Enhancing Mobile Cloud Computing Platform By Using Semantic Technology.
Abstract' Cloud computing has recently evolved as one of the latest emerging industry in today's world.Cloud computing offers us many challenges and there is a strong necessity to overcome all these challenges before adopting the cloud platform.Mobile cloud computing is the combination of cloud computing and mobile networks so as to provide the users rich mobile experience and bring benefits to mobile operators,network providers and cloud vendors.The main challenge in the cloud platform is the interoperability between various vendors.With the help of semantic modelling ,the issue of interoperability can be easily solved and heterogeneous data from different cloud providers can be easily connected and a smooth usage experience can be delivered to the mobile users.Semantic technology also solves the issue of data portability.
Index terms-Semantic Modelling,Interoperability and data portability
Semantic technology in web applications is an information that is linked together and can be easily processed by the machine in the global market.Semantics are usually developed in syntactical approach by making making use of uniform resource identifiers.Semantic technology have ontological models that help in efficient search mechanism which help for data representation in a more compact form.Semantic tools are available autorecognition of topics ,meaning extraction and categorization.
As defined by NIST,the three types of services in the cloud platform are Infrastructure as a service which has a high amount of workload and a little automation with respect to deployment and managing an application.Usually this type of services consists of infrastructure components like servers,storage and networking.The next type of services are platform as a service and software as a service which have less amount of workloads but at the cost of portability and scalability.With this environment set up the cloud consumer faces four challenges.First based on the requirements of the application,legal constraints and other factors the cloud consumer has to choose the respective cloud provider.The details are provided by cloud provider in different formats and different levels.Second the cloud consumer must know the technical aspects and the workflow of the cloud provider.Thirdly the cloud consumer must build an application and customise his application according to his requirements and must also be aware of the technical details like the programming language he is going to use,which will again be specific to the cloud vendor.Finally after the deployment of the application if the cloud consumer wishes to change the service provider which happens most of the times,two major difficulties arise,the cloud consumer must make changes to his application so that it suits the environment of new cloud provider,for IaaS clouds ,the effort is minimal but for the paas and saas data portability requires more effort.Semantic modelling will help us to overcome all these difficulties and provides us a better perception.
Cloud interoperability becomes a major challenge through the different landscapes of the cloud as shown in Fig no 1 from citation.There are basically two types of heterogeneity.One type is the vertical heterogeneity which exists within a single silo and this issue can be solved with the help of middleware or common set of API,the other type is horizontal heterogeneity that exists across the silos.Since they have different abstraction levels ,overcoming this is much more difficult and a high level modelling is necessary especially when an application makes a transit across the cloud silos.
Fig no 1. Cloud Landscape
Many small cloud organizations make the horizontal transit.Paas provides the ready platform for the application set up where as the Iaas clouds provides us a better option when the application scales up and more optimal also.However both types of heterogeneities have the common semantic concepts.Most of the Iaas clouds follow the common workflow type in terms of allocation of resources while the implementation tools differ.
There were many semantic modelling approaches proposed to address the interoperability issue,since earlier models involved a huge investment,they were not implemented much as per citation.However with the advancement in technology,many semantic models arise with new functionality and can provide more details compared to the older modelling approach.These new semantic models provide us a better insight and knowledge.Ex-Ontological semantic models help us in better reasoning and is an ease to make an inference.With an advantage of these capabilities many domain models can be created.Many developers use the web service description language to build syntactic representations but, the WSDL doesn't provide us the required semantic details.Semantic annotations web service description language is developed to overcome such difficulty.
III. PROPOSED METHOD
Based on our study,We basically identify four different types of semantics functional semantics,data semantics,non-functional and system semantics as shown in Fig no 2.Software lifecycle is necessary for modelling the requirements.Software lifecycle has different phases like like design and analysis phase,development phase,deployment and management phase.Some requirements may not be modelled at the time of development but will be taken in to consideration at the time of deployment.Example for such type of requirements are system and non functional requirements.This type of categorization assist us to model the semantics at proper time during lifecycle stage.There are various cloud prototypes for non functional requirements such as elastic deployment
Platform as a service
Limited set of Configurations
Infrastructure as a service
Hardware and system level configurations
Software as a service Configurations speci fic to application
modelling language,elastic computing modelling
language and elastic management modelling
But still there is a high necessity for cloud
models in data semantic stage and functional
semantics stage.By using the semantic web
annotations and web service description language
we will build the required prototype for the data
Fig no 2.Multicloud Model with semantics.
To solve the issue of interoperability among
the heterogeneous clouds we will add semantics to
functional requirements and also to data
requirements.The operations of the cloud
environment are exposed to the user with the help
of web services but different vendors have different
API's.Adding a metadata through the annotations
would help us to solve the interoperability issue
among the heterogeneous clouds.There are many
high level models like UML which describe
functional aspects of an application and use the
artifacts developed form this type of modelling for
development purpose.Often this process is called as
an model driven approach.However transforming
directly a high level model to artifacts creates a
confusion for most of the developers.Also the uml
models depend on the object oriented languages and
heavily dependent on the tools in that
platform,which is an disadvantage of UML.A
suitable form of conversion mechanism is needed
We will address this issue with use of
Domain specific Languages.DSL is a light weight
model specific to the domain with no dependency
on any tools.Here lightweight means that the dsl
models do not use languages pertaining to
knowledge representation.DSL can be used to
connect the facts and the high level
models.Annotations are used to link the models as
shown in Fig no 3.
Fig no 3.Representation of dsl annotations in (a)
and link to high level semantic models.
Semantic Annotations are built form the semantic
meta data models and which are later passed on to
the DSL design and development phase as shown in
Fig no 4 .We also have the middleware to access
the cloud platform services.Since the abstraction
provided by dsl is not enough we have semantic
meta models to describe us the abstractions.The
same concept is for the virtualization but here we
use an generic approach where artifacts are created
instead of virtual machine based interpretation.The
generated DSL scripts is then verified by various
testing tools and graphical abstractions can be
generated. Since dsl is specific to its domain we propose three different types of dsl's.
Type-1 dsl is used for development of an application.Different dsl's describe different applications.
Type 2 dsl is used for non functional specifications.
Type 3 dsl is used for gathering the artifacts of an application which include both functional and non functional configurations
Fig no 4.DSL Model For Semantic Annotation.
The middleware gives us the service level abstractions needed for deploying and managing an application.Also the middleware chooses the respective cloud environment for an application.The middleware also initiates the enforcement policy pertaining to service level agreement.Hence an automated service level agreement operation reduces the burden of the cloud consumers.
IV.CONCLUSION AND FUTURE WORK.
With the help of rich semantic meta data models,we can address the issue of interoperability between the heterogeneous clouds.Semantic data models are used to create the intermediate transformations before directly converting the exectuble artifacts of an application to high level models.Also the semantic meta data models have an advantage of automated service level agreement enforcing mechanism.Semantic models provides us a better reasoning approach.In our prototype we have shown that the semantic models use the middleware layer to operate with the cloud services to assist the data migration operation.The DSL engines can be deployed on any web server.With the help of service level abstractions provided by semantic models we can reduce the number of lines of code
of an application.
In future we would like to model the RDF data in to semantic models so that we can easily use the web services to provide us the semantic abstractions.Also there is a need for an efficient data portability mechanism by utilizing the semantic cloud models.
 A. Sheth and A. Ranabahu, 'Semantic modeling for cloud computing,part 1,' IEEE Internet Computing, vol. 14, no. 3, pp. 81'83, 2010.
 A. Sheth and A. Ranabahu, 'Semantic modeling for cloud computing,part 2,' IEEE Internet Computing, vol. 14, no. 4, pp. 81'84, 2010.
 M. Satyanarayanan, P. Bahl, R. Caceres, and N. Davies, 'The case for vm-based cloudlets in mobile computing,' Pervasive Computing, IEEE, vol. 8, no. 4, pp. 14'23, 2009.
 Zohreh Sanaei,Rajkumar Buyya,Abdullah Gani,Saied Abolfazli,'Heterogeneity in Mobile Cloud Computing:Taxonomy and Open Challenges'.IEEE,2013.
 M. Sabou, 'Smart objects: Challenges for semantic web research,'Semantic Web, vol. 1, no. 1, pp. 127'130, 2010.
 O. Corcho and R. Garc'?a-Castro, 'Five challenges for the semantic sensor web,' Semantic Web, vol. 1, no. 1, pp. 121'125, 2010.
 L.-J. Zhang and Q. Zhou, 'CCOA: Cloud computing open architecture,' in Proc. IEEE 7th International Conference on Web Services (ICWS'09), Los Angeles, CA, USA, Jul. 2009, pp. 607'616.
Design and Analysis
DSL Based Design and development process
Middleware for design phase and development phase
Semantic Meta data models