Posted by Estimating S O A, As Easy as 1 2 3
Software & IT Track
Downloadable Files:
Abstract:
Metcalf’s Law tells us that the value of a telecommunications network is proportional to the square of the number of users of the system [1]. Service-Oriented Architectures (SOAs) capitalize on this phenomenon. Through a set of standard interfaces, services (software based capabilities) are made available to any consumer willing to follow the rules for interface and consumption. The loose coupling provided by the standard interfaces enables this plug and play capability. Taking advantage of such a notion promises great gains in efficiency for anyone looking to create interoperable, scalable applications that share information across boundaries.
According to Gartner, SOA will be used in more than eighty percent of mission critical operational applications and business processes by the year 2010. [2] Analysis of the literature indicates that the SOA vision leads to a belief of implementation efficiencies and cost savings of epic proportions. As the DoD moves forward with its vision of highly distributed net-centric capabilities in current and future DoD programs, it will be impossible to deploy, maintain and evolve capabilities without the benefit that SOA brings to the table.
SOA offers the DOD the promise of cost savings, data sharing, interoperability and increasingly agile operations. Getting ones head around the cost savings possible is not a trivial task. Cost estimators, analysts and project managers need tools and methods beyond traditional software cost estimation techniques to properly assess the costs of SOA projects. While this is not true for all aspects of SOA projects, there are aspects of this new paradigm that require shifts in estimating practices. This paper focuses on three such areas: SOA Governance, Middleware Integration and Service Development.
The proper integration of SOA into an organization necessitates significant changes in governance practices. In order for SOA to be successful, there needs to be governance at an organizational level to ensure services are aligned with business processes; that new services meet organizational needs; and that service reuse occurs effectively. New levels of governance and associated change management have cost implications beyond the scope of traditional estimating practices.
The integration of the middleware components at the heart of SOA solutions presents scoping challenges outside of those encountered with more traditional implementations, even those with a heavy reliance on standard COTS solutions. For the estimator these challenges require a new paradigm to ascertain the ‘size’ of the integration task. While in many ways service development is analogous to new software development, requirements to make services highly reusable and potentially useful outside of the current project can have significant cost implications as well.
This paper reports on research focused on providing the estimating community with solutions to the SOA specific issues outlined above. SOA specific cost drivers will be presented for each of these challenges. New or extended estimating methodologies will be presented along with guidelines for extending existing estimating practices to successfully address SOA projects.
Author:
Arlene Minkiewicz
PRICE Systems
Ms. Minkiewicz leads the Cost Research Department as Chief Scientist at PRICE Systems. In this role, she is responsible for the research and analysis necessary to keep the suite of PRICE Estimating products responsive to current cost trends. She works with industry leaders to collect and maintain cost research data and offers analyses of this data to the cost estimating community through the PRICE products.
Arlene’s most recent accomplishments include the development of a catalog of cost estimating relationships for systems and system of systems projects that will be delivered to the cost estimating community as part of the TruePlanning suite. Her most recent research on software sizing has been published in the March 2009 issue of Crosstalk.
Arlene frequently publishes articles on estimation and measurement in publications such as Software Development Magazine and Crosstalk. She speaks frequently on these topics at conference such as STC, ISPA, SCEA, IEEE Aerospace Conference, SEPG, and many others. Her ‘The Real Costs of COTS-Based Software Systems’ paper was recognized in 2004 by ISPA and SCEA as Best Paper in the Software Track. Her paper “A Case Study and Assessment of a COTS Upgrade for a Satellite Ground System”, co-authored with Marilee Wheaton of the Aerospace Corporation, received Best Paper in Software Track in 2006 by SCEA and her paper “The Evolution of Hardware Estimating” received Best Paper in the Hardware and EVM Track at the 2007 ISPA/SCEA joint conference.