The marketing writes itself.

https://www.youtube.com/watch?v=6QI4KmlcQr4

Microsoft platforms most people think are long dead and buried still get maintenance updates years later.

Guess what? It has been shut down by the end of 2019.

We're back on AWS now. We were (I guess) not an insignificant customer, paying around 8k €/month, with the prospect of growing significantly (which has happened in the mean time, but on AWS).

They still have the GDPR-compliant Frankfurt and Berlin datacenters. [1] I've not see any announcement about closing them.

[1] https://azure.microsoft.com/en-us/global-infrastructure/germ...

Apparently they reversed their decision again, which kind of underlines what I said :)

No one said MS developed new services for years and years and years; they said they supported existing offerings for a long time, which they do.

https://hackernoon.com/microsoft-announces-the-end-of-the-ge...

https://github.com/MicrosoftDocs/azure-docs/issues/14944

If you still don't believe me, please provide your public key or email address so I can publish a response.

Maybe it doesn't stand out: Microsoft Germany delivered the promise that US authorities are technically unable to access stored data. The "new" data centers were without that promise, making the USP obsolete.

> Existing customers can continue to use the current cloud services available today.

> Existing customers are not affected by this and can continue to use the Germany cloud without restriction.

You don't see any Windows Phones being sold, but they still years later being supported with security updates.

No one can possibly fault a business for stopping continuing development of an unpopular offering. It's when it becomes unsupported and leaves people stranded that the the real problems happen.

It’s a matter of visibility and what kind of customers own the products being shut down. From a PR perspective there’s a big difference between killing an enterprise product with few large customers, and killing a consumer product with many small ones. Even if the few complaints are louder, the many complaints around the latter would still totally drown it out. I’m sure they have analytics to know how many people still use Windows Phones and weighing the possibility of an outcry e.g. if all their unsupported phones got 0wned and they collectively took to Twitter about it.

They have a great platform, but I think they've just lost the war by the top brass waffling on this. That's a shame. There could have been a lot of money in third place, and as long as you're in the game there is a non zero probability of moving up the ladder.

The vast majority of startups do not succeed or fail based upon their ability to control their datacenter spend. They succeed or fail based upon their ability to attract customers and grow revenue. Colocating and managing your own servers instead of taking advantage of the ease of using the cloud will only be a distraction for almost all early stage companies.

[1] https://www.bloomberg.com/news/articles/2019-12-03/amazon-tu...

Unless you have some truly massive amount of hardware (in which case it's _definitely_ cheaper to hire someone than pay Amazon), you can generally just set it up and forget about it save for periodic maintenance and emergencies.

So what you really need to pay for is the time of someone who can do these things for you and funnily enough, such a service exists. I believe it's common in the not-startup world: a managed services provider (MSP).

And in the end this really depends on what the business is actually doing. AWS is _really_ price inefficient but in particular for bandwidth. If you're doing anything bandwidth-heavy like video, images or backup, you want to keep the hell away from any kind of cloud.

If Jeff Bezos sent an email to the company saying "We want retail revenue to increase by X% by 2023" would you assume that they'd shutter the whole thing if they didn't meet that goal? Of course not.

Why is a goal and strategy for Google treated differently?

Google migrated from a weird, Apple like "come buy stuff from us, but first solve this riddle" sales process with reps with huge territories to a more traditional enterprise model. Based on Linkedin updates, the Oracle class of 2015 has basically moved in on the sales side. Those guys usually make premium bank, so they need to bang out deals or GTFO.

And the comments upthread said that GCP would be terminated instantly if that target isn't met. You genuinely don't see the disconnect in that interpretation?

Things like Google Reader had healthy user bases when Google came for them in the night.

They could even sell most of the products they murder to someone else for millions.

I wonder if anyone in Google management ever looks back and regrets that decision?

It's still always being brought up, even 6.5 years later.

Most sites have RSS feeds, even the heavy hitters like YouTube.

I couldn't actually read the news in it anymore.

It would have cost Google more money to sell Reader than anyone would have bought it for.

It's hard to even imagine how it could have been split apart from the mothership at all.

No. The technology died. Isn't this clear evidence that the product was dead in the water?

These are not really comparable worlds.

If I also thought it was unlikely that they would hit their target I would be really worried (assuming of course I had any reason to actually worry about how Amazon's business was doing)

If we were talking about a company that had a history of shutting down successful divisions because they were not successful enough I would think "hey 2023 is when the shutting down process is likely to start"

You can't predict the future.

Just build your workloads to be cloud agnostic. If prices ever change, then you can move your workload wherever it makes sense at that time.

Locking yourself into a colo architecture is just as risky as locking yourself into AWS.

That's the game theoretical predicament: Colo's can take advantage of all this and increase their prices as well.

We just don't know how this whole thing will shake out. Given that uncertainty, I think you're proposing the only reasonable way forward for a little guy.

Be agnostic. Meliorate your market position by differentiating in some key area of your business where you can actually secure a competitive advantage.

The only "duopoly" will be in terms of the hyper-scale size of such providers, but very, very, very few customers actually need such hyper-scale. Not to mention, you literally get to pick and choose how much of their tech you use, limiting any potential risk of lock-in. Not to mention, the hyper-scale providers are probably the most expensive ones!

I think this is all very over-blown. AWS's services are fantastic for scaling up services quickly, and their customer focus is great, but if they were all gone tomorrow we'd still be doing the same things as before, just slower and with less reliable infrastructure.

There are literally hundreds of services offered by AWS and Azure outside of what you are thinking of "rented servers". These things cannot be replaced by "managed hosting". Not that they would be slower and less reliable, these services don't even exist outside the cloud providers.

The most "cutting edge" things I remember announced at re:Invent 2019 were machine learning and AI stuff, but it was all stuff integrating with existing 3rd-party services and products. They know you want to be able to work with somebody else's tool, and they want to make it easier for you to spend money using that tool on their systems. The bulk of AWS's income is still EC2 (and data transfer); the rest is window dressing to keep you from looking elsewhere.