توسعه محصول توزیع شده جهانی با استفاده از یک چارچوب مدیریت پروژه جدید

There is a growing pressure on corporations to streamline operations involving multi-site development. In this article, we propose a new project management framework that can be used for managing and tracking distributed development of a large product efficiently. We assume that the development team is organized as Centers of Excellences (CoEs), which may be geographically distributed. The framework described in this article, is capable of presenting diverse views (e.g., Feature, Load, and Release) of a product to its stakeholders seamlessly. It also streamlines communication between the CoEs. A product plan, designed using this framework, offers high resilience to requirements changes during the development cycle. The planning issues for a large (over 200 features) product are described in detail next in the context of the framework. We conclude with a brief scope for future work. The paper chooses examples from the Software domain though the approach is general enough, and is applicable to other disciplines as well.

With technology leapfrogging and ushering its benefits bounteously into society, the expectation of its beneficiaries from the technology-oriented organizations is increasing more than ever before. This is forcing corporations to streamline their development strategies towards achieving faster turn-around time for product development at a much lower cost to sustain their competitive advantage. With globalization at the backdrop, capitalizing on the geographically distributed skills to achieve high quality product deliverables in a much shorter span of time is becoming a necessity today. Literatures abound in the area of distributed project management including tools that facilitate planning. PERT/CPM approaches have been in existence for quite sometime but they are limited in terms of their analytical power as cited by Cho [1]. Cho [1] had primarily concentrated on aspects such as Process Restructuring and Process Modeling of Distributed Project Management. Process Restructuring typically deals with the structure of the information flow in a complex project while Process Modeling captures the behavioural aspects of processes over timelines. Other efforts towards Process restructuring include approaches proposed by Steward [5] such as Design Structure Matrix or DSM that models information flow across tasks based on large-scale system decomposition. There have been focused activities towards development of project management tools. Rojers [3] developed a computer aided design tool to support assisting in large-scale system decomposition. As cited in Stallings [15], some researchers have also come forward with an “Intelligent Agent” based project management framework. This is an integrated project management framework where plan and design may change while the plan is in execution. Notification about any change is passed as `messages' to software agents, thus, minimizing human intervention. According to Petrie et al. [2], the problem of using traditional tools such as Mac Project, or MS Project, is that they are inherently based on single user model of planning, with very primitive change notification mechanisms. Centralized decision-making vs. distributed decision-making is what is highlighted in the first part of Petrie et al. [2]. One of the most important challenges in product development in a distributed environment is to ensure proper co-ordination and control [6]. Many recent studies [7] and [8] emphasize the use of Information Technology in tackling project management issues in a distributed environment. Others [9] and [10] explored the use of diverse Communications Technologies to address the co-ordination issues in a distributed environment. Less technical in nature, Lori Anschuetz [11] elaborates on how sharing of corporate culture, gaining insights into clients, building trust and brining in professional synergy can enable efficient co-operation and effective results for `virtual' organizations. Much of the work in the area of distributed Project management lacks the rigor in dealing with the concurrency issues [12], [13] and [14] when the same project involves development across multiple locations. We propose a simple approach towards distributed project management in this paper that is broadly based on Process Restructuring. We believe that, by establishing a suitable framework that helps in automating and consolidating Project Control related information, the distributed/co-located teams are better informed to take planning decisions, thus, giving them more bandwidth to focus on other issues relating to a new product. We define a protocol as well as an interface for communication between development teams for a product. By virtue of this, the existing skills and framework that an organization may have developed over the years can be readily used.

We have proposed a new project management framework that facilitates development of project plan for a large product in a geographically distributed environment. The methodology developed here applies in the context of product development organizations, which are structured as a collection of CoEs. The framework makes no assumption about the location of the CoEs, and so, they need not be geographically co-located. The framework is capable of integrating diverse views (Feature/Load/Release) of a product plan while enabling coherent, consistent presentation of these views seamlessly to the various stakeholders such as end customer, product management team, development team, and the higher management team. The structure of a product plan (Internal/Export blocks) using the proposed framework ensures minimal coupling and organized communication between the CoEs, offering a great degree of immunity against requirements volatility. It also facilitates easy debugging, as discussed in the paper, during the development of the project plan itself. The use of workgroup feature obviates manual tracking. The project management framework, thus, eliminates any geographical barrier between the CoEs, enabling organizations to capitalize on distributed skills for a successful product development. While we attempted to solve the problems of arriving at and maintaining a large product plan, the issues regarding optimal load contents plan remain largely unanswered. An optimal load contents plan will minimize idle slots between CUT complete and IT within a load (Fig. 3), thereby solving (4.1.1), which may be the future area of work. The examples that we considered in this paper were taken from the Software domain. One can, however, apply the same methodology to other disciplines as well. Finally, we experimented this framework for developing plans for a large software product having around 200 features, which we successfully deployed in our organization.