(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 1, 2011
An Extensible Cloud Architecture Model forHeterogeneous Sensor Services
R.S.Ponmagal
Dept of CSEAnna University of TechnologyTiruchirappalli, Indiarsponmagal@gmail.com
J.Raja
Dept of ECEAnna University of TechnologyTiruchirappalli, Indiaraja@tau.edu.in
Abstract
—
This paper aims
to
examine how the sensorinformation can be shared, through a new resource called“cloud”. The recent research issue in integrating Wireless sensorNetworks with Cloud is to establish a faster communication linkbetween the two. The wireless sensor networks are used to senseand collect information required. The sensor information isdeployed into the cloud through the sensor profile for webservices. The Sensor Profile for Web Services specifies alightweight subset of the overall Web services protocol suite thatis appropriate for network-connected sensors. Cloud generallyoffers resources on demand. Since wireless sensor networks arelimited in their processing power, battery life and communicationspeed, cloud computing usually offers the opposite, which makesit attractive for long term observations, analysis and use indifferent kinds of environments. In this paper a model ispresented, which combines the concept of wireless sensornetworks with the cloud computing paradigm, and show howboth can benefit from this combination. Sensor data access is thusmoved from loosely managed system to a well managed cloud.The integration of sensor information into the cloud through thesensor as a service paradigm proves the faster communicationestablishment. The scalability of this approach seems to beunlimited, since wireless sensor networks operate independently,and are connected to the cloud computing environment through ascalable number of wireless sensor network communicationgateways. The cloud computing environment itself offers ascalable infrastructure, which makes it very attractive. The maingoal is to design a flexible architecture in which sensor network’sinformation can be accessed by applications efficiently.
Keywords-
cloud, sensor profiles, webservices, sensors, resources, information, scalablity
.
I.
I
NTRODUCTION
Wireless Sensor Networks consists of energy constrainedsensor nodes and a Sink node with higher processingcapabilities. The sensors are physically composed of electronicsensing circuitry, a processor and a wireless transceiver, plus apower supply unit (battery). Sensor networks are distributedevent based systems that focus on simple data gatheringapplications and operate notably differently from that of traditional computer networks. The gathered data can be madeaccessible to other nodes, including a specialized one calledsink through a variety of means. TOSSIM is used to generatesensor data. The future sensor networks are envisioned ascomprising heterogeneous devices assisting to a large range of applications. Interoperability is required for suchheterogeneous devices. To achieve this, we propose a ServiceOriented approach for the data acquisition from sensornetwork, and an extensible architecture in which this webservices based deployment is extended to CLOUD. Here thesensor nodes are service providers and applications are clientsof such services. Hosting a web service challenges battery life,bandwidth, processing power constraints of low power sensornodes.The language in which the sensors are proposed tospeak is Sensor profiles for web services. The sensorinformation is also planned to be deployed into the cloudthrough the sensor profile for web services. The Sensor Profilefor Web Services specifies a lightweight subset of the overallWeb services protocol suite that is appropriate for network-connected sensors. The proposed Sensor profiles for webservices reduce the data processing at sensor nodes, whilekeeping the complex data processing at sink. The sensorprofiles reduces the power consumption of sensor nodes, hencemaximizes the network life time. The Sensor Profile prescribeshow to use elements of core Web services specifications toenable these functions: Send more secure messages to and froma Web service, Dynamically discover a Web service, Describea Web service Subscribe to, and receive events from, a Webservice. Complete set of functionalities for sensor integrationand limited constrained functionalities of sensors can bespecified. It further reduces the interdependencies between thesensors. The huge amount of data, which a sensor network isable to deliver, demands a powerful and scalable storage andprocessing infrastructure. Depending on the sample frequency(e.g. from 100 Hz or more down to few samples a day forcalculating observations) of the sensors, the deployedinfrastructure has to scale up memory, storage and processingpower. Today wireless sensor network platforms(e.g.TOSSIM, Crossbow MicaZ, Sentilla JCreate, SunSpot)that perform sensing and complex calculations are most of thetime constrained in their capabilities and therefore is oneappropriate way to solve this issue is to do offline processing of sensor data if the resources are not sufficient.In this paper a model is presented, which combinesthe concept of wireless sensor networks with the cloud
147http://sites.google.com/site/ijcsis/ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 1, 2011
computing paradigm, and show how both can benefit from thiscombination. Sensor data access is thus moved from looselymanaged system to a well managed cloud. The integration of sensor information into the cloud through the sensor as aservice through sensor profiles for web services languages willprove the faster communication establishment.The scalability of this approach seems to be unlimited,since wireless sensor networks operate independently, and areconnected to the cloud computing environment through ascalable number of wireless sensor network communicationgateways. The cloud computing environment itself offers ascalable infrastructure, which makes it very attractive. Hence,the sensed information is deployed into the STAX, cloudarchitecture. The combination of wireless sensor networks,with their huge amount of gathered sensor data and theirlimited processing power, with a cloud computinginfrastructure makes it attractive in terms of
i)
integration of sensor network platforms from different vendors,
ii)
scalabilityof data storage,
iii)
scalability of processing power for differentkinds of analysis,
iv)
worldwide access to the processing andstorage infrastructure,
v)
resource optimization,
vi)
be able toshare the results more easily, and
vii)
using pricing as one morecriteria for the IT infrastructure.The present work defines the proposed architecturalcomponents as well as the protocol stack of the SPWSmiddleware. The paper is organized as follows: Section IIcovers the state of the art. Section III describes the relatedwork. Section IV,V,VI and VII details the system architectureand the related components. Section VIII shows theperformance analysis. Section IX outlines the conclusion.II.
S
TATE OF THE
A
RT
The sensor information can be transmitted to the requestingclient as [1] SOAP messages, which is used to access thesensed information with application independent protocol. Aservice approach for the design of wireless sensor networks isexplained. Services are defined as the data provided by sensornodes and the applications to be executed on those data. Clientsaccess the sensor network by submitting queries to thoseservices.The DPWS proposal is optimized as TinyDPWS [2]with application specific format technique, reduces the energyconsumed by the sensor nodes. An advanced middlewaresolution to the
problem of integrating a Wireless SensorNetwork into the information system of an enterprise at a highabstraction level. This is achieved by using the proposedmiddleware which provides to the wireless sensors a ServiceOriented Architecture connection to the Internet. The proposedmiddleware is based on the Device Profile for Web Serviceswhich is a Service Oriented Architecture technology at thedevice level.A method to access the sensor information usingstructured data [3] and WSDL descriptions is proposed. Thefunctionality and data provided by the new nodes is exposed ina structured manner, so that multiple applications may accessthem. The result is a highly inter-operable system wheremultiple applications can share a common evolving sensorsubstrate. A key challenge in using web services on resourceconstrained sensor nodes is the energy and bandwidth overheadof the structured data formats used in web services.Integrating wireless sensor networks in heterogeneous net-works is a complex task. A reason is the absence of astandardized data exchange format that is supported in allparticipating sub networks. XML has evolved to the de factostandard data exchange format between heterogeneous net-works and systems. However, XML usage within sensornetworks has not been introduced because of the limitedhardware resources. In this paper, an XML tem-plate objectsare introduced making XML usage applicable within sensornetworks. Different optimized way [4] of using XML isspecified. This new XML data binding technique providessignificant high compression results while still allowingdynamic XML processing and XML navigation.The standard device profiles for web services [5]which could be used for wireless sensor networks is proposed.Even if DPWS is the best candidate to integrate WSN inexisting infrastructures, it cannot be applied to WSN withoutresearch efforts, because it addresses softer resource constraintsas required in WSN. But DPWS provides a minimal set of constraints for applications in resource constrained devices. Sothis paper describes an approach that further restricts DPWSfor WSNs, but keeps it still interoperable with DPWS.A cloud storage platform [6] for pervasive computingenvironments such as wireless sensor networks is explained.Data storage and sharing is difficult for these sensors due to thedata inflation and the natural limitations, such as the limitedstorage space and the limited computing capability. Since theemerging cloud storage solutions can provide reliable andunlimited storage, they satisfy to the requirement of pervasivecomputing very well. Thus a new cloud storage platform isdesigned which includes a series of shadow storage services toaddress these new data management challenges in pervasivecomputing environments, which called as “SmartBox”.An efficient way of combining cloud computing andwireless sensor networks [7] is explained. The cloud providesscalable processing power and several kinds of connectableservices. This distributed architecture has many similaritieswith a typical wireless sensor network, where a lot of motes,which are responsible for sensing and local preprocessing, areinterconnected with wireless connections. Since wireless sensornetworks are limited in their processing power, battery life andcommunication speed, cloud computing usually offers theopposite, which makes it attractive for long term observations,analysis and use in different kinds of environments.Several service discovery protocols for wirelesssensor networks [9] are proposed. In addition, to reduce powerconsumption we presented an activation schedule, based on themapping of the nodes’ operational modes to Bluetooth states.By announcing the activation schedule as a service, arepresentation of the state of the nodes is exposed to clientapplications.The proposed work takes into account of deploying the sensed data in STAX cloud using sensor profilesfor web services approach.
148http://sites.google.com/site/ijcsis/ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 1, 2011
III.
SOA MODEL FOR SENSOR INFORMATION SYSTEM
Service Oriented Architectural model for representation of sensor services is shown in Fig.. The SOA model has threeelements namely Sensor Service Provider, Sensor SystemRegistry and Sensor Systems Client. The sensor systemservices are categorized in to Pressure sensing, Temperaturesensing, and Level sensing services. A sensor Service provideroffers the above services and describes the interfaceinformation of the services in interface description languagecalled SensorSDL (Sensor Services Description Language)which is in the form of XML that makes the services availablein the Sensor System Registry.
Figure 1.SOA model for sensor information system
Services are the key building blocks of SOA. A service is areusable function that can be invoked by another componentthrough a well-defined interface. Services are loosely coupled,that is, they hide their implementation details and only exposetheir interfaces. In this manner, sensor system client need notbe aware of any underlying technology or programminglanguage which the service is using. The loose couplingbetween services allows for a quicker response to changes thanthe existing conventional applications for sensor applications.This results in a much faster adoption to the need of applications which makes use of sensor applications.The sensor system clients discover the service available inthe registry by service names and acquire the interfaceinformation by SensorSDL of the sensor services. Based onthis information, the clients have a binding with the sensorservice provider and can invoke services using Simple ObjectAccess Protocol (SOAP).IV.
T
HE
E
XTENSIBLE
C
LOUD
A
RCHITECTURE FOR
S
ENSOR
I
NFORMATION
S
YSTEM
A.
Introduction
The Service Oriented Architectural model for sensorinformation is extended to cloud architecture.
This paperproposes an advanced middleware solution to the problem of integrating a Wireless Sensor Network into the informationsystem of a cloud at a high abstraction level. This is achievedby using the proposed middleware which provides to thewireless sensors a ServiceOriented Architecture connection tothe Internet. The proposed middleware is based on the sensorProfile for Web Services which is a Service OrientedArchitecture technology at the sensor level. Since thistechnology is based on exchanging eXtensible MarkupLanguage documents, a technique is utilized which compressesand reduces the data volume of such documents at a level thatcan be handled by the use of the resource constrainedenvironment of the wireless sensors. By utilizing the proposedmiddleware which implements only the basic functions of thesensor Profile for Web Services, we demonstrate how such aWireless Sensor Network can be connected to the Cloud inwhich all its components conform to a Service OrientedArchitecture standard.
Figure 2.Proposed cloud architecture
B.
System Design1)
Sensor as software
T
he sensor data is obtained from TOSSIM. TinyOSsimulator is run on TinyOS1.7. NesC is the language used tosimulate the sensor nodes. The TOSSIM itself got the packagesto simulate real time sensors. Cygwin is used in windowsplatform to run the TOSSIM. In one cygwin window, thecommands as in the Fig 3. is run. As a result a sensor node issimulated and its sensed parameters are written into tossim.txtfile in the following path: C:\ProgramFiles\UCB\cygwin\opt\tinyos-1.x\apps\Sense\tossim.txt. Thesense folder also contains two NesC files called configuration(Sense.nc)file and Module(SenseM.nc)file.
149http://sites.google.com/site/ijcsis/ISSN 1947-5500
Search History:
Result 00 of 00
00 results for result for
p.
An Extensible Cloud Architecture Model for Heterogeneous Sensor Services
This paper aims to examine how the sensor information can be shared, through a new resource called “cloud”. The recent research issue in integrating Wireless sensor Networks with Cloud is to establish a faster communication link…
(More)
1.4K
Reads
4
Readcasts
4
Embed Views
More From This User
Notes
Load more
