and

semantic reasoning.

Figure 1: The “Front + Back” cloud platform for the next

generation public security system.

Take the video surveillance systems for example,

video data are collected by cameras, in which ARM-

based processing devices are embedded to do

preprocessing such as video encoding, license-plate

recognition, colour recognition under semantic

description models, with the results including

pictures and structured description with standard

format transferred to the back center, and the

original video data stored into databases deployed

near cameras. The center provides more resources to

support deep data analysis and applications for

police.

4 STORAGE, COMPUTING AND

RETRIEVAL OF PUBLIC

SECURITY HETEROGENEOUS

BIG DATA

Data collected from types of public security systems

are huge and heterogeneous, which brings great

challenges for efficient storage and organization, fast

retrieval and computing of data. Cloud computing

technologies such as virtualization, distributed

storage and computing are applied to solve these

problems.

Building the Cloud Platform for the Next Generation Public Security Application

143

Building the Cloud Platform for the Next Generation Public Security Application

143

Figure2: Storage, computing and retrieval of

heterogeneous big data for public security cloud platform.

4.1 Distributed Computing and

Virtualization

To enhance the efficiency of tasks computing and

processing, Mapreduce, Spark, Storm and other

distributed processing model are applied to deal with

corresponding tasks such as content analysis,

semantic modelling and reasoning, other complex

data analysis and mining. Take video retrieval for

example, Mapreduce would be used to support the

task, of which the key is represented by the time in

video, and video data are divided into several parts

by the key, then all tasks execute simultaneously.

Virtualization is adopted to support IT resource

consolidation and optimum use.

4.2 Heterogeneous Data Retrieval

In current information systems with large-scale data,

data retrieval technology is basis for data access,

data analysis and any other applications. Retrieval

can be divided into three types: structured data

retrieval, unstructured data retrieval of which

pictures and text and other unstructured data are

requested and semantic search, which is the results

are the answers of reasoning in association with

repository databases.

Strategies are applied to enhance the efficiency

and availability of data retrieval. For large-scale

structured data retrieval, distributed parallel database

system and query optimization technology has been

concerned and utilized. For retrieval of unstructured

data and high-dimensional data, dimension-

reduction technology is carried out during feature

extraction and index construction, and the above

processes and tasks could run under distributed

computing frameworks.

4.3 Distributed Storage

Data collected from public security systems and

sensors include structured data, video, audio,

images, web pages and other unstructured data, also

the output results from data processing and

computing are heterogeneous. It is very critical and

necessary to make useful strategizes to store and

organize those data to support data analysis, data

retrieval and other computing tasks effectively in

different scenes.

For the structured data, parallel databases are

adopted and the storage strategy is optimized

according to the access frequency and characteristics

of kinds of tables to satisfy the data request from

kinds of applications such as statistics, retrieval,

analysis and visualization. Also, for the analyzed

applications, data could be stored in the Relational

and NoSql databases such as Hbase by columns to

support the tasks running efficiency.

For the retrieval of images and other

unstructured data, Hbase model could be applied to

store those data into distributed file systems such as

HDFS and GPFS, so that other computing models

are able to access data rapidly and efficiently.

5 PREVIOUS WORK

During 2008-2009, the third research institute of

Ministry of Public Security introduced video

structured description technology for the demand in

video surveillance applications, and undertakes a

series national science and technology major

projects including the Ministry of National Science

and Technology Support project, 863 smart city

project and the Core Electronic Devices, High-end

Generic Chips and Basic Software project. Numbers

of public security intelligent video surveillance

systems are carried out successfully, including VSD

based road surveillance video retrieval system in

Shuangliu in Chengdu, Taicang in Jiangsu Province

and so on.

6 CONCLUSIONS

In this paper, we conclude a novel architecture for

next generation public security system, and the

“front + back” pattern is adopted to address the

problems brought by the redundant construction of

current public security information systems which

realizes the resource consolidation of multiple IT

resources, and provides unified computing and

ISME 2015 - Information Science and Management Engineering III

144

ISME 2015 - International Conference on Information System and Management Engineering

144

storage environment for more complex data analysis

and applications such as data mining and semantic

reasoning. Under the architecture, we introduce

cloud computing technologies such as distributed

storage and computing, data retrieval of huge and

heterogeneous data, provide multiple optimized

strategies to enhance the utilization of resources and

efficiency of tasks.

However, some other problems still exist: in

what way the services could be provided to users,

and it still cannot satisfy the routine detection and

application for police. For example, combining

crime prediction results with visualization methods

is necessary for users during detection. These

unsolved problems particularly merit our further

study.

ACKNOWLEDGEMENTS

This work was supported in part by the National

Science and Technology Major Project under Grant

2013ZX01033002-003, in part by the National High

Technology Research and Development Program of

China (863 Program) under Grant 2013AA014601,

in part by the National Science Foundation of China

under Grant 61300028, in part by the Project of the

Ministry of Public Security under Grant

2014JSYJB009.

REFERENCES

Liu,Y., Zhu,Y., Ni, Lionel M., and Xue, G. 2011. A

Reliability-Oriented Transmission Service in Wireless

Sensor Networks. IEEE Transactions on Parallel and

Distributed Systems, 22(12): 2100-2107.

Liu, Y., Zhang, Q., and Ni, Lionel M. 2010. Opportunity-

Based Topology Control in Wireless Sensor Networks.

IEEE Transactions on Parallel and Distributed

Systems, 21(3): 405-416.

Hu, C., Xu, Z. et al. 2014. Semantic Link Network based

Model for Organizing Multimedia Big Data. IEEE

Transactions on Emerging Topics in Computing,

2(3):376-387.

Xu, Z. et al. 2015. Knowle: a Semantic Link Network

based System for Organizing Large Scale Online

News Events. Future Generation Computer Systems,

43-44:40-50.

Xu, Z., Luo, X., Zhang, S., Wei, X., Mei, L., and Hu, C.

2014. Mining Temporal Explicit and Implicit

Semantic Relations between Entities using Web

Search Engines. Future Generation Computer Systems,

37:468-477.

www.nsa.gov

www.nsa.gov1.info/dni/boundless-informant.html

www.accumulo.apache.org

www.prnasia.com

Building the Cloud Platform for the Next Generation Public Security Application

145

Building the Cloud Platform for the Next Generation Public Security Application

145