Decentralize your stuff folks. Or it will go down

@jens Who would have thought about that back in the days when fastly was founded. :D

@mzumquadrat Well, back in that day I was relatively fresh out of kick-starting the streaming revolution with peer-to-peer technology (TL;DR, I have written about it before in more detail), so I would have absolutely thought that.

Bandwidth prices started dropping substantially around 2008, which is when CDN prices dropped, and they started becoming an indispensable part of the web for many people. So the trend towards too much centralization was visible when Fastly was founded.

@mzumquadrat In business, there's this rule of thumb that anything that is core to your business should be invested in via capital expenditure, anything that is not core should be operational expenditure (aka renting third-party services).

In management terms this means that when opex suddenly looks significantly cheaper than capex, you start questioning what is core to your business. Arguably in streaming services it's the tech to get streams to the end-user, but that may be a tech PoV.

@mzumquadrat As an aside that ties in with my quick #docker rant earlier or whatever is going on with #kubernetes - open core is very similar to an "embrace and extend" strategy in that it provides a way to hook into an ecosystem (dev community) and extract value from it until you're in a dominant position and can force the value extraction.

In that kind of world, the only reason to invest anything at all into tech is to be able to do this.

Otherwise it's a distraction that opex "solves".

@jens Thank you for that explanation. That makes sense i guess. But wouldn't the expenditure stuff be weighted against the risk and loss of money due uncontrollable downtimes?

Sure it is nice to not host your CDN yourself. But it comes with the risk that i have to trust another entity that they won't fuck up. And this risk has to get a price tag. And if this risk has a higher price tag than what i am saving with outsourcing it is not worth the risk.

@mzumquadrat We're cross-posting a bit, I hope you're catching my other replies :D

I agree with you!

What I'm driving at is that from a certain point of view, offloading effort to a CDN is a good business decision. What I gave was the overtly stated reason for it, and it's good enough on the face of it.

There's a secondary reason why this kind of decision is attractive, and that's executive bonuses. Bonuses tend to get tied to profits; profits rise when costs go down; costs go down when...

@mzumquadrat ... expenditure is down.

Capex has the downside of front-loading all the expenditure for long term benefits. Opex effectively spreads expenditure out, which means *right now* the costs are lower, so profits are up, so bonuses can be received.

The potential for having to buy yourself out of vendor lock-in or similar issues is dismissed as just that, a potential, some chance that it may happen, but probably to the next CEO.

@jens oh yeah, the short sighted, bonus driven economy. I heard about that. Not an economy i like, to be honest.

@mzumquadrat Neither do I!

The capex/opex rule of thumb isn't terrible per se. As a small startup, keeping your expenditure focused is good.

But IMHO it's the bonus driven economy that keeps the rule in place perpetually, where a strategy to do more capex for longer term benefits would be sensible.

It's not just bonuses, but also shareholders. They also benefit only when profits are up (short-term).

@mzumquadrat Part of why I want to turn my #interpeer project into a public interest company to support it is this complex of reasons.

I need money, but don't need stupid amounts/bonuses (would be difficult under public interest), decent statutes should not only hold me accountable but ensure the project's future, and at the same time there are no shareholders to please because profits cannot be turned into dividends.

I may miss something, but it often feels like a good hack of the system. :)

@jens I really hope that people are more careful now and are better weighting the pros and cons. Thoughtful and informed decisions are always better than some shortsighted, follow the fad ones.

@jens If people were willing to look behind the veil i think. A lot of people saw the opportunity to offload alot to the CDNs. And i think today a lot of CDNs are missused because people tend to solve inefficiancy problems by offloading their problems to the CDN.

Sign in to participate in the conversation

Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.