[ISM3 Users] Tuesday Insight: Environments

[Adrian Wiesmann]

Hello

> - The real world normally changes faster than you can change your
> model of the world. So they tend to stay out of synch.

Agree. Although changes should be noted somewhere. New people starting and
regulars leaving a company are managed in some database. Installed
software should be managed in some software or database. Networks should
be protocolled, systems should be documented. So isn't this only a
question of getting at the data?


> - I will repeat the main message from my original post: A human being
> can't be understood as a collection of cells. A company can't be
> understood as a collection of information system components.

I understood this sentence. I am only not sure if I agree.


> > - Everything in a company is an asset (cable, room, file cabinet,
> > people).
>
> I don't agree. Take me for an example. If you see Vicente as a set of
> organs and let's say, you remove the lungs...oops Vicente is no more.
> But scratch a few cells here and there, and Vicente is still sending
> mail...

Who decides which cells are vital and which not? And based on what
information are you deciding? To stay with your medical examples: Cancer
is known to mutate cells. How can we know we don't have cancer if we
don't look at every cell?


> If you have a Company and remove a switch...hey the Company is still
> there. A poor guy had to go get a less important switch and replace
> that particular one. Now, you send the Production Environment to the
> Mariana's trench...no more Company.

Again, who decides that the switch is not vital and based on what facts?
Probably there is some weird topology and all the traffic flows through
that single switch?


> Simple modelling (environments) takes less effort, and the value of
> the environment won't change wildly from one month to the next.

I completely agree. My point lies in the fact that we base our decisions
on incomplete data. As simpler as things get as more we construe the
situation of an environment.

> Tools are seconday. Good tools + poor ideas = poor results.

Good interpretation + poor data = poor results.


> IT: Logs show we can squeeze 99,99% uptime from standalone servers.
> That's 7 hours dowtime a month.
> Management: The business can't afford more than one hour downtime a
> month. IT: Ok, we will have to go for redundant systems then.

IMHO this is wrong. The management will start to say:

"We want not more than 1hr downtime per month"

and the IT will then look into the logs and say:

"Currently we have 7hrs, so we will need redundant systems"


> The big picture is the picture with less, more significant detail, not
> the picture with all the detail, including the details not wanted or
> needed to take decisions.

Agree, wrong word. I meant if we don't need the info about the cable for
the complete picture?


> Tools are not the solution, better modelling is.

I guess this sums our two positions quite well. You say we need
abstraction or reduction to cope with the data. I say we need more
intelligent systems so that we can cope with the data.


> Again, no. A doctor doesn't want to know what goes on with every
> single cell of yours. What he wants to know is how are you kidneys,
> lungs, etc.

No :)

Your analogy limps, as we say over here. My approach is not about looking
at every single cell. It is about getting all the data from my body.
Without interpretation and in full detail with the intent, that I can
filter out what I am not wanting to know: Heart rate, blood pressure,
components of blood, size of lung, rate of breathing, where in my body all
the blood is circulating, stress level, etc. So if I someday want to see
what risks I am having, that system could tell me that my veins will
soon turn into varices, I just need a simple query and will know the
answer, the data is all there.

Regards,
Adrian