6 ms·

The big problem for me is trust. I don’t care what the feature set or performance is; I don’t trust Google enough to bet a business on it. And I’m not even worr

The big problem for me is trust. I don’t care what the feature set or performance is; I don’t trust Google enough to bet a business on it. And I’m not even worried about Google being malicious; I’m worried about them being mercurial and changing/removing things I need without warning.

bnt 6y ago
Or some ML crob job algo going haywire and closing your account - with no human to contact to get it resolved.
- SteveNuts 6y ago
  Google's services are so perfect they don't even need to invest in customer service! If you have a problem with anything, it's clearly your fault and you're doing it wrong. /s
  - ashtonkem 6y ago
    Sarcasm aside, at $COMPANY we have AWS reps integrated into our Slack, and they work hand in hand with our engineers quite often. Even for things as low down as “why is this query so slow on AuroraDB?” They even hop into war rooms for big events in case we need immediate assistance during high visibility outages. It’s hard to overstate how important this level of close support is for an enterprise that is literally betting the farm on a cloud provider. The idea of counting on Google’s historical level of support is an absolute non-starter for us.
    - theshrike79 6y ago
      Agree, we have an integrated AWS contact too and they're really invaluable for getting insight on what the actual limits are on their services.
    - 0xEFF 6y ago
      The role you describe is a technical account manager and Google has hired many. I’m a Google cloud partner and every project I’ve worked on has had multiple TAMS doing exactly what you describe, working hand in hand with customers, connecting the customer to product engineers, to support, navigating Black Friday and Christmas, etc...
      - SteveNuts 6y ago
        To me, Google has always felt like the Stack Overflow of enterprise companies. Our questions to Google have been met with "You're trying to do it this way, but what you really want to do is this". Or worse, they'll send us a link to a document which we've already read ourselves and then just go silent. They've never tried to deeply understand our use cases or business at all. Hopefully they turn this around but IMO, Google is absolutely horrible at B2B, they're just not currently set up to do it correctly.
        0xEFF 6y ago
        I had this experience too around 2015-2017, but I've noticed a big difference between 2018 and now. They've hired a large number of people whose job is primarily focused on understanding the customer's business and use case. Most of these people have empathy in my experience. In the large enterprise migrations I've worked on the past 2 years, I haven't heard a Google employee or partner say, "you're doing it wrong." Sometimes there is a response along the lines of, "we don't recommend doing it that way for the following specific reasons. For your consideration, here are alternatives we recommend for your use case. We've filed your use case as a feature request with product." Additionally, I've seen issues which block a migration to GCP get escalated and fixed quickly. Issues which in ~2015 might have garnered a, "you're doing it wrong" response. I've noticed a real shift in attitude and execution. There is genuine empathy and an attempt to understand how enterprise customers run their workloads in the cloud, multi-cloud and hybrid-cloud.
    - spinningslate 6y ago
      this exactly. Microsoft has known it for years: to be successful with enterprise accounts, people matter. Account managers, sales/support engineers, direct access to product teams (sometimes). I've never worked with Amazon/AWS but from what I've heard anecdotally, they are of the same mindset. (It would be consistent with Bezos' customer service mantra so perhaps unsurprising). I know of only 2 colleagues who've looked at GCP. Both are medium-large financial institutions. Both said they had very little human-to-human contact with Google representatives. In one case they chose AWS instead; the other is still evaluating GCP. It's interesting that other posts in the thread are complimentary about the quality of the GCP products. I can believe that; Google has built its fair share of impressive technology. But I see no evidence of an enterprise supplier I could trust: one that would be there to help out when the world went belly up. Given that, it doesn't matter how good their products are. I'd actually like to see GCP being successful. As well as bringing some competition and diversity to the market, it would give Google a more honest and above-board revenue stream than ads. I don't see that happening though, at least not without a change in leadership.
    - temp667 6y ago
      A while back before AWS was so popular their support at lowest tier developer accounts was totally ridiculous. I'd messed something up, out of their scope to fix (they could have just said nothing wrong with X). Instead some AWS engineer with my permission jumped in and got application side to fix things up and walk me through the configs etc. What!! So I paid $50 for one month for support, and got a full on support expert who'd bill (at least in my world) $250+/hr. Honestly, the support cost was maybe even LOWER, and I'd just signed up for support to submit my ticket. Google does not care. You GSuite calendar is not accessible on google home devices etc etc despite TONS of requests from paying customers. That same calendar integrates easily with ALEXA from amazon. Huh? Someone is paying attention, and it's not google. I do like project based permission / approach in GCP. I like a lot of other things there too. But I've had stuff running for years on AWS without issue - so the trust with AWS keeping things going mostly is there.
bitcharmer 6y ago
I work for a huge international bank. We've had some of our non-production infra on GCP and it's been a very bad experience for us; mainly reliability-wise. We're moving away from it entirely.
- dijit 6y ago
  I put a production AAA game on it (with horrible constraints for a cloud provider, like highly stateful connections and hard requirement to honour fsync in the underlying hardware)- and we had a truly great experience. Can you go into more detail? Not saying you’re wrong of course, but your experience is massively contra to my own, and like I said, we have production on there with little to no issues.
phreeza 6y ago
I know it's easy for me to say, but shouldn't nobody be building their business to be critically dependent on some other business, regardless of whether cloud provider or other class of service? Especially when the balance of power clearly favours one of the parties, this seams like a bad idea if it can be avoided. I don't know how hard it is to develop architectures which run on several cloud services. If it is possible, shouldn't this be standard practice?
- KronisLV 6y ago
  > I don't know how hard it is to develop architectures which run on several cloud services. If it is possible, shouldn't this be standard practice? I think the answer to this question depends on a multitude of factors. For example, nowadays you can use container technologies to create images of your software, which will run on any bit of infrastructure that supports them, be it AWS, GCP, Azure, or even your on-prem servers with something like Rancher or even Docker Swarm running on it. However, creating software like that needs some special thought put into it, this site covering some of those aspects: https://12factor.net/ https://12factor.net/ And while this will make your code more reproducible and migrating between different clouds more easy (or even running on multiple clouds simultaneously), it'll do so at the expense of making development slower and a bit harder for you. For some, this will be worth it, while for others it'll be less so. There are definitely valid criticisms of container technologies and some of the aspects they don't handle all that well yet (Kubernetes being overcomplicated for some projects, whereas Docker Swarm isn't "trendy" or in active development, then there's managing storage and needing to work with either network file systems, or distributed file systems, or even using bind mounts even though they're considered bad practices, then there's routing and the complexity of service meshes etc.). Of course, there are also people who really don't want to think about infrastructure that much and just want their apps to run in a semi-managed manner, like Heroku does, or perhaps just want to use one of the cloud vendors' managed database or messaging system offerings. Not everyone has a large amount of resources to invest in engineering and running infrastructure.
- shaicoleman 6y ago
  A multi-cloud strategy is generally not a good idea. Some links that will hopefully explain it better than I can: * https://www.lastweekinaws.com/blog/multi-cloud-is-the-worst-practice/ https://www.lastweekinaws.com/blog/multi-cloud-is-the-worst-... * https://www.infoworld.com/article/3428682/your-multicloud-strategy-is-all-wrong.html https://www.infoworld.com/article/3428682/your-multicloud-st...
Spooky23 6y ago
It’s an old and lazy/annoying complaint. There’s a difference between consumer stuff and enterprise stuff with contracts. Grown up services like GSuite, AppEngine, etc have been alive and well for many years and aren’t going anywhere. It makes sense to do risk assessment and avoidance where there is value. General emotional stuff isn’t productive. Building your entire business on AWS Lambda, for example, is a risk you need to understand. In the past, I worked on a team that chose to put a critical business process on a IBM POWER/Aix platform... in our case we went through the options, identified risks/opportunities for standardizing a long lived process on a sole source platform, and made a decision. It was a decision that served us for about a decade before we moved on, so it was very successful.
- jshen 6y ago
  Google has absolutely abandoned things in their cloud. App Engine has been a moving target for years for just one example.
  - Crash0v3rid3 6y ago
    Can you give a more clear example of what GCP has dropped support for?
    - Strom 6y ago
      Google provides a list of App Engine features they've removed. [1] Beyond that there is also a somewhat undocumented phase of working-but-forgotten. Classic App Engine features like the datastore, memcache, Users API, Python 2, Go 1.11 etc go under this category. These are things that still work, but get no updates. Instead you get constant e-mails and other notifications about how you should redesign your app to work with the 2nd generation App Engine system. Which means Firestore (in datastore mode) instead of datastore. Memorystore instead of memcache. Your own solution instead of the Users API etc. -- [1] https://cloud.google.com/appengine/docs/deprecations https://cloud.google.com/appengine/docs/deprecations
      - 6y ago
shaicoleman 6y ago
Not to mention some of the ridiculous price increases, which could be enough to kill your business if you depended on them, e.g. * Google Maps API increasing prices by 1400%+ * GKE price increasing from free to ~$73/month per cluster
throwaway28392 6y ago
Long-time GCP user here, from a big conservative enterprise. GCP is a wonderful technical platform, but if I ran an oil or coal or tobacco or weapons company, or any other business susceptible to future boycotts and political activism, I would only adopt GCP in a cloud-neutral way. Google employees have disproportionate power over what gets done, versus for example AWS where customers seem to have a lot of power.

bnt 6y ago

Or some ML crob job algo going haywire and closing your account - with no human to contact to get it resolved.

SteveNuts 6y ago
Google's services are so perfect they don't even need to invest in customer service! If you have a problem with anything, it's clearly your fault and you're doing it wrong. /s
- ashtonkem 6y ago
  Sarcasm aside, at $COMPANY we have AWS reps integrated into our Slack, and they work hand in hand with our engineers quite often. Even for things as low down as “why is this query so slow on AuroraDB?” They even hop into war rooms for big events in case we need immediate assistance during high visibility outages. It’s hard to overstate how important this level of close support is for an enterprise that is literally betting the farm on a cloud provider. The idea of counting on Google’s historical level of support is an absolute non-starter for us.
  - theshrike79 6y ago
    Agree, we have an integrated AWS contact too and they're really invaluable for getting insight on what the actual limits are on their services.
  - 0xEFF 6y ago
    The role you describe is a technical account manager and Google has hired many. I’m a Google cloud partner and every project I’ve worked on has had multiple TAMS doing exactly what you describe, working hand in hand with customers, connecting the customer to product engineers, to support, navigating Black Friday and Christmas, etc...
    - SteveNuts 6y ago
      To me, Google has always felt like the Stack Overflow of enterprise companies. Our questions to Google have been met with "You're trying to do it this way, but what you really want to do is this". Or worse, they'll send us a link to a document which we've already read ourselves and then just go silent. They've never tried to deeply understand our use cases or business at all. Hopefully they turn this around but IMO, Google is absolutely horrible at B2B, they're just not currently set up to do it correctly.
      - 0xEFF 6y ago
        I had this experience too around 2015-2017, but I've noticed a big difference between 2018 and now. They've hired a large number of people whose job is primarily focused on understanding the customer's business and use case. Most of these people have empathy in my experience. In the large enterprise migrations I've worked on the past 2 years, I haven't heard a Google employee or partner say, "you're doing it wrong." Sometimes there is a response along the lines of, "we don't recommend doing it that way for the following specific reasons. For your consideration, here are alternatives we recommend for your use case. We've filed your use case as a feature request with product." Additionally, I've seen issues which block a migration to GCP get escalated and fixed quickly. Issues which in ~2015 might have garnered a, "you're doing it wrong" response. I've noticed a real shift in attitude and execution. There is genuine empathy and an attempt to understand how enterprise customers run their workloads in the cloud, multi-cloud and hybrid-cloud.
  - spinningslate 6y ago
    this exactly. Microsoft has known it for years: to be successful with enterprise accounts, people matter. Account managers, sales/support engineers, direct access to product teams (sometimes). I've never worked with Amazon/AWS but from what I've heard anecdotally, they are of the same mindset. (It would be consistent with Bezos' customer service mantra so perhaps unsurprising). I know of only 2 colleagues who've looked at GCP. Both are medium-large financial institutions. Both said they had very little human-to-human contact with Google representatives. In one case they chose AWS instead; the other is still evaluating GCP. It's interesting that other posts in the thread are complimentary about the quality of the GCP products. I can believe that; Google has built its fair share of impressive technology. But I see no evidence of an enterprise supplier I could trust: one that would be there to help out when the world went belly up. Given that, it doesn't matter how good their products are. I'd actually like to see GCP being successful. As well as bringing some competition and diversity to the market, it would give Google a more honest and above-board revenue stream than ads. I don't see that happening though, at least not without a change in leadership.
  - temp667 6y ago
    A while back before AWS was so popular their support at lowest tier developer accounts was totally ridiculous. I'd messed something up, out of their scope to fix (they could have just said nothing wrong with X). Instead some AWS engineer with my permission jumped in and got application side to fix things up and walk me through the configs etc. What!! So I paid $50 for one month for support, and got a full on support expert who'd bill (at least in my world) $250+/hr. Honestly, the support cost was maybe even LOWER, and I'd just signed up for support to submit my ticket. Google does not care. You GSuite calendar is not accessible on google home devices etc etc despite TONS of requests from paying customers. That same calendar integrates easily with ALEXA from amazon. Huh? Someone is paying attention, and it's not google. I do like project based permission / approach in GCP. I like a lot of other things there too. But I've had stuff running for years on AWS without issue - so the trust with AWS keeping things going mostly is there.

bitcharmer 6y ago

I work for a huge international bank. We've had some of our non-production infra on GCP and it's been a very bad experience for us; mainly reliability-wise. We're moving away from it entirely.

dijit 6y ago
I put a production AAA game on it (with horrible constraints for a cloud provider, like highly stateful connections and hard requirement to honour fsync in the underlying hardware)- and we had a truly great experience. Can you go into more detail? Not saying you’re wrong of course, but your experience is massively contra to my own, and like I said, we have production on there with little to no issues.

phreeza 6y ago

I know it's easy for me to say, but shouldn't nobody be building their business to be critically dependent on some other business, regardless of whether cloud provider or other class of service? Especially when the balance of power clearly favours one of the parties, this seams like a bad idea if it can be avoided. I don't know how hard it is to develop architectures which run on several cloud services. If it is possible, shouldn't this be standard practice?

KronisLV 6y ago
> I don't know how hard it is to develop architectures which run on several cloud services. If it is possible, shouldn't this be standard practice? I think the answer to this question depends on a multitude of factors. For example, nowadays you can use container technologies to create images of your software, which will run on any bit of infrastructure that supports them, be it AWS, GCP, Azure, or even your on-prem servers with something like Rancher or even Docker Swarm running on it. However, creating software like that needs some special thought put into it, this site covering some of those aspects: https://12factor.net/ https://12factor.net/ And while this will make your code more reproducible and migrating between different clouds more easy (or even running on multiple clouds simultaneously), it'll do so at the expense of making development slower and a bit harder for you. For some, this will be worth it, while for others it'll be less so. There are definitely valid criticisms of container technologies and some of the aspects they don't handle all that well yet (Kubernetes being overcomplicated for some projects, whereas Docker Swarm isn't "trendy" or in active development, then there's managing storage and needing to work with either network file systems, or distributed file systems, or even using bind mounts even though they're considered bad practices, then there's routing and the complexity of service meshes etc.). Of course, there are also people who really don't want to think about infrastructure that much and just want their apps to run in a semi-managed manner, like Heroku does, or perhaps just want to use one of the cloud vendors' managed database or messaging system offerings. Not everyone has a large amount of resources to invest in engineering and running infrastructure.
shaicoleman 6y ago
A multi-cloud strategy is generally not a good idea. Some links that will hopefully explain it better than I can: * https://www.lastweekinaws.com/blog/multi-cloud-is-the-worst-practice/ https://www.lastweekinaws.com/blog/multi-cloud-is-the-worst-... * https://www.infoworld.com/article/3428682/your-multicloud-strategy-is-all-wrong.html https://www.infoworld.com/article/3428682/your-multicloud-st...

Spooky23 6y ago

It’s an old and lazy/annoying complaint. There’s a difference between consumer stuff and enterprise stuff with contracts. Grown up services like GSuite, AppEngine, etc have been alive and well for many years and aren’t going anywhere. It makes sense to do risk assessment and avoidance where there is value. General emotional stuff isn’t productive. Building your entire business on AWS Lambda, for example, is a risk you need to understand. In the past, I worked on a team that chose to put a critical business process on a IBM POWER/Aix platform... in our case we went through the options, identified risks/opportunities for standardizing a long lived process on a sole source platform, and made a decision. It was a decision that served us for about a decade before we moved on, so it was very successful.

jshen 6y ago
Google has absolutely abandoned things in their cloud. App Engine has been a moving target for years for just one example.
- Crash0v3rid3 6y ago
  Can you give a more clear example of what GCP has dropped support for?
  - Strom 6y ago
    Google provides a list of App Engine features they've removed. [1] Beyond that there is also a somewhat undocumented phase of working-but-forgotten. Classic App Engine features like the datastore, memcache, Users API, Python 2, Go 1.11 etc go under this category. These are things that still work, but get no updates. Instead you get constant e-mails and other notifications about how you should redesign your app to work with the 2nd generation App Engine system. Which means Firestore (in datastore mode) instead of datastore. Memorystore instead of memcache. Your own solution instead of the Users API etc. -- [1] https://cloud.google.com/appengine/docs/deprecations https://cloud.google.com/appengine/docs/deprecations
    - 6y ago

shaicoleman 6y ago

Not to mention some of the ridiculous price increases, which could be enough to kill your business if you depended on them, e.g. * Google Maps API increasing prices by 1400%+ * GKE price increasing from free to ~$73/month per cluster

throwaway28392 6y ago

Long-time GCP user here, from a big conservative enterprise. GCP is a wonderful technical platform, but if I ran an oil or coal or tobacco or weapons company, or any other business susceptible to future boycotts and political activism, I would only adopt GCP in a cloud-neutral way. Google employees have disproportionate power over what gets done, versus for example AWS where customers seem to have a lot of power.