Some Gottchas of Estimating Service Oriented Architecture Systems (SOA)
As Fredrick Brooks said “there are no silver bullets.” Unfortunately the software industry continues to tout every incremental step as that silver bullet which will solve the problems of software development. Don’t get me wrong. SOA when applied in the appropriate environment can improve productivity. Galorath analysts have been using SEER to estimate SOA systems for some years now. The basic conclusions are 1) SOA can be the best approach to achieving interoperability and reusability and 2) SEER accurately models the effort, schedule, risk, cost of software development of an SOA system as well as the cost of the IT services and infrastructure to support them. But the reality must be modeled, not just vendor claims.
Grace Lewis, Et Al. of the Software Engineering Institute did a good job of identifying some of the misconceptions of SOA as follow:
SOA Does Not Provide a Complete Architecture: SOA provides a pattern for an architecture, not the whole system architecture itself. Hence you can’t really buy SOA off the shelf. You can buy SOA products to help.
Dan Woods of Forbes points out: “The problem is that most IT departments are used to running applications. We can think of them as automakers assembling a car from many different parts. The auto manufacturer makes sure that all the parts fit together and work as they should so the car runs well. The world of SOA lifts the hood of the car and breaks the rest of the car into parts, that is, individual services that can be reassembled. “
Migrating Legacy Systems Is Nether Automatic Nor Easy… There Are Significant Costs
The Use of Standards Does Not Guarantee Interoperability In An SOA Environment
Testing SOA Systems Is Harder: “Because an SOA environment is distributed, loosely coupled, and asynchronous, testing can be significantly more complex than simply testing a set of known paths in a single system.”
Gartner also raised up the caution flag stating that SOA deployments were slowing down as people begin to understand the real costs.
Also, Forbes Dan Woods post provides a sobering observation: with SOA, IT departments end up bringing much of the work formerly handled by software vendors in house. So for estimation purposes plan on doing more IT development and services in house… And plan on training, retaining, and equipping those personnel.
Oh Glorious Day.. Weapons System Decisions To Be Made Based On Total Life Cycle Costs
The October 13 2008 issue of Federal Times has a story Titled “Better Estimates Urged for Weapons Programs” about John Young, Pentagon Acquisition Executive and his movement for “establishing life cycle metrics early in the acquisition process”
Galorath has been preaching evaluating total ownership costs, not just development, for years. And we understand the pressures to “get it out in the field to support the warfighter” Using all the SEER products such as SEER for Hardware, Electronics and Systems support total ownership costs and tradeoff analysis to determine the most effective system design and architecture from a set of alternatives.
We all realize that there are other issues beyond just total ownership cost to keeping our country and those in harms’ way safe. And I know that combining estimating of total ownership costs early and often can support that goal.
SEER Estimate By Comparison Professional Released To All Users
The SEER Estimate By Comparison (formerly SEER-AccuScope) comes in two flavors, the core and the professional. The professional version allows estimation of anything, such as
- total system cost
- software size
- hardware weight
- system value
- server capacity
- best choice from a set of alternatives
This new feature incorporates sophisticated mathematics for uncertainty and estimation.
For the first year Galorath is providing the professional version to all SEER users. Beginning the second year the professional functionality will be available as an upgrade.
Galorath recommends using an industrial strength database for an enterprise but will allow Microsoft Access for simple desktop installations.
This feature began shipping with the release of SEER for Software 7.3, in October, 2008.
GAO Knowledge Points
A valuable GAO document discusses knowledge points.. the certainty that systems are appropriate. “Because military programs tend to start product development with more unknowns, it takes them additional time, sometimes until well after production begins, to actually discover and capture enough solid information to attain full product knowledge and thereby virtually eliminate risk.” Read more
Estimating Systems Engineering…. The Upcoming Book.. and SEER
I had the opportunity to review Dr. Ricardo Valerdi’s (MIT) upcoming book on the COSYSMO model for estimating systems engineering: Systems Engineering Cost Estimation with COSYSMO. There were some real gems in it.
From the book: ‘It is known that increasing front-end analysis reduces the probability of problems later on (Blanchard and Fabrycky 1998) but excessive front end analysis may not pay the anticipated dividends. The key to accurately estimating early in a program is to estimate the appropriate level of systems engineering in order to ensure system success within cost and schedule budgets’ This chart is included showing a relationship of systems engineering effort. Of course COSYSMO is a model for estimating systems engineering effort parametrically.
Galorath is proud to offer an implementation of COSYSMO for those desiring that methodology for systems engineering.
The SEER Allocation handles much more:
Additionally, for complex systems Our SEER for hardware, Electronics, and Systems estimates for software, hardware and the total system: 1) systems engineering, 2) systems program management, 3) integration Assembly & systems test / 4) System support equipment. SEER-H’s organization is summarized as follows:
SEER Does Systems of Systems (SOS) Estimating
SEER-SEM works hand in hand with SEER for Hardware, Electronics * Systems (SEER-H) to estimate systems and systems of system (SOS), that is the integration of major systems together. For example, a squad car with a fire truck and a foot policeman’s GOS, three systems that need to be integrated and tested together. This capability works for hardware/software integration or just software intensive systems of systems.
Our VP of development, received an email from a customer who was concerned by the statement in a DACS paper that SEER doesn’t have the capabilities to estimate systems of systems. Upon internal review we believe it was a simple misunderstanding by the authors.
Dr. Denton Tarbet of Galorath added I recall at the PSM conference in 06 (Vail) that we had a working session on System of System estimating. Leaders from Lockheed, USC, and a software lead from NGC and Boeing was in the round table discussion. I know that at the time all in the discussion indicated they did not see a significant difference in a systems integration problem, except that for a “mega-project” you would have to be concerned about the major costs related to integration and testing of such a complex set of interfaces between what in reality were major systems on their own. Our SLC supports the estimation and analysis of a Mega-project . it would be a complex analysis to account for rollups of major systems into a mega-system.
PS.. the text of this BLOG disapeared for reaasons unknown. The author recreated the blog but it is not the same text as the original entry.
