Orchestration of Global Software Engineering Projects

Christian Bartelt, Manfred Broy, Christoph Herrmann, Eric Knauss, Marco Kuhrmann, Andreas Rausch, Bernhard Rumpe, Kurt Schneider

Abstract

Global Software Engineering (GloSE) often fails to meet cost and quality expectations due to cultural friction between distributed sites. This paper proposes an "orchestration" framework that integrates project management, development processes, and technical artifacts into a cohesive virtual project environment. By utilizing site-independent and task-driven views, the framework aims to mitigate the communication overhead and process clashes inherent in global collaboration.

Introduction

Global Software Engineering (GloSE) is frequently adopted to leverage global talent and reduce costs. However, these projects often suffer from budget overruns and quality issues. The core challenge is not merely geographical distance or time zone differences, but the friction between disparate software engineering cultures. In co-located teams, "handshake" mechanisms—shared habits and vocabularies—develop naturally. In a global setting, these are disrupted, leading to underspecified components and massive communication overhead that can negate intended cost savings.

The Orchestration Framework

To address these challenges, the authors propose a conceptual framework for "orchestration." This involves (re-)orchestrating three specific layers to create a unified "virtual project environment" that sits above participating organizations.

  1. Project Set-up and Management: Establishing a unified organizational structure for the virtual project.
  2. Processes and Information Flows: Harmonizing development lifecycles and defining explicit interfaces between them.
  3. Artifacts and Product Models: Ensuring consistency and transparency of code, requirements, and documentation across all sites.
Conceptual framework for orchestrating global software engineering across organizational boundaries.

The centerpiece of this approach is the use of "Views"—snapshots of the project that are site-independent, task-driven, and role-specific. This allows developers to see only the code and specs relevant to their current tasks, while managers receive high-level status dashboards, effectively shielding participants from the underlying complexity of the global infrastructure.

Key Insights

  • Culture over Geography: The authors redefine GloSE as the allocation of work to sites with different software engineering cultures. This shift moves the solution space from simple communication tools to deep process integration.
  • The Virtual Project Layer: A global project should be treated as its own entity. This virtual layer acts as a mediator, collecting elements from participating sites and presenting them in a unified way.
  • Visibility of Internal Staff: Orchestration must account for "invisible" internal support staff at each site who, while not part of the core virtual team, are essential to the project's execution.
  • Explicit Interfaces: The lack of defined "handshake tasks" between organizations is a primary risk. Orchestration must make these interfaces explicit to prevent knowledge gaps.

Research Agenda and Limitations

The paper provides a qualitative roadmap for future research rather than quantitative data. Key areas for investigation include consistency control for mirrored artifacts, strategies for "stitching" together disparate company workflows, and risk-oriented management to mitigate the loss of communication richness.

While the framework focuses on professional and technical orchestration, the authors acknowledge that social aspects (like trust-building) and the practical challenges of integrating local practices without interrupting ongoing work remain significant open questions. Empirical validation in real-world industrial settings is required to prove that this orchestration framework can successfully improve the outcomes of global software projects.

Summary updated

BibTeX
@inproceedings{bartelt2009orchestration,
  author = {Christian Bartelt and Manfred Broy and Christoph Herrmann and Eric Knauss and Marco Kuhrmann and Andreas Rausch and Bernhard Rumpe and Kurt Schneider},
  title = {{Orchestration of Global Software Engineering Projects}},
  booktitle = {ICGSE},
  year = {2009},
  url = {https://openreview.net/forum?id=YR4g0YDYZKz},
}
Figure