👐    design: bpm
👐    devops: bpm
Design SDLC - Software Development Life Cycle
Ideas concepts "versions" testing releases deployment
ICT at an organization by layers.
The organisatizion powered by ICT in a ship like constellation.
The engines (data center) out of sight below visibility.
Serving nultiple customers (multi tenancy) for best performance and best profits.
Too fast .. previous
- 2019 week:17
- Changing the temporary page, adding new subpages.
Rework on old DTAP pages and infographics.
- This page cahnged but all subpages are outdated copies of the temporary version.
| Reference || Topic || Squad |
| Intro ||ICT at an organization by layers. ||01.01 |
| DTAP 101 ||DTAP Introduction ||02.01 |
| 👓 layers ||deep dive layers, VMAP ||02.xx |
| Perspectives ||DTAP, Dimensions - Perspectives ||03.01 |
| 👓 Dimensions ||Multiple dimensions layers *stages ||03.xx |
| BP-mdl2 ||Business Proces model-2, ALC ||04.01 |
| 👓 ALC type2 ||Business Apllications - 3GL ALC ||04.xx |
| BP-mdl3 ||Business Proces model-3, ALC ||05.01 |
| 👓 ALC type3 ||Low code Analytics, Business ALC ||05.xx |
| What next ||Change ICT - Transformations ||06.00 |
| ||Following steps ||06.02 |
DTAP Layers Multiple dimensions
The high level concepts, enterprise architect, global guidelines, the governance are leading.
There are three lines for relase-management, versioning, with their dependicies. Those are:
- Analyses of Business Applications. Goal: optimizations
- Business Applications having a/ logic code and b/ data information
- Infra: tools (DBMS, messaging, sFTP) - middelware (ERP, BI, Analytics).
- Infra: hardware, operating system, basic security layers & facilities
The DTAP Develop, Test (integration), Acceptance (Logic -User, Technical, Persons Education) are part of a life cycle.
The waterfall model (figure wikipedia) is well known.
There are several words being different and some are not present in the DTAP staging.
- Requirements = (done before)
- Design = (done before)
- Implementation = Development
- Implementation = Testing
- Verification = Acceptance
- Maintenance = Production
VMAP - DTAP dependicies
Realizing a DTAP implementation using the VMAP model is recommended.
- Allowing agile project planning.
- Acknowledging the dependicies between actions.
- Not following: every stage should be signed off before starting the next one.
The process design in the infographic VMAP
is an old one.
It is combined with TMAP
TMap NEXT is an approach to structured testing.
is a better approach practicing testing.
(vmap deep details, 👓 click figure)
DTAP dimensions - perspectives
PDCA cycle refined with SDLC
Example steps refined cycle:
Not every step needs to be executed, the disposal is only done when it is the result of the evaluation.
Running, Maintaining - Developping Building
Before you can build something you have to organize the working force.
There is a reversed order, starting at the bottom it looks like:
Building, starting at the bottom (right).
Organizing the workforce, starting at the bottom (left).
Building a ICT system is more easy when it is an new one. Maintaining and changing whats is operational is more challenging.
It is like:
Maintaining, starting multiple at the top ltr (right).
New System, starting at the top ltr (left).
Details on perspectives dimensions, 👓 click one of the figures.
Starting Left to right (ltr) working to an end.
- The two Infra (Infrastructure green) lines are being joines although there are important differences in their life Cycle, release management, behavior.
- The Analytics line (orange) is left open for the choice of human algorithmic decisions in the proces "model-2" or rolling out in a robotic proces "model-3".
- business line (blue) is in the diagonal, always having areversal in time somewhere (requirements realizations).
Business Proces model-2, ALC
Deployment - Release - Versions
Deployment (release management) while documenting of the active software versions, what has been deployed and being used, is a business requirement.
In ICT frameworks used in legal directives, versioning - release mangement, is mentioned.
SDLC is touching almost anything in an ICT environment. It is getting most attention in organizing the work getting done.
The following relationships are here in the mindmap approach:
Is used at:
👓 Machine supported develop
Change ML AI.
👓 threats for data & tools
Proces Life Cycle.
👓 resulting Life Cycle
ALM, business Life Cycle.
Details to be found at:
👓 BI Analytics Cycle
👓 Meta data
Versioning is mentioned with ITSM (IT Service Management) for the operational usage (Production), not for what developpers are doing daily.
Classic model-2 Life Cycle
There are two components involved: a/ "business data" and b/ "business logic".
The classic well known SDLC focusses only on the "business logic".
The "business data" set aside, isolated, not approachable.
👓 Click on the figure for details on classic deployment.
- Data : There will be no exchange between the DTAP stages.
Data in the non-production stage are fake, not traceable to natural persons.
- Logic: Artifacts may be promoted by:
- copy everything, also all the ones not in scope for change - maintenance.
In a distributed environment whithout connectivity this is the only option.
- only having those being changed in scope temporarly locked for work.
In a centralized environment very well to govern wihtout the burden of duplicates.
- External Connections are part of the data segregation.
They are not visible with code deployment. Good care is needed.
- For technical performance reasons with differences in CIA (Confidentiality Integrity Availability) there can be several Production machines servicing with Production tools: Production Business Applications.
Business Proces model-3, ALC
The Life Cycle using analytics &robotics is different. The classic approach valid for logic but getting extended.
Modelling (developping) is based on the business production data, with all resulting dependicies and consequences.
The artifacts and deployment must be more clear classified compared to the "business logic only" approach.
new terminology "model"
The modelling part has got his own life cycle:
- data preparation
- extract features
- train model
- evaluate model
Instead of human defined decisons it is humand guided, computer assisited, best decision (champion) too choose by business responsibility.
Low code Analytics, Business ALC
Haven choosen the best decision a deployment is to be set up. The development line connecting to the operational line.
There are up to five components involved.
Details on modernized deployment, 👓 click on the figure.
- Logic Data: The data delivery (ER star-model) is done for modelling and scoring seperately. The data model is almost the same, except:
- Development line: As much as possible PII (Personal Identifiable Information) is excluded, historical data made in extended periods.
- Operational line: PII data is included when needed in deliveries, historical data only for relevant periods.
- Logic Data: Needing a denormalized table is doen as part of modelling (data preparation).
- Logic: Artifacts in an automated deleivery container made available for scoring.
- External Connections, delivery, Are part of the automated scoring proces. Good care is needed. Becoming part of release management.
- The behavior of the model being a result of the modelling phase. Reports on lift, expected errors, reliability performance. Becoming all part of release management.
- Infra: Scheduling availablity for the development line adnoperational prodcution are different. Development-Test environmetns are smart shared.
- Infra: For performance reasons with differences in CIA (Confidentiality Integrity Availability) multiple production machines is an option.
Change ICT - Transformations
Dell Emc Cloud promotion
building yourself anything, everything is costly. The on-premise solutions have that association. Promoting an ICT transformation.
IT organizations need to be able to deliver an efficient, automated, and repeatable environment that is focused on the applications that are being built and delivered.
IT should not focus on how to get the various components of the infrastructure to operate together.
That is a change to the -AAS service delivery. Do not expect it to be easy when multiple "someones else" are running your environment.
These are high level considerations.
Business Intelligence & Analytics 👓 researching operations for innovations.
What is not there are: 👓 the details how you do it
👐    design: bpm
👐    devops: bpm
© 2012,2019 J.A.Karman