Devops Improving Business Solutions
Change, improve business processes.
Information Management.
BI, Analytics is challenging.
Having both ALC type 2 and type 3 in use, behaving different. Requirements differ from the operational line.
Activities for some parts of the operational line are only possible to solve witht information management.
🔰 Too fast ..
previous.
Contents
| Reference | Topic | Squad |
| Intro | Information Management | 01.01 |
| Run&Improve | Business Improvements | 02.01 |
| ItSm | ICT governance. | 03.01 |
| missing PoU | Missing operational Tasks. | 04.01 |
| Operational IM | Operational Information Management. | 05.01 |
| What more | Supporting business, implementing improvements. | 06.01 |
| | Following steps | 06.02 |
| Install | Installing BI - Analytics tools. | 07.01 |
| Running | Running BI - Analytics tools. | 08.01 |
| Government | ICT AI BI governance. | 09.01 |
| IM internal | Information Management, operational internals. | 10.01 |
| IM external | Information Management, operational externals. | 11.01 |
| wrong BI | Confusing, misunderstanding, hidden goals. | 12.01 |
Progress
- 2020 week:37
- The important water strider from agile ideas includes.
- Adding a split in content.
- 2020 week:03
- Changing the first failed one.
Rework on old pages and infographics.
Business Improvements.
The world is changing. Having all activities goals and results tied to that you have to change also. Not adjusting to other conditions can leave you isolated.
For an enterprise or other organisation that has the meaning of non existence.
Continuous Delivery
These words are used for changing software. The idea is that delivering software for business processes faster will solve business problems.
What is needed to be delivered in a continuous flow is the inventory, elements of information needed in an assembly line.
I am seeing an administrative process flow in the same way as an assembly line.
The core business processes are needing regular maintenance and they are needing improvements in:
- new products, replacing old products
- How the assembly line is working
- financial profit optimisation
Software and the applications are building blocks part of the assembly line.
Threats with Changes
Serious problems: doing any not verified change in the assembly line, abandoning an old product when the new one is not ready.
The usual change management process in place (ITIL) does not have a vision for change categories.
Change categories are:
- Not tested, not verified changes. Internal product life cycle
- Certificates licenses. Needed in time, dependencies
- Required security fixes. Ad hoc by external events, urgency
"The road to hell is paved with good intentions."
Stable Operations
Any operational process with a well structured delivery is like an open highway in use.
2 predictable repeatable behaviour for basics
3 defined global knowledge what is going on
4 Measured Managed behaviour
5 Optimised, planned defined controlled improvements
The maturity level is not always what the involved people assumes it would be.
ICT governance.
ICT is used for the operational process (shop floor, assembly line) and is used for analysing processes (BI analytics).
The way they are managed are slightly different because the impact at incidents is different.
Operations core business
ICT service for the operational line is nicely described at:
ITIL in 2018 - cloudy days ahead? (InsightaaS an analyst firm) 👓
Relevance of ITIL to modern IT.
Modern IT requires an expanded range of management capabilities as is illustrated in the diagram below (this is not an exhaustive list of topics, however!).
Three levels of management need to be considered:
- managing the systems from multiple providers,
- managing the services that combine multiple components,
- managing the endpoints (users and "things") that access the services.
ITIL currently addresses some parts of each level.
Operations with BI & Analytics
The ICT service for BI, Analytics is having the challenge of supporting another type of life cycle with a different set of requirements. Those are adding:
- Technology blindness,
- Solving Lack of reproducibility and reusability,
- Solving Lack of collaboration.
- Solving Suboptimal architectural decisions.
Missing operational Tasks.
There are some necessary tasks to do that are impossible for the operational assembly line or for just a part of analyses to do in an ad hoc approach.
The glue in a value stream is having all artefacts that are needed in time available for assembly. In an administrative process that is having all information needed available.
Preparations at processes.
The physical assembly line is not really different to an admministrative one.
Just imaging the artifacts coming in as adminsitrative requests. 👓
With this PoU function in an administrative process the role of IM, Information Management, at the operational level is becoming a clear defined one.
The primary objective of the point-of-use provider is to keep the workers supplied with material. Hence the workers can focus on the work and do not need to leave their post looking for material.
Therefore we achieve the separation of cyclic work (the workers) and non-cyclic work (the point-of-use provider) to improve efficiency.
First of all, do not create an additional kanban loop between the supermarket and the assembly line. The effort would heavily outweigh the benefit, making the whole idea pointless.
Instead, the point-of-use provider is close enough to the line to keep an overview about what is needed.
Rather than handling additional kanban cards, his job is merely to fill up the material at the line to its target levels.
Operational Information Management.
There are some necessary tasks to do that are impossible for the operational assembly line or for just a part of analyses to do in an ad hoc approach.
When connecting information that is under control by different parties and with different responsibilities, coordination with support on high level is needed.
Preparations at processes.
Administrative lines are often more segregated ones in dependent chains. For worse when there are segregations in organisational ones with different responsabilities different goals.
Going to have multiple lines the PoU is more and more getting out of sight for each line. That could be an advantage when they are related as chain dependency. 👓
Many point-of-use providers are circling around one line, with the material being supplied to a supermarket close to the line through a milk run.
If the line does not need much material and the worker has a lot of idle time, he may be able to take care of the material supply for more than one line or cell at the same time.
However, it is also possible to have a point-of-use provider that gets the material from a central warehouse.
This is illustrated below. However, this significantly increases his walking distance.
In these cases, it is highly suggested to give the point-of-use provider additional tools to help him transport more at the same time.
The PoU fucntion getting an IM, Information Management, task at operations solves some common challenges.
Those are for example:
- Who is responsible for information coming in from external but needed at processing.
- Who cares for issues when the assembly artifacts are not consistent tot process.
- There is point in communication where problems and options to improve.
Supporting business, implementing improvements.
Improving business is requiring to know what is going on. There are no generic recipes for that, every organisation has his own challenges to solve.
Patterns at the operational level of Improvements.
There are no detailed information about any of my experiences. Writing details about an organisation would show no responsibility.
With all experience there are many stories and topics that went wrong and could have done better.
This page is describing operational and practical things executed inside an organisation. Becuase of this it doesn´t have patterns of me.
Anonymised patterns are external services, found at &auotdata" and "meta". There are several views on those.
Choices and the effort to implement are to evaluatue at the paatern chapters.
Duality service requests
A lot is outsourced or bought from others. Contacts with external parties for improvement of the internal processes is a special task.
That special task has no clear position but the one that would make most sense is "Information Management (IM)" in an operational setting.
Intermezzo 👓 and elaboration for what I couldn´t classify in the six basic topics.
Getting in external solutions at the operational data and meta level is shopping in an own context.
That context is by goal and size.
Used figure to get an association when used data, meta is used in bi analytics or at the system development life cycle.
sdlc: The view and focus in the request is on the core business process.
bianl: The view and focus in the request is on analytics getting information on improving the core business process.
Association for the size of an organisation doing some choice, used figure.
bpm tiny: the internal organisation is relative small.
bpm big: the internal organisation is relative big.
Following steps
These are practical sdlc experiences.
Data Information generic - previous
sdlc , development life cycle 👓 next .
Others are: concepts requirements: 👓
BPM SDLC
Bianl
Technical Implementations.
Installing BI - Analytics tools.
Running tools that are just only used for optimising the cored business (sdlc) have little impact on the core business itself.
The high risks are using data, information, that is very sensitive. It could be personal data or technical knowledge.
Steps for installing tools.
The first question is where the segregation, decoupling, in responsibilities will have result in a logical borderline.
What are the providers tasks and where are the tenants coming in. A multi tenancy approach looks complicated but when considered from start will simplify things a lot.
Having many tenants supported in an ad hoc way eg using dedicated (virtual) machines for each of them will be more complex. The outsourcing argument for having doing it all by someone else is underpinning this.
Management may also need to be coordinated across multiple owners and operators.
For instance, a service provider´s infrastructure management system and the cloud customer?s application management systems should be connected (see the diagram below).
Running BI - Analytics tools.
The second question is how to implement and improve the service. I don´t agree with the article.
That building something that is not needed and is not able to run well operational is the new agile way feel like falling back into ad hoc, chaotic.
Steps for running tools, using information (data)
Getting the inforamtion (data) can be very hard. It is knowing to be the most time consuming step.
There are several difficulties to overcome.
They are
data and
metadata patterns found at other pages, not unique to an organisation.
CheckPoints:
- searching and finding usefull available information
- getting access to information once for a proof of concept research
- Limiting the needed data while knowing what is available
- getting access to information internal and/or external a regular supported way.
- determining when the information gets obsolete and whether archiving standards are required.
- Having a regular archived environment and a legal audit one if they are expected to be usefull.
- defining an easy to understand explainable acces model that is supported by regular staff.
- have auditing and logging in place that is supported by regular staff.
The diagram below was used in ITIL 2011 to illustrate the relationships between ITIL´s core guidance topics.
The implied model is functional and siloed, i.e., a model based on a sequential plan-build-deploy paradigm.
Today, "mass collaboration" using agile methods and team-based DevOps projects is changing the management model.
Doing this cycle in small steps of small parts and uning the V-map model for that foloowing a defined strategy can result ia real agile as it is intended.
ICT AI BI governance.
ICT is used for the operational process (shop floor, assembly line) and is used for analysing processes (BI analytics).
The way they are managed are slightly different because the impact at incidents is different.
Operations core business
Good old CRISP-DM is too limited, an attempt to get it more structured is an IBM solution method (Asum) 👓
ASUM-DM is part of a more generic framework called Analytics Solutions Unified Method (ASUM) that provides product- and solution-specific implementation roadmaps covering all IBM Analytics products.
I prefer my personal one with all life cycles and dependencies.
Information Management, operational internals.
There are some necessary tasks to do that are impossible for the operational assembly line or for just a part of analyses to do in an ad hoc approach.
When connecting information that is under control by different parties and with different responsibilities, coordination with support on high level is needed.
Data exchange internal connections.
An internal connections, data pipeline, is having the advantage of an visible useable road.
Content, meaning of artefacts, delivery moments all is under control in the same hierarchy.
There can be a distance. The work to get access by a service agreement being hard work.
💡 Support from the hierarchy and having some place where people are managing this should make this well workable.
Information Management, operational externals.
💡 Support from the hierarchy and having some place where people are managing this should make this well workable.
Decoupling will add some additional cost of maintaining copies.
Optimised support is possible by a shared staff. Ad hoc level support is not coöordinated or shared.
Data exchange external connections
An external connections, data pipeline, is having the disadvantage of not having a clear bridge and clear destination.
It is adding complexity on data versions, interface versions, delivery moments, content and artefacts meaning.
Usually there are other data consumers at the supplier. All this results in a lot of dependicies to manage. Problems will arise when there conflicting goals.
Missing the information security management processes (ISM).
Defining the data connection line between information assembly units and analytics as very important.
⚠ The support for designing implementing and monitoring doesn´t have a clear postion.
It doesn´t belong to the operational process neither to financial and generic (bpm) line.
💡 Left is a place the security line along wiht the analytics, information management, business improvement line.
This makes sense:
- shorter control & communication lines for design implementations audit
- A SOC (Security Operations Center) will do a lot of analytics and BI
These are many times the real blocking ones needing some bending to make things workable.
Confusing, misunderstanding, hidden goals.
A tool is just a tool but when coming in with a cross sell approach some strange human politics can evolve out of that.
The battle of BI & Analytical tools.
⚠ Which tools should be used is changing in time, learning what is working and what is not effective by experiences.
The world of BI and Analytics is a challenging one.
Nobody is happy with today´s situation and nobody wants to change.
Everybody wants to change the work of someone else. Arguments by personal reports using dashboards.
BI tools used as part of te core business assembly line
This is often practised because the visual presentation of BI tools are often the best of what is available.
The easy pitfall is: usage of the BI tool is creating dependicies with the life cycle of that tool.
💡 Decoupling using a duplicated package will solve that.
The tool possible gets confiscated by a single department.