High Level Modeling to Support Software Design

Transcription

1 by Gerrit Muller Buskerud University College Abstract High level models are simple models with the primary goal to support understanding, analysis, communication and decision making. The models have different complementary representations and formats, e.g. visual diagrams, mathematical formulas, and quantitative information and graphs. The models are made at different levels to guide software design choices: enterprise level, specification level, and design. Distribution This article or presentation is written as part of the Gaudí project. The Gaudí project philosophy is to improve by obtaining frequent feedback. Frequent feedback is pursued by an open creation process. This document is published as intermediate or nearly mature version to get feedback. Further distribution is allowed as long as the document remains complete and unchanged. status: preliminary draft

2 Figure Of Contents TM reflection on modeling illustration by case problem statement approach technical complexity context high-level conclusions 2 Gerrit Muller HLMSDlogo

3 Ubiquitous Information and Services consumers businesses government business business... information and services anywhere, anytime IT infrastructure IT companies 3 Gerrit Muller HLMSDubiquitous

4 But, Horrendous Failure Rate slow response, outages, human-less helpdesks, silly excuses (the computer could not...), identity-theft, lost privacy consumers businesses government business business... budget overruns total failures information and services anywhere, anytime IT infrastructure IT companies late delivery of new products, poor scaling of new services, interference of features,... 4 Gerrit Muller HLMSDhell

5 Typical Architecture Levels in IT enterprise context enterprise system specification system design enterprise architecture information architecture system architecture local SW choice subsystem subsystem subsystem 5 Gerrit Muller HLMSDlevels

6 reflection on modeling illustration by case problem statement approach technical complexity context high-level conclusions 6 Gerrit Muller HLMSDlogoApproach

7 Simplified Framework for Modeling business : profit, etc. operational costs stakeholder benefits workload risks usage context key performance : throughput, response reliability availability scalability... and their mutual relations (emerging?) properties : resource utilization load latency, throughput quality, accuracy... system enterprise & users business : profit, etc. operational costs stakeholder benefits workload risks requirements black box view creation life cycle business life cycle context design realization technology 7 Gerrit Muller MAOwhatModeling

8 Recommendations for Modeling principles use feedback work incremental work evolutionary be explicit make issues tangible objectives support communication facilitate reasoning support decision making create maintain understanding insight overview translate into help to achieve translate into Time-box Iterate Multi-view recommendations Quantify early Measure and validate Multiple levels of abstraction (Simple) mathematical models Analysis of accuracy and credibility System and its context Visualize 8 Gerrit Muller MAOrecommendations

9 reflection on modeling illustration by case problem statement approach technical complexity context high-level conclusions 9 Gerrit Muller HLMSDlogoTechnical

10 Example Web Shop screen screen screen client client network web server network client consumer browse products order pay track exhibit products sales & order intake order handling stock handling financial bookkeeping enterprise logistics finance product management customer managment customer relation management update catalogue advertize after sales support data base server product descriptions logistics ERP financial customer relations 10 Gerrit Muller MAFTexampleWebShop

11 Web Shop: NFR s, Properties and Critical Technologies system NFR's: performance browsing initial cost running costs reliability/availability scalability order rate maintainability effort product changes effort staff changes security (emerging?) properties : resource utilization server load, capacity memory load, capacity response latency redundancy order throughput product data quality product definition flow staff definition flow security design compartimentalization authentication encryption critical technologies caching load balancing pipelining virtual memory memory management data base transactions XML for customization and configuration firewalls virtual networks Gerrit Muller MASMwhatWebShop

12 Purpose of Picture Cache Model in Web Shop Context screen screen screen client client client response time browse products data base server web server network network exhibit products product descriptions required server capacity required server capacity logistics ERP financial picture cache customer relations 12 Gerrit Muller MASMwebShopPictureCache

13 reflection on modeling illustration by case problem statement approach technical complexity context high-level conclusions 13 Gerrit Muller HLMSDlogoComplexity

14 Level of Abstraction Single System 10 0 number of details system requirements multidisciplinary design static system definition monodisciplinary 14 Gerrit Muller RAPpyramid

15 From system to Product Family or Portfolio number of details system system portfolio systems multidisciplinary monodisciplinary increase multidisciplinary monodisciplinary Gerrit Muller DRALpyramidGrowth

16 Product Family in Context enterprise context enterprise number of details stakeholders systems multidisciplinary design parts, connections, lines of code 16 Gerrit Muller RAPdiabolo

17 RA: Capturing the Essence number of details enterprise context enterprise stakeholders systems multi-disciplinary design parts, connections, lines of code high-level models some context details are essential high-level focus area some technical details are essential 17 Gerrit Muller HLMSDdiabolo

18 reflection on modeling illustration by case problem statement approach technical complexity context high-level conclusions 18 Gerrit Muller HLMSDlogoContext

19 Data Sources of Web Server client content preparation content web server data quality? content content content provider provider provider shop configuration e.g., staff, roles system configuration e.g. resource allocation 19 Gerrit Muller MALCwebServerChanges

20 Example Product Portfolio Change Books sales info product portfolio characteristics O'Reilly Addison Wesley Springer publisher selection system selection depends on business life cycle changes determined by business characteristics new books per year source: UK (1) USA(2) China(3) 206k (2005) 172k (2005) 107k (1996) 68k (1996) 101k (1994) sales WH Smith Amazon "long tail" India(21) 12k (1996) source: item 20 Gerrit Muller MALCbookExample

21 Example Customer Change internet: broadband penetration growth in Q1 '04 Q2 '04 Q2 '04 Asia Pacific total 48M 54M 12.8% China 15M 19M 26.1% India 87k 189k 116.8% What is the expected growth of # customers? What is the impact on system and infrastructure? What is the impact on CRM (Customer Relation Management)? What is the impact on customer, sales support staff? 21 Gerrit Muller MALCcustomerExample

22 Web Shop Content Update screen screen How much time/effort is needed for content updates? client client How much staff is needed? What is the impact of errors in content updates? network How many errors can be expected? What is the impact of content updates on server loads? web server network publication new content and updates content definition data base server product descriptions logistics ERP financial customer relations 22 Gerrit Muller MALCwebShopContentUpdate

23 Web Shop Content Change Effort prepare change 1 review input select info layout&cosmetics check-in prepare change 2 prepare change n effort changes = n changes *(t prepare + t verify )+t commit verify change 1 inspect source inspect result verify change n commit changes #fte = effort changes / hours per day n changes per day effort changes 1 uur 10 uur uur #fte with t prepare t verify t commit = 4 min = 2 min = 1 min hours per day = 8 hours 23 Gerrit Muller MALCeffortContentChanges

24 Example of Client Level Changes main access through search personalization catalogue entries Up-to-date information: Bestsellers What Other Customers Are Looking At Right Now other advertizements styling: frequently updated, fashion! standard boilerplate snapshot of 24 Gerrit Muller MALCmoreChanges

25 Web Shop Security and Changes public internet protected production area web server screen client network screen client What is the security model? What is the impact on server loads? What is the impact on staffing? What is the impact of changes in staff? What is the impact of changes on security? data base server very secure intranet network product descriptions logistics ERP financial content definition customer relations 25 Gerrit Muller MALCwebShopSecurity

26 Web Shop Reliability and Changes new faults = average fault density * #changes severity hit probability detection probability #errors = faults f( severity, hit probability, detection probability) Jansen iso Janssen operator iso sales repr low high high high low medium 26 Gerrit Muller MALCreliability

27 reflection on modeling illustration by case problem statement approach technical complexity context high-level conclusions 27 Gerrit Muller HLMSDlogoHighLevel

28 High Level Models Support Communication number of involved people 10 tangible model based architecting 1 fast but intangible in the architect's head 1 minute 1 hour 1 day 1 week 1 month time per iteration cycle 28 Gerrit Muller MAREcommunication

29 Model Life Cycle understanding exploration optimization verification try out models simple and small models?? most try out models never m leave the desk or computer mm m m m of the architect! abandoned abandoned many small and simple models are used only once; some are re-used in next projects archived archived not maintained not maintained re-use mmm m mm archived not maintained m re-used in next project m re-used in next project m re-used in next project substantial models re-use m m re-used in next project m re-used in next project m re-used in next project substantial models capture core domain know how; they evolve often from project to project. creation and evolution of intellectual property assets 29 Gerrit Muller MAREmodelLifeCycle

30 Example Model Life Cycle understanding exploration optimization verification try out models load/cost simple and small models substantial models (IP assets) function mix customer global distribution load/cost peak impact global customer demographics integral load model load/stress test suite web server performance webshop benchmark suite 30 Gerrit Muller MAREmodelLifeCycleExample

31 reflection on modeling illustration by case problem statement approach technical complexity context high-level conclusions 31 Gerrit Muller HLMSDlogoConclusions

32 Conclusions lack of integral understanding of software in human and business context causes horrendous failure rate of IT projects and systems simple high-level models create understanding across specialties and stimulate fact based decision making 32 Gerrit Muller HLMSDconclusions