Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I think GCP is better than AWS (2020)

nandovillalba.medium.com

Post reply on HN
Print view View original

271–280 of 409 posts

Re: I think GCP is better than AWS (2020)

#271

Post by danielmarkbruce »

white_dragon88 wrote: post #250
danielmarkbruce wrote: post #180

Earlier quoted context omitted.

What folks at google think and what they say publicly aren't the same thing. It's naive to think otherwise. Look to what actually makes sense for them to do and you'll have your answer. Internal folks knew google would walk away in a few years if it didn't work. Stadia was largely a group of people out of the failed and left behind Daydream (VR) project.

"Internal folks" who gives a rats ass about what they think? A company's reputation has little to do with how people on the inside perceive it, its actions and track record do the talking. Stadia was a classic reason I couldn't trust GCP or any Google product aside from Gmail cause I've been using it so long, because unlike you I'm NOT in the know, I just see a bunch of stuff spun out, hyped (calling it 'little stuff…

Agree, nobody cares. But it was pointed out in response to a suggestion it's revisionist to say it was an experiment. It was correct (ie it was an experiment), albeit not understood by people who believe at face value what companies say.

I haven't been in the know for years, but I learned enough - don't look at what they say, look at where the money is. It's the same for almost every company.

Product strategy is a mess at most companies. The developer experience sucks for most products. Stripe built a multi-billion dollar business off of developer experience because the standard experience sucks so bad.

Re: I think GCP is better than AWS (2020)

#272

Post by joseph »

jb_gericke wrote: post #249
crazygringo wrote: post #115

Earlier quoted context omitted.

I think this is an unfortunate conflation of free Gmail/etc. consumer products with GCP. All of what you say is pretty accurate for free-tier Gmail/etc. No customer service, risk of an account getting blocked and nothing you can do, new apps not finding success and getting cancelled. But none of it is the case with GCP. Customer service is great and you can reach real people easily, paid accounts aren't getting shut…

We had our main GCP account suspended because we were running a Lightning node, and some Google automata flagged us as mining. We couldn’t get hold of any actual person at Google, and were told by our Google reseller to buy a fairly expensive support package to have our issue expedited after raising multiple appeals/objections. Suffice it to say, we run out of AWS now. I’ve heard GCP support (in terms of reaching an…

This is kind of a problem with non-paid support I think. I had an AWS account with some credits that were going to expire and decided to use them to spin up a bare metal instance. My account was immediately flagged and I got locked out. I finally managed to reach someone who made me jump through hoops changing my password and rotating my API keys. It still took a few days before I was able to log in again, and by then my credits were expired. Luckily I wasn't doing anything important in my account, but someone running a business would have been screwed.

Re: I think GCP is better than AWS (2020)

#274

Post by clhodapp »

Alupis wrote: post #135
crazygringo wrote: post #124

Earlier quoted context omitted.

Ugh. So it was first reported that Google Domains was being abandoned, but Google Cloud Domains was still just fine. I thought -- exactly, you can count on their paid enterprise cloud services. I looked it up now to reconfirm, and it turns out a few days later it came out that Google Cloud Domains was also being shut down? [1] WTF. OK, well that's idiotic. Maybe I have to take back my parent comment then. That's one…

Google Domains is not being shut down - it's being sold to Squarespace. I would assume Cloud Domains either went with that package, or is being removed as part of the deal. Selling to another company != shut down in my book. Although I'm already in the process of moving domains to another registrar...

I don't see the point of making this distinction.

First, selling it to another company is just shutting it down but doing you the courtesy of migrating the data to another vendor. It still results in all of the management UIs, integrations, and APIs that you rely on going offline. It just avoids creating an immediate outage (but it will create one eventually, unless you start interacting with this other vendor).

Second, the point has repeatedly been made that GCP is fundamentally different from Google's other offerings and essentially immune from being shut down (even in the responses to this article). That may be true for GCP as a whole, but this proves that it's not true of the individual sub-components of GCP, even ones that seem extremely core.

I actually agree with the article that GCP is just good in ways that competitors aren't but... Running on their platform does continue to be a bit nerve-wracking.

Re: I think GCP is better than AWS (2020)

#275

Post by LargeTomato »

simoncion wrote: post #184
mvdtnz wrote: post #165

Earlier quoted context omitted.

Google very clearly stated in no uncertain terms that Stadia was a real product that they were committed to for the long term. Stating, after the fact, that it was "actually" an "experiment" or a "random little thing" is not just dismissive - it's pure revisionist history.

Yeah... Google employees just _do_ that sort of shitty redirection, and it's obnoxious as fuck. Maybe it works in a "We're too polite to disrupt whatever political game you're playing in the hopes that you'll reciprocate when we splash obvious bullshit when we play ours." snakepit that seems to be most of Google, but out here, we're unchained by that political nonsense. Bonus chatter: We get that you'd like die for y…

You were just able to articulate what I hated about Google. Thank you.

Re: I think GCP is better than AWS (2020)

#276

Post by bruce511 »

pknomad wrote: post #121
crazygringo wrote: post #115

Earlier quoted context omitted.

I think this is an unfortunate conflation of free Gmail/etc. consumer products with GCP. All of what you say is pretty accurate for free-tier Gmail/etc. No customer service, risk of an account getting blocked and nothing you can do, new apps not finding success and getting cancelled. But none of it is the case with GCP. Customer service is great and you can reach real people easily, paid accounts aren't getting shut…

> I think this is an unfortunate conflation of free Gmail/etc. I would also say that this is a byproduct of how google treats its customers, even on paid platforms. I used to manage Google Workspace few years ago (2016 -> 2020) and GCP (2018-2019) and I remember the difficulty I faced when attempting to reach a human. Even if I were to able to reach someone, I got someone who clearly was a tier 1 support, who wanted…

To be clear I'm in the don't trust Google camp.

But I am not a Google customer. I am a Google user. I use all sorts of Google properties, but i dont give them any money.

What most Gmail users call "customer support" is really "user support". Your free account has the durability of what you pay - zero.

So should we judge customer support based on user support? Technically no. Do we? Of course we do. We absolutely judge the unseen experience on other things we can see.

Quite literally, we judge a book by its cover.

Google has a reputation for bad support. It has earned this reputation from endless free users. It may not be fair, it may not be true, but it exists.

Equally it has a reputation for shutting down services. Popular services. Gmail is likely safe. GCP is likely safe. Am I gonna bet my business existence on "likely"?

Not likely.

Re: I think GCP is better than AWS (2020)

#277

Post by akira2501 »

foobarian wrote: post #232
next_xibalba wrote: post #194

Earlier quoted context omitted.

> The group [Pichai, Page, and Porat] even talked about—and eventually dismissed—the idea of leaving the market entirely, this person said. [1] That they’ve actually discussed shutting it down in the recent past lends plausibility to this risk. In 2019, they set 2023 as a deadline for moving into the #2 spot ahead of Azure. Obviously, they’ve not gotten anywhere near that. [1] https://www.theinformation.com/articles/…

Wow. Imagine top level execs at Ford discussing, but eventually dismissing--the idea of leaving the automobile market entirely. Their hearts really aren't in it as much as with AWS.

Google is an advertising company. They're planning an exit in a non-core market. It shouldn't be a surprise. The weight of "traditional google" has held zero value to me in measuring their cloud product.

Google's cloud products are built around the needs of advertisers. A large and slow backend that's meant to churn ungodly amounts of data.

Amazon is a retail company their cloud products are build around this need. They have a highly segmented and mostly nimble backend that's mean to build unwieldy amounts of API.

Re: I think GCP is better than AWS (2020)

#278

Post by clutchdude »

div_zero wrote: post #36
craigmcnamara wrote: post #6

I find the GCP documentation to be lacking compared to AWS. Specifically while using GCP with Terraform I hit more frustrating nonsense error messages with little to no google hits or undocumented behavior around IAM config compared to my experience doing similar things on AWS.

I'm a technical writing manager on GCP, and we are actively working on improving Terraform documentation. Thank you for bringing up two significant problems with using TF on GCP. If you are willing (and of course have the time), we would love to hear other issues you - or anyone else who reads this thread - are having with TF documentation on GCP as we try and make things better.

This is probably more a request at the compute team/gke teams but...spot management.

For example, managing a spot fleet in AWS via terraform is filled with features and levers I can pull. It's easy to manage basic scale-out/scale-in procedures and allows me to set thresholds.

In GCP, as I understand it, I can only manage a node pool at a basic level.

EDIT: Also, yell at the people who manage the python library.

https://cloud.google.com/python/docs/reference/container/lat...

Generated samples suck - take a look at the boto3 docs and see if they tell you to go look at code generated samples.

Re: I think GCP is better than AWS (2020)

#279

Post by hncommenter13 »

sovnade wrote: post #188

To everyone saying GCP will be gone - it generates over $8 billion in revenue per year and $400 million in profit. There's no chance they kill it any time soon. Yes it posted heavy losses in the past, but any major infrastructure does, and obviously there is a major amount of potential - and they have around 10% of the total cloud market share, not insignificant by any means.

One correction. It's not $8B per year, it's $8.4B in the last quarter--which is over $32B annualized, especially considering the last quarter's revenue grew over 20% from the same time last year. GCP's profit margins are low (for now) but positive. [1]

Moreover, per its filings, Google had almost $65B of contracted backlog representing customer commitments for future purchases (over multiple years), primarily related to Google Cloud. That is not to say those can't ever be unwound or delayed, but it's a pretty meaningful amount, even to a company the size of Google. [2]

[1] https://www.sec.gov/Archives/edgar/data/1652044/000165204423...

[2] https://www.sec.gov/Archives/edgar/data/1652044/000165204423...

edit: formatting

Re: I think GCP is better than AWS (2020)

#280

Post by nathants »

catlover76 wrote: post #219
nathants wrote: post #139

i’ve built bespoke automation at the api level for azure, gcp, and aws. using azure makes you want to watch the world burn. using gcp makes you want to give alcoholism a fair try. using aws makes you wonder why we can’t have nice things. ime those who prefer gcp tend to use a web interface. those who don’t have a preference tend to use iac. you want to work directly with the api, for innumerable reasons. aws is the c…

I defer to your expertise, esp when it comes to actually using these different services at scale, but as someone who used to deploy only to Heroku and wanted to learn real cloud deployment, AWS was pretty unapproachable. I almost laughed out loud at how unnecessarily arcane and difficult they made things when all I wanted to do was set up a simple Python app on EC2. Google Cloud felt easier and intuitive to get start…

too much maybe? yelling at clouds just feels so good sometimes. apologies. i do actually wish we could have nice things.

this is how aws became easy and fun for me[1]. a similar interface could exist for any provider.

heroku is right about picking the good parts and exposing only that as an interface.

1. https://github.com/nathants/libaws

Post reply on HN