ENABLING SEMANTIC RESOURCES IN THE CLOUD

Salvatore F. Pileggi, Gema Ibañez, Carlos Fernandez-Llatas and Juan Carlos Narajo-Martinez

ITACA-TSB, Universidad Politécnica de Valencia, Valencia, Spain

Keywords: Cloud Computing, Semantic technologies, Virtual Organization, Virtualization.

Abstract: The massive migration to cloud solution appears next to be a fact. Cloud infrastructures could assure a

competitive, scalable and sustainable environment for services and applications. However, next generation

applications have to be able to pervasively meet the needs and requirements deeply different among them.

Applications involving complex virtual organizations could require a higher level of flexibility that could

result by the convergence of migration and virtualization. In this paper, a resource-centric model, is

proposed: file systems, DBs, services and any other class of resources are available in the "cloud" as Virtual

Resources regardless the infrastructures on which they are deployed. Semantics play a critical role in order

to assure advanced and open solutions in a technologic context featured by a fundamental lack of

standardization.

1 INTRODUCTION

The popularity of Cloud technologies is constantly

increasing as well as the interest on this emerging

market from specialized companies that are already

able to offer several solutions based on different

models such as IaaS, PaaS and SaaS (Shuai, 2010).

Even if there are several concerns from involved

customers and stakeholders and open issues (e.g.

privacy (Esteves, 2010), security (Ramgovind, 2010)

and standardization (Ortiz, 2011)) concerning the

massive migration to the cloud, both private and

enterprise cloud solutions are unanimously

considered the "future" of the computation (Kshetri,

2010).

In practice, the cloud approach is actually

referred as the most competitive, scalable and

sustainable solution on the market under the always

more realistic conditions of constantly decreasing

bandwidth price and of always connected users. An

exhaustive analysis of technical (Shuai, 2010) and

business (Chang, 2010) aspects of cloud solutions is

out of paper scope.

This business scenario could quickly change in

the next future if the cloud will not provide high

flexible solutions able to meet the needs of complex

Virtual Organizations (VOs). Analyzing VOs, it is

evident that not all resources normally involved are

suitable to the migration to the Cloud. The level of

flexibility of cloud solutions could be further

increased if the platforms are the results of the

convergence between migration and virtualization.

The key idea is that resources deployed using cloud

infrastructures (migrated) and resources deployed

according to conventional solution could be merged

in a unique virtual environment.

This paper proposes a resource-centric model for

cloud infrastructure in which virtual resources are

provided with a semantic description/specification

able to assure a potential high level of

interoperability among platforms as well as a set of

facilities for the integration of pre-existent or new

resources. As detailed in the following sections,

semantics play a critical role in the proposed model

especially considering the fundamental lack of

standardization the cloud is experimenting (Ortiz,

2011).

As any cloud model, it is potentially independent

from any application domain even if, inevitably,

domain-specific semantic representations could be

required in certain contexts/applications.

A potential application domain for the model is

the Spanish health system. In Spain, National Health

System follows a decentralized model where each

autonomous community manages all the centres,

services and establishments of the Community

Councils, Town Councils and any other intra-

regional governments. An autonomous community

is the first-level political division of the Kingdom of

Spain, established in accordance with the current

Spanish Constitution. Each autonomous community

has many hospital systems of very different natures,

541

F. Pileggi S., Ibañez G., Fernandez-Llatas C. and Carlos Narajo-Martinez J..

ENABLING SEMANTIC RESOURCES IN THE CLOUD.

DOI: 10.5220/0003885405410546

In Proceedings of the 4th International Conference on Agents and Artiﬁcial Intelligence (IWSI-2012), pages 541-546

ISBN: 978-989-8425-95-9

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

being independent of each other.

When a patient must go to different health

centres should generally answer the same questions

in order to open his/her medical records in each of

them. This generates an overall lack of coordination

regarding the interoperability between systems,

redundant data, basic services, etc.

There is a current trend, which aims to unify

those different systems to autonomous community

level. At the moment, interconnection between all

hospital systems is unapproachable, being applied to

European context.

With this purpose, the proposed model aims to

provide interoperability through cloud-based

environments, giving support and tools for the

virtualization, migration of already existing systems

formed by data, apps and/or infrastructure and

creation of new generation services, in order to make

available basic and composite services to consumers

(patients, services,…) and for development and

management of customized e-health services to all

users.

The core part of the paper is structured in 3 main

sections: the section 2 is the natural extension of the

introduction (it provides a detailed description of

approach and scope); in the section 3 the model is

deeply described; finally, in the section 4 a brief

analysis to next generation resources is provided.

2 APPROACH AND SCOPE

The scope of enabling semantic resources in the

cloud is the interconnection of heterogeneous

systems providing the coexistence of different kinds

of heterogeneous resources (Figure 1):

• Internal resources: they are the result of the

common migration process to the cloud.

Resources are hosted by internal infrastructures

and they are available as virtual resource in the

platform.

• External resources: they are hosted by external

infrastructures but they are pervasively

available into the platform as virtual resources

(Virtualization).

• Next generation resources: can be designed and

implemented directly over the virtualized layer

provided by the platform.

The platform model assures a flexible and shared

infrastructure mainly featured by the following

points:

• It enables ecosystems among heterogeneous

systems.



Figure 1: Interconnecting resources in the Cloud.

• It provides a pervasive virtual environment:

resources are managed at virtual levels assuring

high-interoperable capabilities.

• Cloud approach: scalable, competitive and

sustainable solutions.

• Open semantic support: the core semantic

support can be integrated/extended with

domain-specific interoperability capabilities

(e.g. health/medical).

• Cross-domain platform: the core infrastructure

of the platform is not domain-specific; the

potential application range increases with the

expressivity of the semantics.

The diagram represented in the figure 2 shows

the platform conceptualization. The free software

IHMC CmapTools (IHMC, 2010) is adopted for

specifying this compact model as a concept map.

IHMC CmapTools proposes an approach for

knowledge representation similar to semantic

networks that build semantic relations among

concepts through a directed or undirected graph

consisting of vertices, which represent concepts, and

edges.

As showed, pre-existing VOs join the common

cloud ecosystem enabled by the platform. Each VO

provides a set of resources. These resources can be

public (available in the ecosystem for any other

VO), protected (available just for authorized actors)

or private (available only inside of owner VO or

under payment). The platform is able to manage any

kind of resource regardless by the infrastructures on

which they are deployed. This interoperable layer is

assured by the semantic representation of resource.

2.1 A Practical Use Case: Health

Systems Interconnection

Several medical systems (e.g. ORION and IANUS)

ICAART 2012 - International Conference on Agents and Artificial Intelligence

542

Figure 2: Platform conceptualization.

are currently coexisting in Spain providing similar

data and services. They are solutions developed in

different times in order to meet different

requirements and needs from different end-users

(e.g. hospitals) that are progressively converging.

Integrated services as well as the need of a

stronger level of collaboration require a high level of

interoperability. This requirement is completely

missed in current systems that are designed to work

according to stand-alone behaviours.

The interoperability among health systems is

normally solved through ad-hoc solutions (e.g. an

interoperable layer between ORION and IANUS).

This class of solution is expensive and has several

limitations in terms of flexibility because it does not

solve a generic problem but just a concrete/local

problem. Details about will be provided in the

section 3.1.

3 PLATFORM MODEL

The convergence between cloud-based solution and

virtualization is not an absolute novelty. An

exhaustive analysis (Dong, 2010 and Gmach, 2011)

is out of paper scope but, recently, this approach was

used in order to reach different goals in the context

of different domains.

The last generation of virtual platforms

(Siddhisena, 2011), virtual organizations (Li, 2010),

virtual services (Fu, 2010), techniques for virtual

resources management (Hien, 2009)/ optimization

(Khatua, 2010) and virtual infrastructure (Keller,

2010) are, implicitly or explicitly, referred to the

cloud.

Also a semantic specification of resources is a

well known topic for both general purpose and

specific (e.g. industrial resources (Khriyenko, 2004))

purpose.

The proposed model is composed of three

converging perspectives:

• Interoperability model (Section 3.1)

• Business model (Section 3.2)

• Technical perspective (Section 3.3)

3.1 Interoperability Model: Vertical

Approach

One of the critical and key issues for the

improvement of the current interoperability model is

the conceptual evolving from a “horizontal” to a

“vertical” approach.

Actually medical systems are logically part of

virtual organizations characterized by different

complexity in terms of structure and distribution.

Each virtual organization has its own technological

environment.

During the last few years, a progressive

convergence among these environments was aimed

(see introduction). The problem is normally

approached trying to provide added (or improved)

capabilities among existent systems (Figure 3). This

“direct” solution is effective and efficient but it is

just a local solution: the “integration” of a new

system (or resource) implies the need of a “new”

solution.

Furthermore, if a new system/resource is

integrated in the ecosystem and it has to be

interoperable with the existent ones, an ad-hoc

component (proxy) that assures the interoperability

has to be provided for interfacing each existent

system/resource.

This last situation is expressed by (1) where n is

the number of independent systems/resources and k

is the number of proxies. As showed, if a new

resource or system is integrated, a full-interoperable

solution implies the deployment of O(n) proxies.

k(i) = k(i-1) + n(i-1) → O(n-1) = O(n)

n>2

(1)

The virtualization of resource enables a model of

interoperability based on a vertical approach (Figure

3): resources are available at virtual level and the

interoperability among systems is approached at this

abstracted level.

A solution based on the vertical approach for the

interoperability is intrinsically simplest because the

integration of a new system/resource just implies the

interfacing with the abstract layer (2).

k(i) = k(i-1) + 1→ O(1)

n>2

(2)

ENABLING SEMANTIC RESOURCES IN THE CLOUD

543

Figure 3: Horizontal vs vertical approach for

interoperability.

In practice the vertical approach can be assured

according to two different approaches:

1) Standards. Solutions based on shared layers that

impose resources to be described according to a

well defined set of standards. This is the

simplest solution by a technical and conceptual

point of view but, unfortunately, is not always

easy to be applied in the context of complex

virtual organization. This is mainly because

standards are hard to be imposed and the

data/knowledge from systems can be hard to be

converged in well defined standards even

considering domain-specific solutions.

2) Open models. Dynamic solutions based on open

models (e.g. semantics). Semantic

representations guarantee the definition of local

knowledge environments that can be centrally

managed without the need to share standards.

Furthermore, resources could interact among

them interchanging semantic data. A completely

open model is complex to be proposed and

managed. As explained in the section 4, realistic

solutions can be designed according to hybrid

approaches based on core ontologies that can be

extended and/or particularized.

3.2 Business Model: Merging

Migration and Virtualization

Merging migrated and virtualized resources provides

a high level of flexibility respect to both technologic

and business perspectives. Migration assures a

scalable environment, the potential reduction of

maintenance costs, as well as the other vantages

typical of the cloud approach.

Virtualization allows a further degree of

flexibility for resources that owners cannot migrate

or do not want to migrate: resources are available

into the ecosystems but they are not migrated to the

cloud. Motivations for preferring a virtualized

resource can be related to law restrictions, business

constraints or any other real situation that does not

match a full cloud approach.

3.3 Technical Overview: Functional

and Semantic Support

Platform designed according to the proposed

approach should include at least a set of core

functionalities, as well as a full semantic support. A

reference model can be logically structured as in the

follow (Figure 4):

• CORE Infrastructure/Functional support:

infrastructures and any other functional support

to the ecosystem.

• Management support: set of functionalities for

the management of the platform.

• Semantic support: models for knowledge

representation.

Figure 4: Technical overview at the platform.

Furthermore, in order to assure a realistic

exploitation plan a support for developers is

required: a set of APIs and interfaces that support

the developer to migrate and virtualize existing

ICAART 2012 - International Conference on Agents and Artificial Intelligence

544

resources as well as the design and implementation

of new resources.

The platform is evidently the result of the

convergence among cloud technologies,

virtualization techniques and semantics.

4 NEXT GENERATION

RESOURCES

A pervasive virtual environment designed according

to a cloud approach can provide a solid support for

the development of a new generation of resources

(e.g. services and applications).

These resources can be developed directly on the

top of the virtual layer provided by platforms

supporting a high level of abstraction (e.g. role-

driven development).

As introduced in the section 3.1, the key issue is

the efficient and effective application of open

models for the knowledge specification and

representation.

The use of “standard” ontologies could be the

most immediate solution: rich data models could be

enough expressive to represent the knowledge as

well as to assure inferred knowledge and an

interesting set of interoperability capabilities. But it

could limit the advantages and benefits provided by

open solutions as well as the problems related to the

knowledge convergence could not be solved or

skipped.

On the other hand, a completely open model that

assumes each local system/resource described

according its own ontologies could be hard to be

applied in real systems. Typical problems in multi-

ontology computation (e.g. correctness and

ambiguities) both with the objective difficulty to

provide a centralized management for resources

advise more realistic approaches.

The current idea is the use of shared

vocabularies. These vocabularies should provide the

basic concepts making possible the definition of

independent local knowledge environments that can

be globally linked and processed. In practice, shared

concepts have to be used in order to link local

ontologies to the platform. Further concepts, as well

as rules and relations among them, can be provided

by local ontologies. This approach is equivalent to

object extension in object-oriented environments.

At the moment of designing a new resource,

developers could have a full functional support

provided by the platform and a dynamic semantic

support. The developer can choose the deployment

model (migration or virtualization) that better

matches the business needs, link the resource to the

platforms through concepts from the shared

vocabulary and make available the knowledge

required (local ontologies). Further advantages are

provided at the moment to design resources that

assume the coordinated/uncoordinated use of other

resources that can be directly managed at high level.

5 CONCLUSIONS

The convergence between cloud and virtualized

solutions in a semantic context provides improved

interoperability capabilities as well as a competitive

environment for resources integration.

The flexibility assured by open models for the

knowledge definition and representation could play

a key role in several concrete environments (e.g.

Spanish health system) involving complex virtual

organizations.

The power of integrating existent resources (as

well as the design of new ones) directly on the top of

an abstracted layer provides a new vision at the

cloud and its exploitation model.

Finally, a semantic layer able to link resources to

the global environment (platform) and to support, at

the same time, local knowledge representations

could provide a dynamic support for the effective

convergence of dynamic resources in the cloud.

REFERENCES

Shuai, 2010. Shuai Zhang; Shufen Zhang; Xuebin Chen;

Xiuzhen Huo, "Cloud Computing Research and

Development Trend," Second International

Conference on Future Networks, 2010. ICFN '10, 22-

24 Jan. 2010.

Esteves, 2010. Esteves, R. M.; Chunming Rong, "Social

Impact of Privacy in Cloud Computing," 2010 IEEE

Second International Conference on Cloud Computing

Technology and Science (CloudCom), Nov. 30 2010-

Dec. 3 2010.

Ramgovind, 2010. Ramgovind, S.; Eloff, M. M.; Smith,

E.,"The management of security in Cloud computing,"

Information Security for South Africa (ISSA), 2010,

vol., no., pp.1-7, 2-4 Aug. 2010.

Kshetri, 2010. Kshetri, N., "Cloud Computing in

Developing Economies," Computer, vol.43, no.10,

pp.47-55, Oct. 2010.

Ortiz, 2011. S. Ortiz Jr., The Problem with Cloud-

Computing Standardization, Computer, July 2011,

pp.13-16.

Chang, 2010. Chang, Victor; Bacigalupo, David; Wills,

Gary; De Roure, David, "A Categorisation of Cloud

Computing Business Models," 2010 10th IEEE/ACM

ENABLING SEMANTIC RESOURCES IN THE CLOUD

545

International Conference on Cluster, Cloud and Grid

Computing (CCGrid), 17-20 May 2010.

IHMC, 2010. IHMC CmapTools [online at: http://cmap.

ihmc.us/], accessed on: 24.09.2010.

Dong, 2010. Hanfei Dong; Qinfen Hao; Tiegang Zhang;

Bing Zhang, "Formal Discussion on Relationship

between Virtualization and Cloud Computing," 2010

International Conference on Parallel and Distributed

Computing, Applications and Technologies (PDCAT).

Siddhisena, 2011. Siddhisena, B.; Warusawithana, L.;

Mendis, M., "Next generation multi-tenant

virtualization cloud computing platform," 2011 13th

International Conference on Advanced

Communication Technology (ICACT).

Li, 2010. Jianxin Li; Bo Li; Zongxia Du; Linlin Meng,

"CloudVO: Building a Secure Virtual Organization for

Multiple Clouds Collaboration," 2010 11th ACIS

International Conference on Software Engineering

Artificial Intelligence Networking and

Parallel/Distributed Computing (SNPD).

Fu, 2010. Jicheng Fu; Wei Hao; Tu, M.; Biao Ma;

Baldwin, J.; Bastani, F. B.; Virtual Services in Cloud

Computing, 2010 6th World Congress on Services

(SERVICES-1).

Hien, 2009. Hien Nguyen Van; Tran, F. D.; Menaud, J.-

M., "SLA-Aware Virtual Resource Management for

Cloud Infrastructures,"Ninth IEEE International

Conference on Computer and Information

Technology, 2009. CIT '09.

Khatua, 2010. Khatua, S.; Ghosh, A.; Mukherjee, N.,

"Optimizing the utilization of virtual resources in

Cloud environment," 2010 IEEE International

Conference on Virtual Environments Human-

Computer Interfaces and Measurement Systems

(VECIMS).

Keller, 2010. Eric Keller, Jakub Szefer, Jennifer Rexford,

Ruby B. Lee, NoHype: Virtualized Cloud

Infrastructure without the Virtualization, Proceedings

of the 37th annual international symposium on

Computer architecture, ISCA '10.

Khriyenko, 2004. Khriyenko, O.; Terziyan, V.,

"OntoSmartResource: an industrial resource

generation in semantic Web," 2004 2nd IEEE

International Conference on Industrial Informatics,

2004. INDIN '04, 26-26 June 2004.

Gmach, 2011. Gmach, D.; Rolia, J.; Cherkasova, L.,

"Resource and virtualization costs up in the cloud:

Models and design choices," 2011 IEEE/IFIP 41st

International Conference on Dependable Systems &

Networks (DSN).

ICAART 2012 - International Conference on Agents and Artificial Intelligence

546