AWS celebrated its 10th birthday last week, and it has grown incredibly in that time. Lately there have been a couple of pretty big shifts in the cloud as well, and today some things happened for Apple.

I decided to include the Apple stuff at the last moment, because of the upset in the Apple vs FBI case mentioned at the end. Because of that it’s a fairly long note with a lot of long articles linked.

10 Years of AWS

On Monday the 14th AWS celebrated it’s 10th birthday. As someone who likes and uses the product (although definitely not for 10 years), I’m happy to see that it’s doing very well. There are a lot of interesting services they’ve introduced in that time as well and I’m hoping that they’ll keep on growing their product offering.

To cloud or not to cloud

Over the past couple of weeks there have been 3 major announcements concerning companies changing their cloud strategy.

First was Spotify announcing they moved to Google Cloud. Originally they had their own data centers, but they’ve now changed this in favor of running on Google’s Cloud services so they can focus on their core competencies. Then last week Dropbox announced that they had moved a lot of their AWS infrastructure to their own hardware in order to have more control over their whole stack.

Both of these are completely valid reasons for making these opposing moves. In fact, they both make a lot of sense for the companies involved. Of course, if the moves had been reversed they still would have made a lot of sense with these same arguments. That, I think, is currently the biggest issue with cloud infrastructure. There are good arguments in favor of it, as well as good arguments against it. For startups and other (potentially) rapidly growing entities using the cloud is the most logical thing to do. You don’t want to run your servers all over the world and then discover on the day of a big event that you don’t have enough of them so your whole service dies. Once you reach a certain scale however, there is a very good case to be made that having your own hardware is worth it.

Which brings us to the third announcement. Apple is going to use Google Cloud. Previously Apple relied on Azure and AWS for its cloud backend, and reportedly spent about 1 billion dollar on AWS in 2015¹. That’s not going to change though. While obviously they will reduce their spending on AWS this to me sounds mostly like a diversification strategy than anything else. Apple has realized that their cloud services are very important, and the whole idea of using the cloud is that you are more flexible. Which obviously includes not relying on a single provider.

Of course, if we’re speaking of a company whose scale should be big enough for their own infrastructure it’s Apple. In fact, they do have a number of data centers, and they are looking into running it all on their own hardware. That’s still some time away though, and if you read the article from Dropbox you can clearly see that it’s not an easy process. Which in a way might mean that the move to Google Cloud is a bit of a test case for this as well. I doubt that Apple is going to completely stop relying on AWS, Azure, and Google Cloud even after they mostly shift to their own hardware, even if it’s just as a backup system or for peak periods (after all, why pay for 10 times the amount of servers all year if you can just rent those when needed).

Apple updates

While technically this happened today, instead of last week, I didn’t want to let this wait for another week. Two things happened today that are related to Apple.

First, there was a product announcement. The new hardware introduced are a smaller iPhone and a smaller iPad Pro. Both are very interesting from a technical perspective, but as someone who is wondering if I should switch to the Plus size next² a 4" iPhone doesn’t hold much interest. Similarly I’ll probably pick up the big iPad Pro at some point, as I’ve found that I’ve started doing a lot more work on it lately so a bigger screen would be really nice. Which means that the hardware wasn’t all that interesting for me personally.

The software however is a different matter. ResearchKit and the new CareKit are very important from a health perspective, and it’s great that they are receiving updates. For me personally though, it’s the new iOS that is most immediately interesting. I’ve been looking forward to the release of iOS 9.3 because of Nightshift, the feature they copied from f.lux. Which I’m quite happy about that they copied, as something as important as this should not be limited to an app only technical people will install. This is basically a health feature, as it shifts the color tones away from the blue that impacts your circadian rhythms, which is the reason why looking at a computer/phone/tablet screen before going to bed has an impact on your sleep quality³. For years I’ve used f.lux on my Mac, and I’m happy to see this extend now to iOS with official support.

To be clear, f.lux was never allowed on the App Store due to its strict (and sometimes annoying) policies. Android recently did get f.lux, but you can only use it with a rooted device. If you’re willing to wait⁴ a similar feature is in the beta of the next Android version. The health benefits are well researched and the app is free. Of course, installing it on your computer is therefore also a good idea. Although, I hope that iOS 9.3 (and Android N) will be taken as an example and this will become a standard feature in other operating systems as well.

The FBI blinked

The second news concerning Apple is that the FBI has requested that tomorrow’s hearing is “postponed” while they investigate whether the solution a different source provided will work.

There are two ways to read this, first that somehow magically the FBI suddenly got access to a 0-day they didn’t have before. Or, that they realised the chances of winning tomorrow’s hearing were basically nil and that would set the wrong precedent. As I usually give people the benefit of the doubt, I’ll do so here as well. It seems to me that they did indeed come to their senses. It probably didn’t help that not only was Apple’s case stronger they also had all the people who know what they’re talking about on their side (including a former NSA director). Additionally it was becoming pretty clear that if the FBI won the engineers who were capable of building GovtOS would quit.

That means that the FBI’s options for tomorrow were either: Lose and set the wrong precedent, or win a hollow victory that still didn’t give them what they claimed to want.