Automatic Web Service Composition Models, Complexity and - - PowerPoint PPT Presentation

Automatic Web Service Composition Models, Complexity and Applications

PhD Defense Presentation PhD candidate Paul Diac

Alexandru Ioan Cuza University of Iasi Faculty of Computer Science

Supervisor: Prof. Dr. Dan Cristea

26 June, 2020

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

3/42

Automatic Web Service Composition - Models, Complexity and Applications Introduction

What is (Automatic) Web Service Composition?

web services provide a simple fine-grained functionality functionally described by input and output parameters composition of services enables new functionality automation of composition based on rules

4/42

Automatic Web Service Composition - Models, Complexity and Applications Introduction

Motivation

services are a popular and growing style of software design too many services to handle manually there is potential for automation interesting computational problems

5/42

Automatic Web Service Composition - Models, Complexity and Applications Introduction

Example - services

getCityCenter getMap locatePhone getWeather getLatLon

web services repository

nearbyStreet trafficInfo getCityDistrict

service providers

5/42

Automatic Web Service Composition - Models, Complexity and Applications Introduction

Example - the new request

getCityCenter getMap locatePhone getWeather getLatLon

web services repository

nearbyStreet trafficInfo getCityDistrict

service providers

getDrivingConditions

request

user

5/42

Automatic Web Service Composition - Models, Complexity and Applications Introduction

Example - can we solve the request?

getCityCenter getMap locatePhone getWeather getLatLon

web services repository

nearbyStreet trafficInfo getCityDistrict

service providers

getDrivingConditions

request

user composition

?

5/42

Automatic Web Service Composition - Models, Complexity and Applications Introduction

Example - yes, using some of the services

getCityCenter getMap locatePhone getWeather getLatLon

web services repository

nearbyStreet trafficInfo getCityDistrict

service providers

getDrivingConditions

request

user composition yes: 1 2 4 5 6 3

6/42

Automatic Web Service Composition - Models, Complexity and Applications Introduction

Example - conclusions

the composition is a sequence of services satisfies the user request and follows some rules it may not exist, or it may not be unique if these rules are formalized, automation is possible

7/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity

Models and Complexity

by model we refer to the methods of representing the system in service composition most often refers to parameter matching expressivity is good, complicated models are hard to adopt models generate problems of various computational complexities

8/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

9/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Name Matching Model

parameters generally have a short and intuitive name name coincide ⇒ parameters most likely refer the same concept simplest and first formalism for matching parameters knowledge is a simple set of known and learned parameters

10/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - services

getCityCenter getMap locatePhone getWeather getLatLon

web services repository

nearbyStreet trafficInfo getCityDistrict

service providers

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - services, with parameter descriptions

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - parameters with the same name

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - matching parameters

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - matching parameters

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - the request

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers

input = { phoneID, diameter }

getDrivingConditions

• utput = { map, traffic,

weather }

request

user composition

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - request parameter matchings

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers user composition

input = { phoneID, diameter }

getDrivingConditions

• utput = { map, traffic,

weather }

request

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - request parameter matchings

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers user composition

input = { phoneID, diameter }

getDrivingConditions

• utput = { map, traffic,

weather }

request

11/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Example - composition based on matching rules

input = { city, district }

getCityCenter

• utput = { lat, lon, altitude }

input = { lat, lon, diameter }

getMap

• utput = { map }

input = { phoneID }

locatePhone

• utput = { state,

city, district } input = { state, city }

getWeather

• utput = { weather }

input = { state, city }

getLatLon

• utput = { lat, lon }

web services repository

input = { city, lon, lat }

nearbyStreet

• utput = { street, number }

input = { city, street }

trafficInfo

• utput = { traffic }

input = { city }

getCityDistrict

• utput = { district }

service providers user composition

input = { phoneID, diameter }

getDrivingConditions

• utput = { map, traffic,

weather }

request

1 2 4 5 3 6

12/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

The Composition Problem

input: repository, request; output: valid composition or inexistent shorter composition preferable, shortest is NP-Hard the first Web Services Challenge - benchmark, Blake et al. 2005 many solutions translate the problem to a planning instance

13/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Algorithm Prerequisites

traversal of the parameter (hyper)graph O(R·logR+P), with R services and P total number of parameters efficient data structures, like map<int, set<int>> for parameters score-based heuristic chooses next service to add

14/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Polynomial Algorithm for Name-Match, Diac 2017

Algorithm Name-Match Composition Builder.

1: function FINDCOMPOSITION(R, i, g) // (repo, initial, goal) 2:

sol ← empty list // sequence of services

3:

K ← i.O // set of known parameters

4:

newServ ← NEWACCESSIBLESERVICE(R, sol, K)

5:

while newServ = NULL do

6:

sol.add(newServ)

7:

K ← K ∪ newServ.O

8:

if (g.I ⊆ K) then

9:

return sol // and exit

10:

newServ ← NEWACCESSIBLESERVICE(R, sol, K)

11:

return NULL

12: function NEWACCESSIBLESERVICE(R, sol, K) 13:

// if heuristic is used, simply return callableServices.first()

14:

for n ∈ R do

15:

if (n.I ⊆ K) ∧ (n ∈ sol) then return n

16:

return NULL

15/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Name Matching

Results on Generated Tests, relative to Zou et al. 2014 method

R

solution length

• pars. per

service Alg NameMatch GraphPlan Fast-Fwd time (s) len. time(s) len. time(s) len. 300 100 15 0.07 50 1.5 51 3.3 50 300 100 40 0.2 95 61 98 43 95 200 150 70 0.3 141 >3h ? 22 min 141 1000 200 50 1.6 130 39.9 133 error ? 1000 500 20 0.5 314 >3h ? error ? 1000 100 20 0.4 86 4.8 87 error ?

16/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Hierarchical

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

17/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Hierarchical

Hierarchical Model

first step to semantics, concepts are organized hierarchically subsumption: concepts can replace their parent concepts another edition of the Web Services Challenge, Bansal et al. 2008 parallel execution modeled by execution path metric

18/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Hierarchical

The Composition Problem

problem complexity does not significantly increase minimum execution path is polynomial Euler traversal of the concept tree is used for subtype queries O(R·logR+P+C) algorithm, where C is the number of concepts

19/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Hierarchical

Euler Traversal pre-computation ¸ Tuc˘ ar and Diac 2018

d b c f e a g 1 2 3 4 5 6 7 8 9 10 11 12 13 14 d g g b c c f f b e a a e d

Algorithm Euler traversal ⇒ entry and exit time

1: time ← 0 2: function LINEARIZATION(concept) 3:

time ← time + 1

4:

entryTime[concept] ← time

5:

for subconcept ∈ subconcepts[concept] do

6: LINEARIZATION(subconcept) 7:

time ← time + 1

8:

exitTime[concept] ← time

9: LINEARIZATION(root)

20/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Hierarchical

Algorithm Results ¸

Tuc˘ ar and Diac 2018, Rodriguez-Mier et al. 2011 +6 points for min. services and path; {+6, +4, +2} for fastest three if one min. reached

Tsinghua University University of Groningen Penna. University ¸ Tuc˘ ar algorithm result points result points result points result points Test 4

• min. serv.

10 +6 10 +6 10 +6 10 +6

• min. path

5 +6 5 +6 5 +6 5 +6 time (ms) 312 +4 219 +6 28078 34 +6 Test 5

• min. serv.

20 +6 20 +6 20 +6 20 +6

• min. path

8 +6 10 8 +6 8 +6 time(ms) 250 +6 14734 +4 726078 87 +6 Test 6

• min. serv.

46 37 +6 45

• min. path

7 +6 17 no result 7 +6 time(ms) 406 +6 241672 +4 132 +6 Score 46 Points 38 Points 24 Points 48 Points

21/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Relational

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

22/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Relational

Relational Model

multiple relations over concepts and parameters relation properties and inference rules multiple objects of the same type in the composition process which enables multiple calls of the same service relations in service definitions are either preconditions or effects

23/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Relational

Example, Diac et al. 2019

input = ∅ userInput (initially known)

.O=

• pers, homeUniv, foreignUniv,

isEmployeeOf(pers, homeUniv), hasDestination(pers, foreignUniv)

• .I=
• X, Y, Z, isEmployeeOf(X, Y),

isLocatedIn(Y, Z)

• locatedAtWorkRule

.O={isLocatedIn(X, Z)} pers(X), homeUniv(Y) .I={univ} getUniversityLocation .O={city, isLocatedIn(univ, city)} .I=

• X, Y, Z, hasDestination(X, Y),

isLocatedIn(Y, Z)

• destinationGenRule

.O={hasDestination(X, Z)} .I=      pers, source, dest, isLocatedIn(pers, source), hasDestination(pers, dest)      getAirplaneTicket .O={airplaneTicket} .I={airplaneTicket} userOutput (required) .O = ∅ foreignUniv(1) h

• m

e U n i v

( 2 )

p e r s ( X ) , f

• r

e i g n U n i v ( Y ) city(1)(Z) city(2)(Z) hasDestination(pers, city(1)) isLocatedIn(pers, city(2)) pers(pers) city(2)(source), city(1)(dest)

24/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Relational

The Composition Problem

finding objects to match input parameters is NP-Complete brute-force algorithm in O(K inputs · inputs 2) with K objects generated tests reveal: properties, rules, composition size

25/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Relational

Input ↔ Knowledge Matching, Diac et al. 2019

WebService1 Input Params p1 : type1

:

p2 : type2 p3 : type3 p4 : type4

• 1 : type1
• 2 : type4
• 3 : type5
• 4 : type1
• 5 : type2
• 6 : type2
• 7 : type3
• 8 : type6
• 9 : type3
• 2 : type4

rel1(p1,p2)

rel2(p2,p4) r e l 3 ( p 3 , p 2 )

type1 type2 type5 type4 type6

types hierarchy

r e l 1 (

• 1

,

• 5

) rel1(o3,o6) rel2(o6,o2) rel3(o9,o6) rel2(o5,o2)

type3

current level completed levels next levels

26/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Relational

Results on Generated Tests

• repository
• solution
• number of

rules applied run time (seconds)

| solution |

(ignoring rules)

63 11 0.07 11 30 14 74 0.30 15 30 8 3 0.04 13 46 4 0.02 4

27/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Object-Oriented

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

28/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Object-Oriented

Object-Oriented Model, Netedu et al. 2019

motivated by: OpenAPI, schema.org concepts have properties similar to class members in OOP

• nly one object of each concept ⇒ linear runtime complexity

properties: inherit downstream, learn upstream in the hierarchy

29/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Dynamic

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

30/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Dynamic

Dynamic Composition Problem

not a matching model, but a higher-level new view on WSC maintain a dynamic repository and a set of composition requests prepare for failover using backup compositions, if possible queries and updates interleaved + unpredictable events

31/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Dynamic

Use Case Diagram, Diac and Onica 2019

users . . . composition engine

repository

service providers . . .

s e r v i c e r e g i s t e r service remove service monitor

register service compute composition delete service drop compositon

. . . in

• ut

composition (in, out) find

si sj sk sl

32/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Dynamic

Failover recovery, Diac and Onica 2019

service wsp is deleted or breaks try to save both prefix ws1 . . . wsp−1 and suffix wsp+1 . . . wsk

• therwise, try to save only the prefix ws1 . . . wsp−1

save the first type of backup found, if any

user input user requested ws1

In O u t

ws2

In O u t

wsp

In O u t

wsk

In O u t

. . . . . . {known} {required}

service

\ {gen} \ {known}

suffix query composition query

32/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Dynamic

Failover recovery, Diac and Onica 2019

service wsp is deleted or breaks try to save both prefix ws1 . . . wsp−1 and suffix wsp+1 . . . wsk

• therwise, try to save only the prefix ws1 . . . wsp−1

save the first type of backup found, if any

user input user requested ws1

In O u t

ws2

In O u t

wsp

In O u t

wsk

In O u t

. . . . . . {known} {required}

service

\ {gen} \ {known}

suffix query composition query

32/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Dynamic

Failover recovery, Diac and Onica 2019

service wsp is deleted or breaks try to save both prefix ws1 . . . wsp−1 and suffix wsp+1 . . . wsk

• therwise, try to save only the prefix ws1 . . . wsp−1

save the first type of backup found, if any

user input user requested ws1

In O u t

ws2

In O u t

wsp

In O u t

wsk

In O u t

. . . . . . {known} {required}

service

\ {gen} \ {known}

suffix query composition query

32/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Dynamic

Failover recovery, Diac and Onica 2019

service wsp is deleted or breaks try to save both prefix ws1 . . . wsp−1 and suffix wsp+1 . . . wsk

• therwise, try to save only the prefix ws1 . . . wsp−1

save the first type of backup found, if any

user input user requested ws1

In O u t

ws2

In O u t

wsp

In O u t

wsk

In O u t

. . . . . . {known} {required}

service

\ {gen} \ {known}

suffix query composition query

33/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Others

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

34/42

Automatic Web Service Composition - Models, Complexity and Applications Models and Complexity Others

Different aspects of the composition problem

quality of service metrics (QoS) stateful services privacy and security parameter types and constraints executing compositions

35/42

Automatic Web Service Composition - Models, Complexity and Applications Applications Software Development

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

36/42

Automatic Web Service Composition - Models, Complexity and Applications Applications Software Development

Software related to Service Composition

execution is more frequent than automating the composition mashup tools: Yahoo! Pipes, Intel Mash Maker, Zapier1 more recent: Altova MapForce2, a graphical data mapping tool

1https://zapier.com 2https://www.altova.com/mapforce

37/42

Automatic Web Service Composition - Models, Complexity and Applications Applications Natural Language Processing

Contents

1

Introduction

2

Models and Complexity Name Matching Hierarchical Relational Object-Oriented Dynamic Others

3

Applications Software Development Natural Language Processing

4

Conclusion

38/42

Automatic Web Service Composition - Models, Complexity and Applications Applications Natural Language Processing

NLP related services

NLP services provide a particular use-case produce new information about texts based on previous data elements required and added are similar to service parameters but they are represented differently systems like ALPE (Cristea et al. 2007) have similar models and

• perations: annotation schemas ↔ parameters,

pipeline ↔ matching, etc.

39/42

Automatic Web Service Composition - Models, Complexity and Applications Conclusion

Conclusion

Conclusion

automatic WSC is a complex research area contributions: algorithmic optimizations and semantic modeling

Future Work

algorithmic: address new composition models and graph problems semantics: extend expressiveness, avoiding complicated models applications: framework, applications in the NLP domain

40/42

Automatic Web Service Composition - Models, Complexity and Applications Conclusion

References, (co-)authored papers

Paul Diac. ‘‘Engineering Polynomial-Time Solutions for Automatic Web Service Composition’’. In: International Conference on Knowledge-Based and Intelligent Information and Engineering Systems. 2017 Liana ¸ Tuc˘ ar and Paul Diac. ‘‘Semantic Web Service Composition Based

• n Graph Search’’. In: International Conference on Knowledge-Based

and Intelligent Information and Engineering Systems. 2018 Paul Diac, Liana ¸ Tuc˘ ar, and Andrei Netedu. ‘‘Relational Model for Parameter Description in Automatic Semantic Web Service Composition’’. In: International Conference on Knowledge-Based and Intelligent Information and Engineering Systems. 2019 Andrei Netedu, Sabin Buraga, Paul Diac, and Liana ¸ Tuc˘

• ar. ‘‘A Web

Service Composition Method Based on OpenAPI Semantic Annotations’’. In: International Conference on e-Business Engineering (ICEBE). 2019 Paul Diac and Emanuel Onica. ‘‘Towards Integrated Failure Recovery for Web Service Composition’’. In: International Conference on Software

• Technologies. 2019

41/42

Automatic Web Service Composition - Models, Complexity and Applications Conclusion

References

M Brian Blake, Kwok Ching Tsui, and Andreas Wombacher. ‘‘The EEE-05 Challenge: A New Web Service Discovery and Composition Competition’’. In: International Conference on e-Technology, e-Commerce and e-Service. 2005 Dan Cristea, Ionut Cristian Pistol, and Corina For˘

• ascu. ‘‘ALPE as LT4eL

Processing Chain Environment’’. In: Natural Language Processing and Knowledge Representation for eLearning Environments (2007) Ajay Bansal et al. ‘‘WSC-08: Continuing the Web Services Challenge’’. In: IEEE Conference e-Commerce Technology, e-Commerce and e-Services. 2008 Pablo Rodriguez-Mier, Manuel Mucientes, and Manuel Lama. ‘‘Automatic Web Service Composition with a Heuristic-based Search Algorithm’’. In: International Conference on Web Services. 2011 Guobing Zou, Yanglan Gan, Yixin Chen, and Bofeng Zhang. ‘‘Dynamic Composition of Web Services Using Efficient Planners in Large-Scale Service Repository’’. In: Knowledge-Based Systems (2014)

42/42

Automatic Web Service Composition - Models, Complexity and Applications Conclusion

Questions?

Thank you for your attention!