Does someone know why Google Cloud can afford to be so much cheaper compared to AWS? Are the costs to Amazon for AWS so much higher than the costs to Google for Google Cloud? Why?
If you'll allow me to get a bit more wibbly: I've been an engineer at both companies, and from a technical perspective, Google is much better at this stuff. In order to provide their own services, they've been operating at much larger scale for much longer than anyone else. There's a pretty long record of the new hot thing in the industry being something Google's been doing on the order of a decade. The basic deficit in experience is exacerbated by Amazon's relatively poor retention of its best talent.
Putting all of this together, it is my expectation that Google be able to build better services faster, and to operate them more efficiently. AWS started with a massive first mover advantage, but my expectation is that in the long run Google will have little difficulty closing that gap and pulling ahead, should it wish to.
Correct, see how much profit AWS contributed to Amazon in the past quarters.
> I've been an engineer at both companies
Same here, so I offer my perspective.
> Google is much better at this stuff
Not much, but noticeably better across the board. Overall it's substantially better when combine everything together. But as others point out, the technical superiority does not account 100% of the quality of product offering. Being in 2 companies, I think you have similar experiences. Otherwise, GCP should already eat AWS in 2012.
> Google will have little difficulty closing that gap and pulling ahead
I think so. But my expectation is that Google needs operate Cloud as its top priority, and treat AWS as a tougher competitor than MSFT in the search/browser war. The thing is, that Cloud is a much much bigger market than internet ads. The race will be long and difficult to predict. Amazon is a much more formidable competitor than MSFT. They have proven to be the best in retaining customer and build a brand. They are also extremely fast in execution. And they are not afraid to try crazy ideas and abandon them if proven wrong.
Edit: posts further down from mine seem to suggest my sentiment is outdated and Google takes cloud support very seriously in their teirs. https://cloud.google.com/support/
The individual teams however have been fantastic, available on both Slack and Google Groups.
I don't know about that. Going further, majority of the spend is going to come from established businesses. They need software and not just infrastructure. And msft has those accounts. Has the reseller network. Has the software. I really can't think of any good b2b/productivity software that google wrote. Probably they haven't tried, but looking at what they announced last - data studio, survey and optimize - the talk was like they were super serious about it. The result is very, very far from the competition.
The only thing I can think of is the google apps for business (Gmail/Docs).
Also worked at both companies. Google is not better. Google is a retirement home for engineers. Few creative things happen there (anymore--wasn't always true). If GCP is so star spangled awesome why is it still so far behind AWS? Why did AWS start literally 10 years before GCP?
Google reacts to competition while Amazon proactively responds to customers -- this, as always, is the difference that allows amazon to succeed.
That seems to answer the first question. AWS had a multi-year head start, but GCP is catching up and has the fundamentals covered, and the primitives they offer are very strong.
A "bias" disclosure would be helpful here.
(work at Google Cloud and don't get paid to post on HN)
Totally possible I didn't get to see the full benefit of the culture there. I have however seen friends disappear into google over the years and emerge with a distaste -- obviously the same can be said for amazon. A constant bias disclosure is cumbersome but you're right that it would have been helpful, you've forgotten it on some of your posts as well.
My comments were off base sorry, obviously google does a ton of cool stuff. I let my emotions around this ridiculously biased post get in the way of my reason.
You also tend not to address the merits of the post, instead using ad-hominem and misdirection. I'm actually noticing a pattern here :)
And what are you doing?
I certainly am not in a habit of disparaging my employer's competitors' cultures and technologies through anecdotes (the last ex-AWS bit was the notable exception in the heat of the moment).
I urge both you and "ignoramous" to get into this habit as well :)
(work on Google Cloud)
Basically, they can do a better job than Amazon predicting current and future workloads. As a result, they can more efficiently place workloads and literally do more with equal or less hardware. That and the fact that they have been building their own hardware better means their hardware is also very likely more advanced. Those together mean they can do the same thing for cheaper and still turn a profit.
Amazon optimizes for doing something first and doing the most. Google optimizes for doing things smarter and as a result often does them cheaper. However, it also means that Google provides massively fewer services than Amazon.
They have better hardware utilisation and a lot of their infrastructure was fully-depreciated before they opened GCP.
Unlike AWS, Google couldn't design systems assuming a customer would bear the price of any moderate inefficiencies. So they've had more design pressure on cost from the beginning.
I also suspect their cost of networking is much lower -- they own their own fibre network and a lot of it will also be fully-depreciated.
From Amazon however I have even heard about them forgive bills where the customer was really to blame.
All this is hearsay though and I don't think Amazon does this out of the goodness of Jeffs heart but as long as enough people think this is how it works it might be part of why people will use AWS even if it is more expensive.
I'll also comment that Google's Eng/PM org is very engaged. It's not unusual for an engineer in charge of a service to help resolve a customer an issue.
[0] https://cloudplatform.googleblog.com/2016/10/introducing-a-n...
(work on Google Cloud but do not get paid to post here)
This is not related to Google Cloud, but to GSuite: Or, if you get an engineer on the phone, through the support number, they just insult you, and hang up on you, as has happened to me once (back when the free tier of GSuite was still a thing).
I know Google support spends gobs of time quantifying customer success and happiness, gets measured and rewarded by these numbers. Nature of support anywhere is that sadly you'll always have some folks with a bad experience, all you can do is try to make it right by all.
This is the norm with AWS, in case you didn't know.
Like someone pointed out, having a deeply engrained customer-focussed culture is different from having a massive support org that quantifies customer happiness as datapoints on a review board somewhere in cozy SV offices.
I am also not going to get into a discussion with an AWS employee about culture health between the two organizations. Both are trying to do it right for the customers, and both have unique challenges, regardless of the amount of coziness of their offices :)
It's unfair to paint Google as "out of touch" and "not customer focused", just like it is unfair to bring up horror stories about Amazon culture, especially if you're employed by a competitor.
Apologies, again, but are you an AWS employee? I don't want to inaccurately label you as such, but your posts hint at it.
(work at Google Cloud)
I work for Amazon, yes.
I do assure you that Google Cloud invests a whole lot of time and money thinking about customer happiness, as should all organizations. In the public cloud business, customer happiness is everything.
In the last three days I have been able, in my day job, to personally ask a Google PM to pull Google Cloud engineers into a shared slack channel to help.
It took us longer to set up the slack permissions for the googlers to join than it took for Google folk to identify who to pull in.
We are also a major partner of Microsoft.
We've also been on AWS, spending a shit ton of money, for a long time.
And as whelmingness goes, my personal experience with AWS has remained squarely in the "under-" camp.
I am not an executive, or director, or anyone with any political or financial pull. I am a plain engineer. Yet Google folk came to our aid, quickly, when we needed it, without needing to stump up buckets of cash or wait for someone somewhere to take our ticket.
I wonder what you should be disclosing.
Assuming that this is true, having no customer support is still better than the AWS support ;)
--
I do not want to draw any conclusion here but when it comes to which company gets my money decision I always prefer Amazon.
The following applies to customer trust too:
"It takes a lifetime to build a good reputation, but you can lose it in a minute."
― Will Rogers
Under the Obama administration we didn't see any antitrust/antidumping actions because of how much Google managed to embed themselves into the administration. Don't know if it will change.
EDIT: fixed a typo
In my view, there's really only two things that can reduce cost enough to make a difference: 1. Reducing energy consumption, 2. Reducing drive / component failure.
I believe Google may be better at both of those than Amazon.