Patrizio Pelliccione , Massimo Tivoli Software Engineering and - - PowerPoint PPT Presentation

Università degli Studi dell’Aquila Università degli Studi dell’Aquila

Software Engineering and Architecture Group Computer Science Department, University of L’Aquila, Italy

Marco Autili, Vittorio Cortellessa, Davide Di Ruscio, Paola Inverardi, Patrizio Pelliccione, Massimo Tivoli

March 20, 2012 - Monterey Workshop

2

Users produce their own piece of software, by mainly reusing existing software with a goal oriented and opportunistic use in mind The focus of software production shifts from domain specific software applications to software integrator systems Software has to be able to evolve, react and adapt to a continuously changing environment, while guaranteeing dependability

Users role Integrator synthesis Evolution

We refer to the general notion of dependability, as defined by IFIP Working Group 10.4: “the trustworthiness

• f a computing system which allows reliance to be justifiably placed on the service it delivers”.

3

Creationistic view

• A producer is the
• wner of the artifact,

and with the right tools she can supply any piece of information

Experimental view

• The knowledge of a

software artifact is limited to what can be

• bserved of it
• Theoretical barrier

limits the power and extent of observations

Galbraith’s definition of uncertainty, it In the EAGLE context uncertainty corresponds to a measure, in a given metric system, of the incompleteness and inaccuracy of the models with respect to the goal G, which is due to

4

5

• M. Autili, V. Cortellessa, D. Di Ruscio, P. Inverardi, P. Pelliccione, and M. Tivoli.

Eagle: engineering software in the ubiquitous globe by leveraging uncertainty. In Proceedings of the 19th ACM SIGSOFT symposium and the 13th European conference on Foundations of software engineering, ESEC/FSE ’11, pages 488–491, New York, NY, USA, 2011. ACM.

Explore

• Exploration of available software and makes explicit the

degree of uncertainty associated with it in relation to an

• pportunistic goal G

Integrate

• Assistance of the producer in creating the appropriate

integration means towards G

Validate

• Validation of the integrated system to assess its quality with

respect to G and the current context

6

7

The integration solution compensates the lack of knowledge of the composed software by adding integration logic

• Existing services are discovered out of a goal-oriented

choreography specification as suitable participants

• Coordination of the interaction behavior of the

participant services in order to fulfill the specified choreography

• Research in this domain does not explicitly account for

neither uncertainty about the behavioral suitability of the discovered services nor the uncertainty of the context changes

8

9

• While attending a conference in London,

Patrizio was notified by the airline company, through SMS, that his flight back to Rome was cancelled

• Flights across Europe were cancelled due to

the ash cloud

• Patrizio shared the situation with tens of

thousands of people: uncertainty and chaos were the only certainties

• This situation lasted for six days and people

were forced to self-organize their travel back home

10

Patrizio through phones, stuttered internet connection and social chatting

• rganized an expensive, non-optimal, and 30 hours long multimodal travel

which included buses, trains, ships, and rented car

• EAGLE (specialized for transportation) would have

supported Patrizio in producing a software system

• riented to organize and support his travel

11

12

The goal G is a combination of a functional and a non- functional properties, that are respectively: (i) to achieve a successful interaction among EcommerceWS and a client of it, i.e., the client always progresses on buying items, (ii) to achieve a certain level of reliability of the whole system, where this attribute is given by the combination of the client and the web service reliabilities.

13

• We consider a version of the StrawBerry tool that, for the

EAGLE purposes, is enhanced in order to deal with the uncertainty degree of the elicited models

• The StrawBerry method automatic discovers the behavior

protocol of a WS

• StrawBerry derives from the WSDL a partial ordering relation

among the possible invocations to the WSDL operations

• BPA models the interaction protocol that a client has to abide

by to correctly interact with the WS

14

15

Antonia Bertolino, Paola Inverardi, Patrizio Pelliccione, and Massimo Tivoli. 2009. Automatic synthesis of behavior protocols for composable web-services. In Proceedings of ESEC/FSE '09. ACM, New York, NY, USA, 141-150.

E-commerce web service: Explore step

E-commerce web service: Integrate step

E-commerce web service: Validation step

19

Università degli Studi dell’Aquila Università degli Studi dell’Aquila

Questions ?

Software Engineering and Architecture Group Computer Science Department, University of L’Aquila, Italy

March 20, 2012 - Monterey Workshop

Marco Autili, Vittorio Cortellessa, Davide Di Ruscio, Paola Inverardi, Patrizio Pelliccione, Massimo Tivoli