March 9, 2022

s11e12: Forget about infrastructure, I just want my tea

0.0 Context setting

It’s 8:37am on Wednesday, March 9 2022 and I have standup in about 22 minutes. It’s a bright sunny morning following light rain overnight. And I’m going to hit fifteen minutes again today.

1.0 Some things that caught my attention

1.1 Forget about infrastructure, I just want my tea

Like I was saying last episode, with one of my clients I’m doing strategy work around their technology infrastructure (the whole “let’s do DevOps or even DesignDevSecOps or even ResearchDesignDevSecOps” and become a mature, frequently shipping valuable software organization), and one of the Jobs To Be Done is to take a look at their current applications and figure out if, well, there’s a better way of doing that.

Invariably there’s this whole issue of We Should Be Using Shared Services Across The Enterprise and no I will not make a Star Trek joke, at least not for now.

But what are those shared services supposed to be? What is the infrastructure supposed to be, other than “my stuff is working and I don’t have to yell at anyone so let’s just pretend it’s all invisible and everything’s good?” In that view, the infrastructure in the Old View is whatever hosting and “servers”, or Whatever It Is People Do On AWS or Azure, or worst case, it’s Whatever This Particular Application Uses.

But the allure of “shared services” makes sense to people who take a step back and look at all these different systems and applications and think to themselves: hang on, I’m paying millions of dollars for this stuff and even I can see that they’re all sending email and they’re all doing it in their own way, and maybe I don’t want to pay multiple times for this stuff to work?

And then they ask you: okay, what should this infrastructure be?

And you have to think to yourself: well, that’s a complicated question, and one that can only be really answered by understanding what it is that all of these applications are doing and how they’re doing it.

“I don’t understand,” say those people, and they’re being perfectly reasonable in saying that because really, why should you expect them to have knowledge about their application architecture if, say, they are Relatively High Up in the Hierarchy and honestly, it’s someone else’s job to know. So at some point, you should probably speak to those people and find out. But anyway, these people still want examples.

And you think for a bit and you say something like “well, you told me that all your applications send email and we know that they all probably send notifications, so you could imagine a notification service that all of your different applications could use and that notification service then abstracts away the business of sending a text message or an email or a mobile push notification so that your application teams don’t have to figure out how to do it or support it or manage it or even procure it, and everyone benefits from things like scale and saving money which on the one hand aren’t necessarily the best reasons for doing something, especially if in the service of scale and saving money you actually end up with something that’s worse for everyone. Which, you will not be surprised to hear, is something that happens!

But then you think a bit more and you realize, no really, this is quite hard and you talk to a bunch of engineers and infrastructure engineers and there are rough categories like “compute” and “storage” and “identity and access management” and a thing called “single sign on” that by definition needs to work across a bunch of things and maybe even if you squint someone might even say “oh, throw it all in a data warehouse/lakehouse/cubby/tinyhouse and be done with it and everyone can use that”, and that person might be from Oracle and you might want to question your life choices, at least for maybe five minutes.

But anyway.

They still want to know what “infrastructure” means and you’re struggling because you need something real and not something abstract because real generally beats abstract and helps people understand in the context in which they have to operate.

But maybe not this time. This time, you tell a story about making tea.

Here is a drawing I made:

And this is the conversation that goes with it:

Me: Hello, I’m British and I like tea and here’s my tea for this meeting! [Hold up tea]. It is good tea. It is in a Calamityware Mug and you should get one too, they’re great.

People: nodding and murmurs along the lines of “we do love a good English accent, just keep going please”

Me: Yes, it is rather good isn’t it. Just keep listening and pay attention, okay?

People: Just keep talking, we love it already.

Me: So let’s talk about what I needed when I made my cup of tea this morning. I needed a cup, which I have. I needed some tea. And I needed some hot water. The hot water meant that I needed, well, water, and the hotness meant that I needed in this particular case a kettle.

People: We love it. We know we’re on zoom so just assume standing ovation.

Me: Okay, so the kettle needs electricity and the water needs pipes. Those need wiring, and, well, someone’s got to make electricity, right? And the pipes, those have got to be machined! Which needs metal. And we need to move metal around, so let’s just say we need roads as well as, I don’t know, oceans and ships and ports. And then we need to do some mining, right?

People: I mean, you’re not wrong.

Me: Right?

People: We’re with you. Got it. Need stuff to make tea.

Me: But we don’t think about the roads and the electricity and the metal and the generating electricity and the power grid or even the tools needed to machine pipes or making all that PVC and all the oil.

People: No, who does that?!

Me: Me, obviously.

People: Yes, point taken.

Me: All I care about, all that I see is the kettle, the mug, the tea and the water, right?

People: We are familiar with this situation and if we close our eyes and we’re the kind of people who can imagine things visually yes, we can see those things.

Me: So the electricity and the wires and the pipes and the machining and the metal, they’re all infrastructure, right?

People: Yes, that much is obvious now.

Me: So one thing about them is that for the person doing the job at hand, they’re largely invisible.

People: Yes.

Me: And what they really care about is stuff like the tea and water and the kettle and the mug and whether I can get combine all three to make something that is simply sublime.

People: I mean you’re English, so yes.

Me: So when we’re thinking about what infrastructure might make sense to exist or create or support in your organization, it would help to do a diagram like this for each of those applications. And not even, really, just the “software application” but ideally, the whole thing about getting the particular job done, which in this case was sitting down to have this meeting with a nice, good, excellent hot cup of tea.

People: We’ve got you.

Me: The thing about these invisible things, this electricty and the pipes and everything also is that, well, lots of other things need them too, right? It’s not like it’s worth my while to go out and try to make my own kettle or dig up ore in my back yard and then learn how to refine it and construct some sort of CNC or whatever out of my bare hands. It’s not like this is Minecraft, right?

People: Ha, you made a Minecraft reference. That is a pop culture thing.

Me: Yes, it is. But the point is that the pipes and all that other stuff, because everyone else also needs them for Other Things that are not making tea (and therefore not as important), they’ve become commodities. They are not done from scratch by the person who wants tea anymore.

People: Aaaaaahhhh we get it.

Me: Do you?

People: No.

Me: Okay, here. You have this application. You have this team of developers or this vendor or whatever. And we don’t know if they’re doing the equivalent of digging up ore, refining it, moving it somewhere else, making their own roads and maintaining the roads to move that refined ore, and then turning it into pipes just to move some data through a pipe from one place to another.

People: ohhhhhhhhhh

Me: Would it make sense for them to do it?

People: Well now you put it that way, no.

Me: Especially if there were a way for Someone Else to take care of it, because you don’t need to do it custom or do it from scratch, right?

People: Because… then those people could spend more time on learning how to make truly excellent tea?

Me: Exactly.

People: Right, who wants to waste time machining metal when you could be making tea?

Me: You’ve got it.

And that, my friends, is your introduction to Wardley Mapping.

1.2 Some small things that caught my attention

Just one thing today.

Organizing the world’s information is hard, let’s go shopping!

So here’s a tweet about what Google Lens which is Google’s thing that uses lots of mathematics to make guesses about what is in a picture:

It’s about trying to use Google Lens to figure out where an unattributed photo of a woman being arrested at an anti-war protest in Russia, which is the kind of thing we used to do with reverse image search, but now we use something that has a proper branded Product Name. But the thing about Google Lens is that instead of helping figure out where the image might have come from, it “decided the right play was to try to sell me the woman’s coat” which is basically everything that is wrong and an example of stupendously misaligned incentives or, if you’re playing the other side, perfectly aligned incentives because clearly the technology behind Google Lens and the scale at which it operates would never have been possible without being funded by selling ads against what you might buy that machine vision hallucinates and guesses might be in an image.

Join me in just having a long, long sigh please.

---

OK, that’s it. I went over. I’m sorry. I didn’t think this through. Maybe there are some things I can’t write in 15 minutes and I would need to break the bigger pieces up into smaller episodes. Which, honestly, is a bit anxiety inducing.

How are you?

Best,

Dan