THE VALIDATION OF A DYNAMIC SERVICE COMPOSITION

MODEL USING A SIMPLE GAME

Abhishek Srivastava and Paul G. Sorenson

Department of Computing Science, University of Alberta, Edmonton, Canada

Keywords:

Service composition, Dynamic systems, Validation, Simple game.

Abstract:

Service composition is fast becoming a dynamic process where the selection of service elements and their

subsequent composition happen at run time. A number of models have been proposed that support dynamic

composition. The big issue lies in the validation of such models. A straightforward method of validating such

models is to put together a service composition comprising of actual service elements and testing the efﬁcacy

of the application. However, this is not always feasible because of the expenses involved in such an experiment

combined with the legal and privacy issues with the service providers.

In this paper, we propose a novel idea wherein the validation of a dynamic service composition model can be

done using a simple ‘Ambitious Traveler’ game. The game comprises the selection from a set of predeﬁned

transportation services using a dynamic composition model and separately by a group of volunteers who serve

as customers. The selections of the two are compared and whosoever makes a better selection gets a higher

score.

The game is simple and may be modiﬁed to validate dynamic systems of other kinds as well. The speciﬁc

approach and game described in this paper is currently under development.

1 INTRODUCTION

Service composition is the combination of simple ser-

vice elements to form larger, more complex com-

posite applications. The composite application so

formed could further be combined in a recursive man-

ner with other service elements or compositions to

form even larger applications. The service elements

that partake in such compositions are usually deter-

mined in advance, and the composition process com-

prises monitoring the ﬂow of data/messages between

these services. Such composition processes with pre-

determined service elements are called static compo-

sitions.

The bigger challenge is in putting together ser-

vice compositions where the selection of service el-

ements is done at the time of composition. Such

dynamic compositions deal with the issue of appro-

priate service selection based on the requirements

of the customer along with monitoring the ﬂow of

messages/data between the selected service elements.

The fact that the attributes of the service elements

are varying as are the requirements of different cus-

tomers, makes this task even more complex.

A number of service composition models have

been proposed to tackle this issue (Silva et al.,

2009)(Colombo et al., 2006). These models typically

arrive at selection decisions based on a logical assess-

ment of the customer requirements and their confor-

mance with the attributes of the potential service el-

ements. Our research also comprises a service com-

position model. In our approach, we mathematically

combine multiple Quality of Service (QoS) attributes

of the potential service elements based on the im-

portance placed on the respective QoS attributes by

the customer. Further, we also incorporate the inter-

actions between the service elements in the domain.

All this together gives us a factor called Afﬁnity upon

which dynamic service selections are made. Details

of our technique are given in Section 3 of this paper.

This brings us to the main topic of this paper

which is the validation of such a composition model.

An intuitive approach to this would be to utilize the

model to dynamically put together an actual service

composition. This however is not always practicable,

especially in an academic set-up, given privacy and

legal issues that govern the usage of commercially

available service elements.

We therefore synthesized a game which we call

the Ambitious Traveler, that involves putting together

260

Srivastava A. and G. Sorenson P..

THE VALIDATION OF A DYNAMIC SERVICE COMPOSITION MODEL USING A SIMPLE GAME.

DOI: 10.5220/0003143402600266

In Proceedings of the International Conference on Knowledge Management and Information Sharing (KMIS-2010), pages 260-266

ISBN: 978-989-8425-30-0

 2010 SCITEPRESS (Science and Technology Publications, Lda.)

individual service elements at run time just like the

dynamic service composition scenario. Furthermore,

the game assigns ‘scores’ to players on the basis of

their performance. A group of volunteers are re-

quested to play the game and their respective scores

are noted. Subsequently, the game is played with

the service selection decisions being made using our

composition model. The score of our model is com-

pared with those of the volunteers. If the model con-

sistently outperforms the players, the model’s efﬁcacy

is validated.

The remainder of this paper is organized as fol-

lows: Section 2 is a discussion on other techniques of

validation for dynamic service composition systems.

Section 3 is a description of our ‘Afﬁnity’ model for

dynamic service composition. This is the model that

needs to be validated. In Section 4, we describe the

Ambitious Traveler game which will serve as a means

to validate our model and which can be modiﬁed to

validate other dynamic systems as well. Section 5

discusses possible improvements and enhancements

which we envisage in the game in the future. Finally,

Section 6 concludes the paper.

2 RELATED WORK

In this section we discuss a few strategies for valida-

tion of dynamic systems that have been adopted by

researchers in the past.

Colombo et al. (Colombo et al., 2006) tackle the

issue of dynamic composition where the service el-

ements do not always behave along expected lines.

They provide an extension to the BPEL language in

the form of the ‘SCENE platform’ which addresses

this issue. The proposed platform is validated form-

ing an application using a set of real services and ob-

serving the behaviour of the application.

Silva et al. (Silva et al., 2009) propose the Dy-

namiCos framework which incorporates the require-

ments of different customers to dynamically put to-

gether personalized services. To validate the proposed

framework they put together an extensive prototype of

the framework which enables services to be deployed

and be published in a UDDI-like registry.

In their paper Eid et al. (Eid et al., 2008) de-

scribe a set of benchmarks against which to assess

various frameworks of dynamic composition. The set

of benchmarks is comprehensive and is broadly cat-

egorized into three parts: input subsystem, compo-

sition subsystem, and execution subsystem. A com-

position model or framework that scores well against

these benchmarks is considered good.

Shen et al. (Shen et al., 2007) present the Web

Service, Role and Coordinator (WSRC) model to han-

dle dynamism in web service compositions. In this

model, the process of service composition is divided

into three layers: Service, Role, and Coordinator. To

validate the model, the authors describe a case-study

of a vehicle navigation system which comprises a

global positioning system and a trafﬁc control service.

These are a few of the validation techniques in use

for dynamic composition models and frameworks. Of

these techniques, the ones that are accurate are quite

complex like the prototype that the authors of Dynam-

iCos put together or the service composition put to-

gether out of real services for the SCENE platform.

On the other hand, the validation techniques that are

simple like the case-study in the WSRC model seem

ineffective.

The validation technique proposed in this paper,

the Ambitious Traveler game, on the other hand is

quite simple as it comprises an elementary game with

straightforward steps and instructions. Despite this, it

is quite precise and is able to quantitatively express

the performance of the dynamic composition model.

3 THE ‘AFFINITY’ MODEL FOR

SERVICE COMPOSITION

In this section we describe our ‘Afﬁnity’ model for

service composition. It is speciﬁcally for the vali-

dation of this composition model that the Ambitious

Traveler game (described in Section 4) is used. The

game can, however, be modiﬁed appropriately and be

used for the validation of other dynamic systems as

well.

3.1 Assumptions on the Service Domain

In our ‘Afﬁnity’ model for service composition, cer-

tain simplifying assumptions have been made about

the service domain from which the service elements

are selected. First, our approach assumes that the ser-

vice elements taking part in the composition process

are already available in the service domain. The ap-

proach therefore does not look into the issue of ‘dis-

covering’ the service elements. The reader interested

in the process of service discovery is pointed to (Wu,

2008) for an exhaustive discussion on the subject.

The potential service elements, which are as-

sumed to have been already discovered, are repre-

sented as arranged in a tiered fashion as shown in

Figure 1. Each tier or level corresponds to a speciﬁc

functionality of the composite application. All the

discovered service elements that correspond to the re-

spective functionality are placed at that level. The ar-

THE VALIDATION OF A DYNAMIC SERVICE COMPOSITION MODEL USING A SIMPLE GAME

261

Functionality Level 2

Functionality Level 3

Functionality Level 4

Functionality Level 5

Functionality Level 6

Functionality Level 1

Figure 1: The service domain.

rangement of the functionality levels from top to bot-

tom follows the order in which the functionalities are

invoked in the composite application. This means that

the functionality of a given level follows the function-

ality of the level immediately above it, and is followed

by the functionality of the level immediately below it,

in the composite application. An example of a Trip-

planner application is shown in Figure 2 to give an

idea of what a real world service domain might look

like (the service names in Figure 2 are ﬁctitious).

TravelPedia FlightsCheap Book-a-Flight

QuickTaxi TaxiOnline

EasyHotels BestHotels FastHotelBooking HoteloCity

SecureTransac PaymentGate EasyPay

Log

out

Figure 2: The service domain of the trip-planner applica-

tion.

We introduce a factor called ‘Coupling’ which

connects a service element at a given functionality

level in the domain to a service element at the next

lower functionality level as shown in Figure 3 . Cou-

pling is an expression of the ‘working compatibil-

ity’ that exists between service elements at adjacent

functionality levels and depends on high level factors

such as business relationships between the providers

of the respective service elements, as well as low level

technological compatibility between the potentially

interacting elements. Coupling values are assigned

through expert judgement based on a comprehensive

assessment of the service element and its relationship

with the service elements at the adjacent levels. A

Coupling value of 1 is assigned if no special relation-

ship exists between two service elements, and higher

values are assigned for stronger relationships (a Cou-

pling value of 0 is assigned if one service element

does not acknowledge the presence of the other).

Figure 3: The service domain with arcs representing Cou-

pling and the effective workﬂow.

3.2 Service Selection and Composition

In our approach the selection of service elements from

the functionally equivalent ones at each level of the

domain is done utilizing the Quality of Service (QoS)

attributes. The service element whose QoS attributes

conform most closely with the requirements of the

customer is the one selected.

The ﬁrst task is to enable the customers to express

their QoS requirements. Our approach requires the

customer to attach a ‘weight’ to each QoS factor de-

pending on the extent to which the factor is desired in

the composite application. For example, if one of the

QoS factors is reliability, and the customer requires a

highly reliable service, the parameter (say α) repre-

senting the weight of reliability needs to be assigned

a high value. This is akin to a ‘tuning knob’ mecha-

nism wherein the QoS requirements of the customer

are looked upon as a set of control knobs. One knob

is assigned to each factor. The customer may turn the

knobs depending upon the degree to which he/she de-

sires the QoS factors in the composite application.

The next task is to make the appropriate selections

of service elements based on the expressed QoS re-

quirements of the customer. To do this the value of

the QoS factors in each of the potential service ele-

ments needs to be calculated. Any number and type

of QoS factors may be considered. In our work, we

have chosen four QoS factors: Reliability, Waiting-

time, Reputation, and Cost.

Reliability (Rel.) is the ﬁrst QoS factor consid-

ered. In our work, the reliability of individual ser-

vice elements has been calculated taking into account

the inﬂuence that the presence of other service ele-

ments in the domain have on them (Srivastava and

KMIS 2010 - International Conference on Knowledge Management and Information Sharing

262

Sorenson, 2010a). Waiting-time (Wait-time) is the

next factor considered. The waiting-time for a ser-

vice element is the time that a request for the service

has to wait from the time of invoking the service ele-

ment to the time the service begins. The waiting-time

for a request at a service element is calculated as the

ratio of the length of the ‘request-queue’ to the av-

erage completion-rate of the service element (Srivas-

tava and Sorenson, 2010b). The Reputation (Reput.)

of a service element is considered to be a valuable at-

tribute and is the next factor that is incorporated in

the service selection process. In our work, the reputa-

tion of a service element is calculated on the basis of

feedback collected from customers who have used the

service element. Finally, the Cost of the service ele-

ment is utilized in making an appropriate selection.

The cost of a service element is calculated taking into

account the functional and non-functional attributes

of the element along with the inﬂuence of its environ-

ment.

The four QoS factors so calculated for each ser-

vice element along with the weights assigned by

the concerned customer to each QoS factor are used

to calculate a factor called ‘Afﬁnity’ between every

pair of potentially interacting service elements. The

A f f inity

i j

between a potential invoking service ele-

ment i and a potential invoked service element j is a

function of the values of the QoS factors calculated

for the invoked service and the value of coupling be-

tween the two services. It is shown in equation (1)

A f f inity

i j

= Coupling

i j

{Rel.

}

· {Reput.

}

{Wait-time

}

· {Cost

}

(1)

α, β, γ, and δ are the weights assigned respectively by

the customer to the QoS factors: Reliability, Waiting-

time, Reputation, and Cost. With the value of Afﬁnity

so calculated, the service elements for the composite

application are selected following a Greedy approach

(Cormen, 1990). The service element selected for

each functionality is the one with the maximum value

of Afﬁnity with the previously selected element.

4 THE ‘AMBITIOUS

TRAVELLER’ GAME

In this section, the ‘Ambitious Traveller’ game is de-

scribed which serves as a simple and inexpensive

means of validating our service composition model.

This is only a preliminary description of the game.

We expect the game to go through two or more evolu-

tions before a production-ready version is completed.

The emphasis here is on the idea and efﬁcacy of using

games like these to validate dynamic systems.

4.1 Description of the Game

The motive of a player in the Ambitious Traveller

game is to travel as much as possible around the globe

given a limited amount of money and limited time.

The winner is the player who manages to travel the

maximum distance. The structure of the game is de-

scribed in Figure 4. The player starts from an ‘origin’

and has to travel to the ﬁrst city on the list: Paris in

the ﬁgure. To travel to Paris, the player has the option

to choose between taking a ﬂight, a car, a train, or a

boat. Each of these means of transportation has its

pros and cons. For example, a ﬂight would save time

but would probably be expensive and possibly require

the player to wait longer before getting a reservation;

a train on the other hand would be much more afford-

able but would be much slower.

The Game ...

Origin

Paris

London

NYC

car ﬂight train boat

car

ﬂight

train

boat

x km

y km

z km

Figure 4: An overview of the Ambitious Traveler game.

On getting to Paris, the next destination for the

player is London. To get to London, again the player

has to take a decision on the means of transportation.

This is continued until the player either runs out of

money, or time. In either case, the distance travelled

by the player until then is noted. Every player is given

a large number of travel opportunities. On each op-

portunity, the player starts from the origin with a lim-

ited amount of money and time and attempts to travel

as far as possible. After the player is done with all his

travel opportunities, the average of the distances trav-

elled on all the opportunities is the ﬁnal score of the

player. The larger the ﬁnal score the better the player

has done in making his transportation selections.

The service domain in this game is the transporta-

tion service (ﬂight, car, train, and boat) between every

pair of travel points. These transportation services are

analogous to the service elements at each function-

ality level in the service domain of our composition

model described in the previous section. Just like in

the composition model where a service selection has

to be made at each functionality level to put together

THE VALIDATION OF A DYNAMIC SERVICE COMPOSITION MODEL USING A SIMPLE GAME

263

the composite application, in this game the player has

to make a selection between the available transporta-

tion services to travel further. Every transportation

service, like the service elements in the composition

model, has an associated reliability, a waiting-time, a

reputation, and an associated cost. The player has to

make a trade-off between one or more of these QoS

attributes every time he makes a transportation selec-

tion.

In the following portion, we describe how the

selection of one transportation service over another

based on its QoS attributes impacts the ﬁnal score (the

total distance traveled) of the player.

Reliability. Every transportation service in the

game has an associated ‘fail-rate’. The transporta-

tion services are randomly made to fail at rates pro-

portional to their respective fail-rates. The value of

the fail-rate of a transportation service is modiﬁed ev-

ery time the service fails. There are no implications

of a failure of a transportation service if the service is

currently not transporting any player. If however, the

transportation service is indeed transporting a player

when it fails, the player loses all the money invested

in invoking the service and also wastes time as he is

sent back to the previous level and has to once again

make a transportation selection for the same destina-

tion. Therefore, it is beneﬁcial for the player to make

a transportation choice which advertises a high degree

of reliability.

Waiting-time/Completion-time. Whenever a

player with any of his travel requests makes a

transportation choice, there is a possibility that the

transportation service is currently in use by another

request and the player has to wait. Waiting for a

large time period results in the player wasting time

which is a limited resource in the game and adversely

affects the score of the player. It is important there-

fore for the player to select a transportation service

incorporating the length of the waiting queue at the

service.

The other factor that inﬂuences the time resource of

the player is the travel time or the service completion-

time. A good choice here results in the player wasting

very little time in covering the distance between two

cities and thus has a favorable inﬂuence on his score.

A bad choice has the opposite effect. The service

time of the transportation should also therefore be

taken into account while making a service choice.

Reputation. Every player that uses a transportation

service in the game is expected to give his feedback

on the service. This feedback is usually a ﬁgure that

the player feels appropriate out of a maximum of 5.

The feedback of the players forms the basis of deriv-

ing a reputation value for a service. The reputation is

a quantitative expression of the overall ‘feel’ that the

player gets while using the service. Each service has

a reputation value which is the average of the feed-

back values assigned by all the players that have used

the service. The reputation value of a transportation

service serves as an indicator of the quality of previ-

ously experienced service to subsequent players while

making their service selection.

Cost. Cost is an overarching factor that has a direct

inﬂuence on the choice of the player. A cost is as-

sociated with each transportation service based on its

overall functional and non-functional attribute values

as well as on the environment in which the service

is being used. The cost is a dynamic factor and its

value keeps varying. A player who selects a trans-

portation service that has a high cost, risks wasting

money which is one of the two limited resources in

the game; however by choosing an expensive service

the same player could be gaining in terms of time, the

other limited resource. Therefore, there is a trade-off

which the player has to decide upon using his judge-

ment.

4.2 The Game as a Validation Tool

To use the Ambitious Traveler game described in the

previous sub-section as a validation tool for our ser-

vice composition model, the game will be played in

three different modes: (i) by volunteers making their

own service selection decisions at each stage of their

travel; (ii) by volunteers providing their quality factor

weightings and have our service composition model

make service selections; and (iii) by playing the game

making random service selections at each travel op-

portunity.

The scores of each of the three scenarios are noted

for a large number of travel requests. A high score

by our composition model when compared to the vol-

unteer(s) score or the random-selection strategy score

serves to validate the model. In the following sub-

sections, we will describe how the game is played by

each of the participants mentioned above.

4.2.1 Game Played by Volunteer(s)

We request one or more volunteers to play the game.

The ﬁrst task that the volunteer needs to do is men-

tion the importance that he places on each of the QoS

attributes: Reliability, Waiting-time, Reputation, and

Cost. If the volunteer customer is comfortable doing

KMIS 2010 - International Conference on Knowledge Management and Information Sharing

264

this, he is asked to express this importance in terms

of values assigned to respectively α, β, γ, and δ. If

the customer is unable to do this, he is asked a set of

questions on whose basis values are assigned to the

four parameters.

Next, the volunteer starts sending service requests

into the transportation service domain. Every time

that the volunteer has to make a decision regarding

the means of transportation, he is shown the following

information regarding the potential transportation ser-

vices available: (i) the current fail-rate of each of the

transportation services, (ii) the queue length of other

waiting requests at each service, (iii) the travel time

of each of the potential transportation services, (iv)

the reputation of each of the services, (v) the cost of

the services, and (vi) a map showing the current lo-

cation of the customer and the location that he needs

to travel to (to give him an idea of the geographical

features separating the locations). This is shown in

Figure 5. On the basis of all these factors the vol-

unteer makes an intuitive judgement on which of the

available services to select. The volunteer is appealed

to take his decision conforming as far as possible to

the weights (α, β, γ, and δ) assigned by him to the

four QoS attributes.

This process is repeated for the volunteer for a

large number of travel requests. The total distance

travelled by each request is noted and ﬁnally the aver-

age of these distances is the score of the volunteer. If

there are several volunteers, the same is done for the

other volunteers also.

Selection by the volunteer

car ﬂight train boat

car

ﬂight

train

boat

Origin

Cost

Queue-length

travel time

Fail-rate

Reputation

intuitive

judgement

Volunteer is requested to take

decisions conforming to his values

of !, ", #, $ as far as possible

!, ",

#, $

Figure 5: The Ambitious Traveler game as played by the

volunteer.

4.2.2 Game Played by our Composition Model

We next play the game with our Afﬁnity model for

service composition taking the selection decisions. At

each service level, the value of Reliability for each of

the potential transportation services is calculated us-

ing the technique described in (Srivastava and Soren-

son, 2010a); the waiting-time is calculated as the ra-

tio of the queue-length of the waiting-requests at each

service to the completion-rate (in this case travel rate)

of the service (Srivastava and Sorenson, 2010b); the

reputation for each transportation service is calculated

using the feedback collected from previous requests;

and ﬁnally the cost is dynamically calculated for each

of the transportation services taking all the functional

and non-functional attributes and the environment of

the service into account.

The Coupling between the service elements at ad-

jacent levels is determined by logical judgement. For

example, in Figure 4, the Coupling between any of the

transportation services to London and the ‘car’ ser-

vice at the level of NYC (New York City) would have

a value of 0. This is because it would be absurd to

use a car as a means of transportation between Lon-

don and NYC given the expanse of the Atlantic Ocean

between the two cities!

With the values of the QoS attributes and Cou-

pling calculated, the equation for calculating the

Afﬁnity between service levels shown in equation (1)

is utilized. The values of α, β, γ, and δ as speciﬁed by

the volunteer are used.

The selections using the composition model are

done for travel requests that are identical to those sent

out by the volunteer(s) and under identical circum-

stances. The total distance covered by each request is

noted and the average of the distances of the requests

gives the ﬁnal score of the composition model.

4.2.3 Game Played at Random

To ensure that the scores of the above two selection

techniques are actually relevant and meaningful, the

game is also played with the service selections being

made at random. The random selections are done for

requests that are identical to the requests of the volun-

teer(s) and the total distance travelled in each case is

noted. The average of these distances gives the score

when the selections are made at random.

The scores of the three selection techniques are

compared. If the scores of our Afﬁnity model for ser-

vice composition are statistically signiﬁcantly better

than the other two, this serves to validate the tech-

nique. Even if the scores of our composition model

are comparable with those of the intuitive judgement

of the volunteer, it suggests that our automated selec-

tion method is useful given the signiﬁcant overhead

associated with the thought and judgement selection

process of individual volunteers.

THE VALIDATION OF A DYNAMIC SERVICE COMPOSITION MODEL USING A SIMPLE GAME

265

5 FUTURE ENHANCEMENTS TO

THE GAME

The Ambitious Traveler game in its present form is

quite simple and is robust as a validation tool for dy-

namic systems from a customer point of view. We

however envisage a much greater role for the game

in the future with possible extensions and enhance-

ments.

As the game stands now, it comprises volunteers

who serve as customers that make selections given a

service domain that is static. An extension that could

have a far reaching impact could be to also have a set

of volunteers who would serve as providers. These

providers would be responsible for setting up trans-

portation services that would be used by customers to

move from one city to another. The providers could

adjust the cost of the services to ensure a comfort-

able proﬁt margin even while keeping up with com-

petition. The other quality attributes of the services

could also be adjusted depending upon the resources

available and with an eye on the level of demand. The

provider would beneﬁt immensely from the game by

being in a position to try out various marketing strate-

gies in a synthetic set up and getting a handle on what

to expect.

A more ambitious enhancement to the game could

be by developing it as an application for a web map-

ping service like Google maps (GoogleMaps, 2010).

Prospective service providers could set up ﬁctitious

services at different locations on the map. Prospective

customers could use such services, provide feedback,

choose between competing services, and much more.

A virtual world on the web mapping application could

be set up with prospective customers and providers in-

dulging in transactions and activities that would nor-

mally occur in a real situation. All this would give

both the customer and provider a ‘feel’ for the situa-

tion and ultimately enable them to make much better

business decisions.

6 CONCLUSIONS

The Ambitious Traveler game described in this paper

serves as an inexpensive and convenient means to val-

idation of a dynamic service composition model. This

game may easily be modiﬁed to suit the purpose of

validation of any other kind of dynamic system model

also. Although the game described is still under de-

velopment, the idea of utilizing such a game for vali-

dation purposes is unique and novel.

The game described is currently good for validat-

ing the service composition model from the customer

point of view. It ensures that the interests of the cus-

tomer are upheld and the service composition formed

conforms to the requirements of the customer. We

hope that with a few enhancements, the game could

be used to validate a model that allows the provider

also to make changes to the available service domain.

Such a game could be used to uphold the interests of

both the customer and provider and would therefore

ensure a degree of co-value to both sides.

REFERENCES

Colombo, M., Nitto, E. D., and Mauri, M. (2006). Scene: A

service composition execution environment support-

ing dynamic changes disciplined through rules. In In-

ternational Conference on Service Oriented Comput-

ing (ICSOC). Springer Berlin / Heidelberg.

Cormen, T. H. (1990). Introduction to algorithms. The MIT

press.

Eid, M., Alamri, A., and Saddik, A. E. (2008). A reference

model for dynamic web service composition systems.

In International Journal of Web and Grid Services.

GoogleMaps (2010). http://maps.google.com/.

Shen, L., Li, F., Ren, S., and Mu, Y. (2007). Dynamic

composition of web service based on coordination

model. In The Joint International Conferences on

Asia-Paciﬁc Web Conference and Web-Age Informa-

tion Management.

Silva, E., Pires, L. F., and van Sinderen, M. (2009). Sup-

porting dynamic service composition at runtime based

on end-user requirements. In 1st International Work-

shop on User-Generated Services (UGS).

Srivastava, A. and Sorenson, P. G. (2010a). The inﬂuence of

service interactions on individual service reliability in

a composition scenario. In International Conference

on Web Information Systems and Technologies (WE-

BIST).

Srivastava, A. and Sorenson, P. G. (2010b). Utilizing the

waiting-time criterion for selecting services in a com-

position scenario. In International Conference on Ser-

vices Computing (SCC).

Wu, C. (2008). Service distribution and service discovery

through a public web services platform. In Ph.D. The-

sis, Curtin University of Technology, Digital Ecosys-

tems and Business Intelligence Institute.

KMIS 2010 - International Conference on Knowledge Management and Information Sharing

266