Simon Wardley likes a good rant on twitter occasionally (they’re often quite fun). He’s often rather good at it too. A recent one caught my eye the other day. Click on the tweet and you’ll see the follow ups that go with it (there’s quite a few) to get the full context.
The really interesting thing about this for me is that it’s even a question. I don’t genuinely understand what the issue is with serverless for other people.
Except maybe I think I do.
And I think so does Simon.
Simply put, Serverless is disrupting the status quo of software engineering practice in quite a fundamental way.
And when you change the status quo, people don’t like it.
Unfortunately, Serverless disrupts the status quo on a number of levels.
It disrupts development frameworks (it changes/removes them)
It disrupts DevOps to become DifferentOps or LessOps (we’ve only just got used to DevOps, so this one is very much disliked)
It disrupts development team structure (people have to be aware of more than just their “bit”)
It disrupts the business model (maintenance considerations are changed so your model has to take that into account)
It disrupts your hiring (you can’t just hire a “<name of framework> engineer”)
And the problem is, that when you disrupt on this kind of level, you really do disrupt people’s jobs, and therefore their salary comes under threat due to possible and unknown future changes.
But maybe, the software development community needs this disruption.
Enterprise v Startup
Over the past few years, I’ve become more and more aware of a disconnect between how startups “do tech” and how enterprise organisations “do tech”.
UPDATE: This is a generalisation! There are some amazing enterprise organisations out there doing brilliant things with serverless.
And this is almost entirely down to one word.
Enterprise organisations hate Risk.
So they will do everything they can to minimise it.
And startups are all about Risk.
And they actually will do everything they can to minimise it.
But that’s not always straight forwards in a startup.
You have to calculate your Risk more carefully.
Which is interesting. Because I sometimes think that Enterprise organisation are so bothered about Risk that they will mitigate by throwing money at it.
Whereas a startup doesn’t.
Which is where Serverless comes in.
Because Serverless mitigates Risk by spreading the Risk to more services.
And those services (if chosen well) are often provided at a significant reduction in cost to “build your own”.
Which makes it a great idea for startups.
Because you can limit the Risk to the problem your startup is trying to solve.
Which is a good thing.
Which is why, when Simon rants about it on twitter, saying that it’s rather simple, I think he is trying to state something like the above.
Serverless and the Enterprise
Serverless should bring in a new era of tech for Enterprise level companies. It should bring about a highly cost conscious and greater ability to assess and manage Risk.
It should, but it won’t.
Because of the naysayers.
We’ve had them for years (and years and years).
Every single time a new technology comes along, the Enterprise basically sits and watches, until someone else has taken the Risk and then they’ll hire a consultancy to implement it for them.
I would make the point (and so would Simon) that, to be honest, that’s just way too late.
Serverless is here. It’s ready. It’s imperfect for sure, but it’s here.
And it should give very very good opportunities for mitigating Risk and sharing the Risk profile of a project for Enterprise.
But it won’t. Because they won’t jump at it yet.
But someone will, and when they do, those of us who have jumped in a while ago will be saying “I told you so”.
And that should mean that services should become cheaper over time.
And that consultancies working with Enterprises should not be able to charge as much (!) for delivering development for these companies.
Which would be a good thing all round (except for the consultancies).
And should lead to greater innovation I think.
So, my suggestion to Enterprise is to just get on with it now. It’s not hard.
Just ask me!
Everyone should join the Serverless Revolution.
Onwards and upwards!
PS I never thought a Serverless approach was going to be this good for our business, but as far as I’m concerned, this should be the first thing you look at for all greenfield projects. If not, you’re going to miss the boat.