Stress-testing centralised model stores

Abstract

One of the current challenges in model-driven engineering is enabling effective collaborative modelling. Two common approaches are either storing the models in a central repository, or keeping them under a traditional file-based version control system and build a centralized index for model-wide queries. Either way, special attention must be paid to the nature of these repositories and indexes as networked services: they should remain responsive even with an increasing number of concurrent clients. This paper presents an empirical study on the impact of certain key decisions on the scalability of concurrent model queries, using an Eclipse Connected Data Objects model repository and a Hawk model index. The study evaluates the impact of the network protocol, the API design and the internal caching mechanisms and analyzes the reasons for their varying performance.

Publication DOI: https://doi.org/10.1007/978-3-319-42061-5_4
Divisions: College of Engineering & Physical Sciences
Additional Information: Funding: EPSRC, through the Large-Scale Complex IT Systems project (EP/F001096/1) and by the EU, through the MONDO FP7 STREP project (#611125).
Event Title: 12th European Conference on Modelling Foundations and Applications, as part of the conference on Software Technologies: Applications and Foundations
Event Type: Other
Event Dates: 2016-07-06 - 2016-07-07
Uncontrolled Keywords: Computer Science(all),Theoretical Computer Science
ISBN: 978-3-319-42060-8, 978-3-319-42061-5
Last Modified: 26 Mar 2024 08:07
Date Deposited: 22 Dec 2016 09:20
Full Text Link: http://link.spr ... 3-319-42061-5_4
Related URLs: http://www.scop ... tnerID=8YFLogxK (Scopus URL)
PURE Output Type: Conference contribution
Published Date: 2016-06-23
Published Online Date: 2016-06-23
Accepted Date: 2016-04-01
Authors: García-Domínguez, Antonio (ORCID Profile 0000-0002-4744-9150)
Barmpis, Konstantinos
Kolovos, Dimitrios S.
Wei, Ran
Paige, Richard F.

Download

[img]

Version: Accepted Version


Export / Share Citation


Statistics

Additional statistics for this record