Rule-based Cloud Service Localisation

Luke Collins

and Frank Fowley

and Claus Pahl

School of Computing, Dublin City University, Dublin 9, Ireland

CNGL, School of Computing, Dublin City University, Dublin 9, Ireland

Keywords:

Cloud Service Localisation, Cloud Computing, Cloud Architecture.

Abstract:

The fundamental purpose of cloud computing is the ability to quickly provide software and hardware resources

to global users. The main aim of cloud service localisation is to provide a method for facilitating the interna-

tionalisation and localisation of cloud services by allowing them to be adapted to different locales. We address

lingual localisation by providing a service translation using the latest web-services technology to adapt ser-

vices to different languages and currency conversion by using realtime data provided by the European Central

Bank. Units and Regulatory Localisations are performed by a conversion mapping, which we have generated

for a subset of locales. The aim is to provide a standardised view on the localisation of services by using

runtime and middleware services to deploy a localisation implementation.

1 INTRODUCTION

Distributed web services can provide business and

private consumers with computing abilities which

may not be feasible for them to develop in-house

(Armbrust et al., 2010; Voorsluys et al., 2011). How-

ever, cloud computing introduces new issues, for ex-

ample in Europe, where there is a wide range of lan-

guages spoken, services are very often only developed

for single language. Often, companies do not have the

resources to develop multilingual products. Locali-

sation encapsulates a large number of issues which

need to be addressed. These include Language Trans-

lation, Regulatory Compliance, Currency Conversion

or Units Conversion. Localisation is typically per-

formed on textual content (i.e. strings) and refers to

either languages only or physical location. However,

the purpose of this work is to develop a method of lo-

calising services by implementing a ’mediator’ type

service which interacts between the APIs of the cloud

service provider and the requester. We are going to

focus on a number of locale dimensions such as lan-

guage, taxation, currency and units. We aim to pro-

vide service-level language translation techniques to

localise services (including API interfaces) to differ-

ent languages. Regulatory translation which includes

currency, units and taxation among other legal gov-

ernance and compliance rules will be provided by

standards-based mappings. Regulatory translation is

important for applications to comply with varying re-

gional laws and regulations.

A speciﬁc need is to make localisation techniques

available at runtime for dynamic localisation, which is

required for instance for currencies and other variable

aspects. Thus, Cloud Service Localisation (CSL) pro-

vides a mechanism for converting and adapting ser-

vices to the locale of the requester. A Localisation

Provider act as an intermediary between the service

provider and the requester. In our proposed platform,

this is supported by a mediation service. By generat-

ing a common platform solution for these localisation

issues, we allow the ability to dynamically localise

cloud services to be made with little effort. The key

contributions of our work are:

• Software Localisation at Service Level - the main

concern is a standardised mapping within a poten-

tially heterogeneous environment.

• Adaptation and Integration - the main concern

is the maintenance of service quality after it has

been localised through adaptation.

2 SERVICE LOCALISATION

Our focus is a platform for cloud service localisation,

which makes a shift from a ”one size ﬁts all” scenario

towards an end-to-end personalised service scenario.

Currently, cloud computing services suffer from lo-

calisation and adaptability issues for multiple users in

220

Collins L., Fowley F. and Pahl C..

Rule-based Cloud Service Localisation.

DOI: 10.5220/0004366802200223

In Proceedings of the 3rd International Conference on Cloud Computing and Services Science (CLOSER-2013), pages 220-223

ISBN: 978-989-8565-52-5

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

different regions. These issues could be overcome if

a multi-lingual and multi-regional solution was devel-

oped (Pahl, 2012; Wang et al., 2010).

A scenario which illustrates the beneﬁts of locali-

sation could be a cloud service provider that manages

company accounts for its customers, with ofﬁces in

different global locations.

• Regulatory: Conversion of data between stan-

dards and their variants, e.g. based on different

units of measurement Metric → Imperial.

• Currency: Conversion of between currencies, e.g.

Euro → Dollar.

• Lingual: Translation of service related data be-

tween languages. This could include free text, but

also speciﬁc vocabularies based on business prod-

uct and process standards such as GS1 or EAN-

COM for documents.

• Taxation: Different customers have different taxa-

tion requirements, e.g. different VAT rates. Local-

isation of the accounts software should take this

into account for each locale.

3 PLATFORM ARCHITECTURE

Localisation of cloud services requires an architecture

framework (Pahl et al., 2007) to be implemented to

facilitate various localisation methods. These various

methods, implemented as services in our proposed lo-

calisation platform, are used to facilitate the localisa-

tion of localisable elements or artefacts. This paper

focuses on the dynamic localisation of service level

descriptions.

With every cloud service there are various el-

ements which may be localised: Service speciﬁ-

cations/descriptions (APIs), Models (structural/be-

havioural), Documentation (for human consumption),

and Messages exchanged between services. Services

are normally written to be independent of locales,

however localisation is often needed. A localisation

platform should be based on attributes which vary

from locale to locale, like time or date format.

A service localisation platform requires a number

of elements. These elements can be pre-translated

fragments in static form or can be dynamic transla-

tion systems. Figure 1 aims to demonstrate the con-

cept of a policy and mappings based system, which

can be scaled when additional processes are attached

to the mediation process. In the platform architec-

ture, user-speciﬁc locale policies are applied to ser-

vice endpoints. For example, in a WSDL ﬁle we

may localise messages and operation names. Rules

for each type of translation would be stored in a rules

Figure 1: Architecture of the Localisation Platform.

database (General Rules Repository). Similarly, map-

pings between common translations would be stored

in a mappings database (Translation Memory).

A mediator operates between users (with different

locales) and several service providers (with different

locales) by providing core localisation services, such

as currency conversion and language translation. The

architecture supports the following:

• Static Mappings: these could be the mapping of

one language to another or one unit to another,

pre-translated in translation memories.

• Dynamic Localisation: when translation map-

pings are not stored, then a dynamic localisation is

required in order to obtain the correct translation

and store the mapping.

• Policy Conﬁguration: in order to conﬁgure the

various locale policies, we must generate partic-

ular translation rules, supported by a logical rea-

soning component.

• Negotiation: this is the exchange of locale policies

through the form of XML and SOAP from a web

services point of view.

• Localisation of Services: the mappings between

the remote service and the localised service de-

scription must be stored in a mappings database

(Translation Memory) so the localised service has

a direct relationship with the remote service.

A sample locale setup is

</ R e q u e s t e r L o c a l e >

</ P r o v i d e r L o c a l e >

Rule-basedCloudServiceLocalisation

221

</ L oc a l e s>

</CSLContext>

A mismatch between provided and required locales

needs to be bridged by a mediator service.

In order to provide the localisation platform ser-

vices, we need to implement a number of classes to

enable a modular service localisation platform. Their

interaction is summarised below.

4 LOCALISATION RULES

At the core of our service localisation platform is

a language to specify the rules in relation to local-

isations. In most cases, languages like WSDL and

other XML languages provide information regarding

the services that are provided via an API. However,

in order to encapsulate localisation information, there

is a necessity to provide a language which will con-

tain details in relation to the locales of the requester

and the provider. For the purpose of our localisation

platform, we use a policy language based on the Se-

mantic Web Rule Language SWRL, which is based

on the propositional calculus.

A localisation layer is used to encapsulate the var-

ious forms of translations. It is used to describe the

relationships between localisable elements. For ex-

ample, it contains the details of items which can be

translated. For our localisation model these are: Doc-

umentation & Descriptions and API Messages & Op-

erations. The rule language is used to deﬁne policies

of two types: ﬁrstly, locale deﬁnitions and, secondly,

conversion (translation) rules. We motivate the rule

set through selected examples. There are a number of

locale deﬁnition rules provided, like Loc or hasCur,

by which locales for speciﬁc regions are described.

A locale can also be described by other rules such as

hasTax, hasLang and hasUnit:

IELoc(?l) ←

Loc(?l) ∧

hasLang(?l, ?z) ∧ hasCur(?l, ?c)∧ hasU nit(?l, ?u) ∧

?z = en∧ ?c = EU R∧ ?u = metric

The beneﬁt of a formal framework for the rules is that

other rules can be inferred by from partial informa-

tion. For example, if we knew that a locale had USD

as its currency we may be able to infer its country

from it: ?c = USD →?l = USLocale. These inferred

rules do not apply in general - this may not work if we

know a currency is Euro in which case it could be one

of many locales in Europe. The purpose of these rules

could be to determine inconsistencies, however. Pre-

conditions can clarify the remit of these rules. A con-

version between locales, e.g. Locale A → Locale B, is

given by the following general conversion rule:

IELoc2USLoc(?l1, ?l2) ←

hasLang(?l1, ?z1) ∧ hasLang(?l2, ?z2) ∧

hasCur(?l1, ?c1) ∧ hasCur(?l2, ?c2) ∧

hasU nit(?l1, ?u1) ∧ hasU nit(?l2, ?u2) ∧

?z2 = convertLang(en, en, ?z1) ∧

?c2 = convertCur(EUR, U SD, ?c1) ∧

?u2 = convertCur(metric, imperial, ?u1)

5 EVALUATION

Scalability becomes more important from a cloud

computing perspective when a service may have large

multiples of users. Scalability has not been empir-

ically addressed for this phase of research and will

be evaluated in later prototypes. Many aspects of the

platform would require modiﬁcation to effectively al-

low the infrastructure to vertically scale-up or scale-

out efﬁciently. However, horizontal scalability - i.e.

the addition of more localisation concerns - is concep-

tually easily supported by the modular mediator ar-

chitecture, which we will address further below in the

extensibility context from an implementation view.

Performance is one of the CSFs of this platform.

Poor performance often tends to affect software expo-

nentially as multiples of users consume a cloud ser-

vice at the same time. As the application scales up to

multiple concurrent users, the latency would increase

due to extra loads placed on the platforms services.

This makes latency one of the KPIs of the project.

Latency, is also an area to be assessed as adding the

localisation platform to the workﬂow of an existing

process has the potential to add lag-time. This lag-

time exists due to time required to compute and also

the time to initialise the various variables. As a strat-

egy, we have aimed to improve performance by using

pre-translated aspects (through stored mappings).

CLOSER2013-3rdInternationalConferenceonCloudComputingandServicesScience

222

6 RELATED WORK

Our platform addresses the need for dynamic localisa-

tion of various artefacts by use of a translation mem-

ory and a set of logical rules.

• Software Localisation: this usually refers to hu-

man consumption of data which are produced by

the software - namely messages and dialogues.

Our focus is on the localisation of the service

level. Service internationalisation is supported

by the W3C Service Internationalisation activity

(W3C, 2005; Phillips, 2005).

• Adaptation and Integration: it is not uncommon

for software adaptation at service level, but the

adaptation of services based on locales and using

a translation memory with rules and mappings is

new (Truong and Dustdar, 2009). The problem of

multi-tenancy is a widespread issue in the area of

cloud computing (Wang et al., 2009; Wang et al.,

2010). This is an area where a lot of research is

being invested in order to provide a platform for

different users with different business needs to be

kept separate and their data to be kept private.

• Semantics: involves the matching of services with

various locales using mappings and rule-based

system (Bandara et al., 2009; Anastasiou, 2011;

Fujii and Suda, 2009).

IBM (IBM, 2010) presents a localisation solution for

its WebSphere platform. It uses static localisation as

it requires the WSDL ﬁles to be generated prior to de-

ployment. This differs from our proposed localisation

platform as our solution aims to perform transforma-

tions between locales dynamically.

7 CONCLUSIONS

A Cloud Service Localisation (CSL) implementation

should allow for a seamless and transparent solu-

tion by automatically adjusting and adapting services

to the requesters’ own locales. We have presented

a modular CSL implementation which can enable

cloud services to be introduced into emerging markets

which have localisation issues. Localisation hence

provides a mechanism to widen a service provider’s

target market by enabling multi-locale solutions. The

easiest solution is for a cloud service provider to pro-

vide a ’mediator’ service which could act as middle-

ware between a requester and the service provider. By

allowing services to be localised, we are enabling the

provision of multi-locale cloud services to create in-

teroperable clouds. The localisation of services also

allows the geographic scope for service providers to

broaden and users to have the ability to combine ser-

vices from different providers from different regions.

These clouds can be heterogeneous in nature. Thus,

the platform described in this paper must support that

ability. Due to the nature of third-party services, it is

more intuitive for service localisation to be performed

dynamically through the use of a mediator service.

REFERENCES

Anastasiou, D. (2011). The impact of localisation on se-

mantic web standards. European Journal of ePractice,

12:42–52.

Armbrust, M., Fox, A., Grifﬁth, R., Joseph, A., Katz, R.,

Konwinski, A., Lee, G., Patterson, D., Rabkin, A.,

Stoica, I., et al. (2010). A view of cloud computing.

Communications of the ACM, 53(4):50–58.

Bandara, K., Wang, M., and Pahl, C. (2009). Dynamic in-

tegration of context model constraints in web service

processes. International Software Engineering Con-

ference SE’2009. IASTED.

Fingar, P. (2009). Cloud computing and the promise of

on-demand business inovation. InformationWeek, July

13,2009.

Fujii, K. and Suda, T. (2009). Semantics-based context-

aware dynamic service composition. ACM Transac-

tions on Autonomous and Adaptive Systems (TAAS),

4(2):12.

IBM (2010). IBM WebSphere Developer Technical Journal:

Developing internationalized Web services with Web-

Sphere Business Integration Server Foundation V5.1.

Pahl, Giesecke, S. and Hasselbring, W. (2007). An

Ontology-based Approach for Modelling Architec-

tural Styles. European Conference on Software Ar-

chitecture ECSA2007. Springer.

Pahl, C. (2012). Cloud Service Localisation. European

Conference on Service-Oriented and Cloud Comput-

ing ESOCC 2012. Springer.

Phillips, A. (2005). Web Services and Internationalization.

Whitepaper.

Truong, H. and Dustdar, S. (2009). A survey on context-

aware web service systems. International Journal of

Web Information Systems, 5(1):5–31.

Voorsluys, W., Broberg, J., and Buyya, R. (2011). Cloud

Computing: Principles and Paradigms. John Wiley

and Sons.

W3C (2005). Web Services Internationalization Usage Sce-

narios.

Wang, M.X., Bandara, K.Y. and Pahl, C. (2009). Inte-

grated constraint violation handling for dynamic ser-

vice composition. IEEE Intl Conf on Services Com-

puting. pp. 168-175.

Wang, M.X., Bandara, K.Y., and Pahl, C. (2010). Process

as a service distributed multi-tenant policy-based pro-

cess runtime governance. International Conference on

Services Computing (SCC), pages 578–585. IEEE.

Rule-basedCloudServiceLocalisation

223