Towards a network ecology of software ecosystems: an analysis of two OSGi ecosystems

Publikation: Bidrag til bog/antologi/rapportKonferencebidrag i proceedingsForskningfagfællebedømt

Klaus Marius Hansen, Konstantinos Manikas

"Software ecosystems'' are gaining importance in commercial software development; the iPhone iOS and Salesforce.com ecosystems are examples of this. In contrast to traditional forms of software reuse, such as common platforms or product lines, software ecosystems have a heterogeneous set of actors sharing and collaborating over one or more technological platforms and business model(s) that serve the actors. However, little research has investigated the properties of actual software ecosystems.
In this paper, we present an exploratory study of software ecosystems using the formalizations and metrics of the "network ecology'' approach to the analysis of natural ecosystems. In doing so, we mine the Maven central Java repository and analyze two OSGi ecosystems: Apache Felix and Eclipse Equinox. In particular, we define the concept of an ecosystem ``neighborhood'', apply network ecology metrics to these neighborhoods (including a keystone index that identifies the importance of elements in the ecosystem), and compare the ecosystems.
OriginalsprogEngelsk
TitelThe 25th International Conference on Software Engineering & Knowledge Engineering SEKE 2013 : proceedings
Antal sider6
ForlagKnowledge Systems Institute Graduate School
Publikationsdato2013
Sider326-331
ISBN (Trykt)978-1-891706-33-2
StatusUdgivet - 2013
BegivenhedThe International Conference on Software Engineering & Knowledge Engineering (SEKE) 2013 - Hyatt Harborside at Logan Int’l Airport, Boston, USA
Varighed: 27 jun. 201329 jun. 2013
Konferencens nummer: 25

Konference

KonferenceThe International Conference on Software Engineering & Knowledge Engineering (SEKE) 2013
Nummer25
LokationHyatt Harborside at Logan Int’l Airport,
LandUSA
ByBoston
Periode27/06/201329/06/2013
NavnProceedings of the International Conference on Software Engineering and Knowledge Engineering
Vol/bind2013
ISSN2325-9000

ID: 120839340