Transitioning Google Cloud leadership after three great years
cloud.google.com
cloud.google.com
Based on my own experience as the CTO: After being an Azure shop for a year we've migrated ~50-100 VMs to GCP and I love the GCP products.
GCP is:
- Simpler to use
- More tailored to people with Linux environment
- Leader in K8S
- Has good support
- So much cheaper (in our case we saved ~60%)
- Has great UI and understandable primitives.
My only pet peeve is the fact that exporing your spend is practically impossible unless you're a BQ guy that can work directly with report exports.
PS: We're building our future infra on K8S to allow us to migrate more easily to a different could if something goes awry with GCP, I really hope there won't be a need to migrate back to Azure and its arcane and high pricing, strange UI, worse tooling...
That blog post is good, but it is over built in my opinion, all you really need is a single daily sum of all spend across your account and chart that. Maybe a stacked column with product color or something. This will allow you to quickly see if one product consumption is shooting up unexpectedly at a glance vs checking something at the end of the month and working with your account reps to figure something out.
6k z z
5k z z
spend 4k z z z
3k z z z
2k x x x x x x x x x
1k x x x x x x x x x
date 1 2 3 4 5 6 7 8 d+
x = expected
z = unexpected
[1] https://cloud.google.com/billing/docs/how-to/export-data-big...
[2] https://medium.com/google-cloud/visualize-gcp-billing-using-...
1: https://cloud.google.com/resource-manager/docs/creating-mana...
And one of the advantages of having a zillion different products approach is that you can trust AWS to grow with you and improve their products over time. Same can't be said of GCP.
Compute is compute. Both providers offer similar instance classes for the most part. GCP offers sustained-use discounts, in addition to committed use discounts that match up to reserve instances, plus billing by the minute.
Spot instances can be hugely variable, but cheaper than prremptible. Preemptible offers predictable costs with a guarantee that the instance will go away every 24 hours (and come back).
Network IO is way better.
Inter-region networking comes for free.
GCP ssd persistent disk lands between AWS GP2 and IO1 and gives way more predictable performance.
Local disks can be attached to any instance and end up giving similar performance and being way cheaper than I2 and I3 instances.
Wait until the next post comes up about “serverless” and someone mentions that there are servers involved....
There is no guarantee that the instance will come back. And it can be shut down in under 24 hours. Not that that's a problem compared to AWS or Azure.
>Because preemptible VMs have no availability guarantees, you should design your system under the assumption that any or all of your Compute Engine instances might be preempted and become unavailable. There are no guarantees as to when new instances become available.
https://cloud.google.com/kubernetes-engine/docs/how-to/preem...
There's also reOptimize for in-depth reporting with a solid free plan: https://www.reoptimize.io/
Azure's VMs are definitely overpriced compared to both GCP and AWS, but most other services are reasonable.
Not sure what you mean about Azure's "strange UI"? Personally, I find the Azure Portal to be far more consistent than AWS' or GCP'. I also think it looks and feels great, and I'm obviously not the only one - I literally have customers asking me to replicate the Azure UI for their web apps!
Lastly, I find it somewhat incredible that you would complain about Azure's tooling - I've found it to be excellent: a snappy web UI, cross-platform Powershell, cross-platform Azure CLI, REST API, ARM templates... seriously, I'd be interested to hear what your beef is with the tooling?
You need to drill down to the network card primitive, figure out which subsection you need to configure and applying the same rule is again an exercise in clicking.
In GCP: Apply a tag. Create firewall rule that targets the tag.
But then again, this is higly subjective and I do understand that somebody would prefer Azure way.
Re: tooling, when I was using CLI was in both Node and Python, docs were lacking and the strange Resource Manager vs. Legacy had so many quirks we had to run windows to get proper powershell for configs of VMs.
Leader in k8s? Not for production. I defy you to do a gcp gke deployment without using beta or alpha features (that specifically say don’t use this for production). Stackdriver error reporting springs to mind. And working with SSL certs/let’s encrypt is a close second.
If you want a couple more peeves for your list how about documentation? I regularly find documentation heavy on theory light on specifics.
GCP is my favorite thing that is almost good enough to use. I’m hoping the change of leadership actually improves things. GCP could stand to re-focus on a “customer first” philosophy.
I don't think we use any Alpha features, I know there are a couple Beta APIs from Kube we use, but we would be using those no matter where we ran our Kube deployment. We use fluentd for our log collection and Prometheus for metrics and alerting combined with Elastalert. So don't have any experience with Stackdriver.
Why is working with SSL certs an issue? We use the NGINX Ingress with cert-manager and have yet to have an issue. We only have a couple domains pointed at the cluster though, not sure if thats where your issue comes from. The two just worked though with very little tweaking.
Average availability for the year across those environments is 99.99%.
I'd say that's decent enough to use.
Their support is hilariously bad at times, though.
I find it a bit ironic that the mobile and web side of the house are pushing very heavily for IPv6 adoption and yet they expect their cloud customers not to need it.
That was my experience when we needed something.
Dont get me started on docs and operations.
Eg: Using Azure we found out the hard way that the Firewall was cutting connections after 5m of inactivity. Changing that setting? Their docs are wrong, stack overflow kindof helps, but to fix it we had to use windows based tooling in a vm since their linux based stuff didnt work.
I know this will happen sooner or later with anybody, but on average the experience with GCP was superb.
I do agree re: certs. That is the only thing I really hate, doing LE when they could run this for you. But they will launch an SSL solutio since they have one now for AppEngine.
Could you elaborate on this? Which firewall is this? I remember something similar with their load balancer but I thought this was fixed years ago. Might you have some link to the issue? Thanks.
StackDriver error reporting is a lite product intended mostly for folks with small applications on AppEngine.
There is a recent post by Corey Quinn entitled CloudWatch Is of the Devil, but I Must Use It discussing the extremely poor usability of Amazon CloudWatch [0]. Here’s another story about the general weirdness of the product [1]
I would strongly recommend using a dedicated exception tracker like Sentry.io, they have an excellent full featured product at a extremely reasonable price.
GCP has clearly documented Platform Launch stages for Alpha, Beta, and General Availability [2].
The Kubernetes project itself follows the same definition for these identifiers as GCP.
Specifically Beta is considered to be acceptable for most production use-cases, but it is not guaranteed that the API will remain backwards compatible.
So if you’re running the majority of cloud services, beta is fine. If you’re running a system that will cause permanent harm to humans, don’t use beta.
One last note, GCP has just released a fully managed certificate product for their global load balancer. There is fully integrated Kubernetes support which has not yet been announced formally [3].
Certificates for Kubernetes on Amazon are less integrated than the above. There’s several options, it’s not clear without extensive research and testing what the tradeoffs are, and in the end it turns out that there is not an officially supported Amazon option.
[0] https://www.linuxjournal.com/content/cloudwatch-devil-i-must... [1] https://www.circonus.com/2016/10/no-fixed-glitch/ [2]https://cloud.google.com/terms/launch-stages [3] https://github.com/GoogleCloudPlatform/gke-managed-certs
I used to do $100,000 a month in revenue on Amazon with one HP skew. The product grew to #1 in its category within a couple months, another month after which Amazon started selling it to, "shipped from sold by Amazon".
Amazon is happy to help be your growth partner until you prove worth devouring.
AWS is their most profitable business in terms of margin. It makes sense to avoid funding your competition if possible, which is a problem for Amazon given how many business divisions they have and continue to add.
Google is very different internally from the enterprise loving Microsoft, and they're more technically fluent than the customer focused Amazon. Googles edge targeting and process scheduling are also very different.
These differences make GCP a very different product than it's competition. Azure is essentially leveraging the ecosystem fracture that Microsoft already holds, and the fact that Microsoft knows how to speak the slow drawl of enterprise. GCP is situated to solve things the Google way, which isn't the way most people run their infra.
Not sure that I've ever seen this articulated so brilliantly.
Citation needed -- from my knowledge, the culture shifted dramatically -- she did succeed, and now that's done.
Granted, I'm referring to Google culture within Cloud/GSuite... obviously she didn't affect culture at YouTube or Search because those weren't her areas.
And once IBM launches their cloud I would actually put them ahead of GCP given how useful their Compose suite is.
Compose is just a managed database service which IBM also acquired. I don't see how that alone would put them ahead of any cloud, especially when they're already lacking in infrastructure and software.
And Compose manages databases on AWS and GCP too, similar to Aiven and others, so I don't see how that alone makes IBM special.
Not that their infrastructure is insignificant ... but it's insignificant
Interesting. Is it just another rename (Softlayer -> BlueMix -> ?). Or something else?
https://9to5google.com/2018/11/16/google-cloud-diane-greene-...
But I will have to admit that AWS seems a lot more committed to their cloud products than Google. New GCP products and features take forever to be introduced or phased out of beta.
For example, cloud functions entered beta some 2 years ago and went out of beta a year later. Even today it's only possible to use Node and Python. Only recently the Node runtime was upgraded from old Node 6 to Node 8.
https://cloud.google.com/functions/docs/writing/
During that same time, a small company like Zeit has created a number of complete cloud products for developers. Not only that, but Zeit Now v2 is better in many aspects than Google Cloud Functions.
Why isn't google invested in GraphQL? They could have come up with something like Prisma with all their talent and resources.
In any case, people should be aware of those limitations when choosing a provider.
One of my projects needed to be used by some users in Cuba, and guess what? GCP doesn't work from Cuba.
I had to adapt one of the features of the project to upload files to AWS S3 which works perfectly there.
Even if GCP accepted packets from Cuba, your problems probably would have persisted on the Cuban end too.
It's ironic that it's fine for Google to provide it's products in those countries, but not for me, if I host my app on their cloud.
US sanctions do not require an infrastructure provider to do this, much like they do not require a broadband provider to filter packages destined to those countries.
Instead the gold podium is occupied by a book re-seller. But the entire Alphabet product line, from the amazing distributed Big Tables to the stitching of 3rd party satellite polygons, is just a front-end cloud use case.
So why didn't it happen? Departing CEO Greene writes in the OP link that she was only supposed to be running GCP for 2 years. Why on earth was an interim CEO there in the first place
This was pivotal moment: backwards Cloud PA nobody cared about (fun story: Cloud PA was once forgotten at company-wide key result meeting, it was fun to see how couple hundreds of people watching video-cast in the auditorium looked at each other and like "meeting is over, they didn't mention or PA once, what is going on!") suddenly became the priority, was merged with TI (Tech Infrastructure) and everything became "Cloud! Cloud! Cloud!".
IMO, Google was just too focused on solving problems for human beings vs. outside institutions; that's why they neglected to invest in making Borg into a public platform.
1. Google saw their infrastructure as being a key competitive advantage, not something to be commoditized.
2. More importantly: Google's core business has ridiculously high profit margins, which made them institutionally incapable of seriously pursuing low-margin business opportunities.
[1] https://www.businessinsider.com/diane-green-donates-150m-to-...
I did interview with Microsoft Global Network Services several years ago (before Satya Nadella) and got a sense on the infrastructure and money they were pumping into Bing.
They were replicating the style of data center deployments that Google was known for in terms of scale and efficiency. I wasn't really privy to any specific details beyond the Clos architecture they'd adopted for their underlay and the fact that they rolled their own switches and SDN.
I don't recall specifically when this was, but I do recall being impressed that they had something deployed and working. The only folks anyone were certain had done this at the time was Google. Facebook would later announce something similar with their Wedge switch and now there's a fully baked OCP standard.
They point of all this I suppose is that they were heavily investing in scale for Bing and Office 365. I suspect the product of all that work is what became Azure.
https://www.quora.com/Have-you-worked-for-Thomas-Kurian-at-O...
Google has no cloud counterpart to that level of service. If TK brings an equivalent orientation that makes customers believe that GCP will do whatever is necessary to resolve a problem, he'll have contributed significantly to its success, IMHO.
I knew there must have been a good reason they overcharged everyone for everything...
In all seriousness, I imagine this is the type of thing you can do when you have (or traditionally had) close to a captive market for specific things (government contracts) that allows for massive profits. It's easy to throw a few tens of thousands of dollars at black swan support events so you can point to those as examples of what you're paying for when people point out that you've been repeatedly found to violate contracts and fraudulently misrepresent yourself.[1]
1: https://www.mercurynews.com/2010/07/30/whistleblower-details...
TK is very much old school Oracle though. Doesn't feel like a good cultural fit with Google.
Disclosure: I work at OCI.
A lot of lock-ins, contracts and security... for Oracle.
During her tenure it moved from something I wouldn't even consider using, to first choice for my new startup.
- Much less people with proper platform knowledge (not even talking about certified).
- Not much trust on deprecation policies (AWS keeps running very old and deprecated services virtually forever, there’s no guarantee GCloud would do the same given Google history).
- Not from personal experience but customer support is not top notch as with other providers from what I have read and heard.
In my opinion those three reasons alone are big flags for many corporations, which might prevent them from getting the big contracts.
Edit: formatting
That said, I'm not sure I'd use it for anything that might scale heavily with datastore writes. They get expensive very quickly if you can't cache and group the writes.
On top of this, there is the not-so-distant issue with the sudden price hike of Google Maps API, which made many companies struggle (in this case especially small ones).
The engineering resources to migrate was also a problem with the whole stack having needed to be updated.
We've been changing a lot of infrastructure, adding features, fixing bugs, all without users being affected.
I'm personally very excited about the Second Generation Runtimes, which have been worked on in parallel, and run the runtimes unmodified.
Most enterprises don't have a clue about computing though, so I can buy that they would prefer bling like certificates to getting a contact who is a real engineer.
As far as i heard, the VMWare people Diane brought into the Google Cloud org were screaming from the pain of being an adapter between the hard place of the rest of the Google (i.e. the infrastructure, etc.) and the sledgehammer of enterprise customers who, still mentally being unenlightened dwellers of the Dark Ages, fail to understand the 30 seconds shutdown notice suddenly coming from deep inside the Google infra guts as an "advanced maintenance notice".
Lets see what happens to the Oracle guys who are already coming upon the Google Cloud like Viking drakkars upon Northumbria shores.
James Mattis will use his knife hands to convince enterprise customers of the seriousness of Google’s support team.
Do they really think they were going to become market leaders in three years? Come on. The market for cloud infrastructure is in its infancy. The opportunities are there. But you don't build a large business within a potentially multi-trillion dollar industry by being impatient.
And hiring a former Oracle executive just shows how impatient they are... It a complete change of the playbook and it likely means that they are scrapping a huge chunk of Greene's original go to market strategy for Google Cloud.
However there is that Maps pricing change that should be noted, and perhaps counts the same as forced deprecation for many clients.
They haven't done any pricing change like that, nor anything else effectively equivalent to forced short-notice deprecation, for generally available Google Cloud Platform services. (The famous Google App Engine pricing change happened upon GAE's departure from preview.)
Hopefully the new CEO can disentangle the horribly messy branding conflation of orgs which work rather differently.
Google is now effectively a conglomerate, even ignoring Alphabet, maybe even multiple conglomerates within Google. The impacts of their org chart on their users shouldn't be this opaque to people who haven't worked at Google. (Disclosure: I have worked there, though not since 2015.)
However, I was referring more to Google's general tendency to shut down projects across the board. The list of consumer products shut down by Google is pretty long (Google Reader, iGoogle, Aardvark, Google Inbox, Orkut, etc..)
I think both Fabric and Apigee API BaaS where indeed Google Cloud offerings. Google Search Appliance if I'm not wrong was part of something called Google Enterprise which eventually folded into GCP.
>> Apigee API BaaS, Fabric, Google Search Appliance...
Curious, what usually happens to the engineers on projects that get shut down? Do they get redeployed to other areas? What if the team is in a small office?
I've walked past the Apigee office in Reston, VA...what would happen to those engineers...i imagine there arent many other options for redeployment short of moving cities...
Thanks for pointing out my hypocrisy and why I don't have the right to have an opinion about anything anymore.
I also checked your profile and you're pretty good at defending everything Google does, but I don't judge you for having those opinions, nor I think you're a hypocrite for thinking in a certain way.
If I criticize Company A for doing something, and goes around and does the same thing then it'll be very hypocritical of me.
The product-market fit you mentioned doesnt just happen for startup, but for big companies as well. A dollar is a dollar everywhere you go. Just like you found it was OK for your company to sunset a certain product, big companies also decide to do that.
Also, please dont put words into my mouth. I didnt say you dont have a right to say things. You do. But when you do, and please be open to replies.
Companies have the right to sunset products if they're not profitable. Of course, they do. I think the problem is that in general people have high expectations from companies like Google, Amazon, and Microsoft. I certainly do, and that's why I think Google should be more committed to a long-term vision in their Cloud business.
Diane Greene only planned to be there for two years, so three years in, this change is overdue. If anything, keeping her longer demonstrates the opposite of what you're saying.
It's the first sentence in the article.
I would be curious to know if her passions are intertwined with her departure, specifically if she felt that Google had structural issues with women in leadership roles. She is essentially being fired as the lead of this organization and being replaced with a fairly prototypical male lead. I am not implying that Thomas isn't qualified, his LinkedIn profile suggests he have very relevant experience, Diane also has the experience in her resume as well. What will help Thomas succeed in this role?
The bottom line is I would like to see more women in leadership positions and I am sad to see us lose one. I would like to understand (although I realize that isn't possible) if her departure was preventable, and if so what would have had to be different.
Nothing in her resume actually suggests she had the experience for what Google needed. VMware became popular in a similar way to Redhat, in that it solved a technical issue nobody else was solving at the time. It didn't have explosive growth until EMC took over and with it EMC management style/structure. She was just along for the ride the second the company was sold to EMC.
>What will help Thomas succeed in this role?
He was in a position of power in what is one of the most successful sales organizations in the history of tech. I don't like Oracle, in fact my post history will show that I loathe them. But their salesforce has absolutely dominated the tech industry for 2 decades+. Google needs someone that knows how to sell to enterprise, they don't just need another nerd at the top. Based on my experiences with Diane she was an extremely gifted technical talent and an extremely poor saleswoman and leader.
This is not true. Diane was the founder of VMware. VMware is a virtualization technology that runs independently of underlying hardware and infrastructure. It could succeed only because it was able to convince its customers and partners that it is neutral to all forms of infrastructure, whether it is compute, networking or storage related. Most of said partners - companies as IBM, HP, Netapp etc. - are direct competitors of EMC.
This is the reason that EMC allowed VMware to operate independently, be listed on the stock exchange as a independent entity and have its own operating culture. And Diane was the main reason that VMware maintained this kind of independence. Source: I was a VMware employee.
Diane was a founder of VMWare. VMWare was very successful at getting Enterprise customers to convert their massive single use servers into "Liquid IT" where VMWare instances ran the world. NetApp, when I was there, was selling a bunch of NAS into clouds of servers running VMWare. Then she left VMWare and started BeBop, a company that was dedicated to getting Enterprises to move their applications into virtualized compute clusters. When she was acquired/hired by Google is seemed like a slam dunk that she was the person who was going to make Google Cloud the premier service for Enterprise companies to host their infrastructure on.
As I read it, that was very much in line with where Google saw Google Cloud going.
Then we see from the market analysts that Microsoft has (according to the analysts here, I can't really vouch for them) edged Google into third place in the 'Cloud' market by capturing Enterprises moving their applications to the enterprise space.
Then we see Google replacing their head of cloud with a high power sales guy.
And as always, "reading the tea leaves" as an analyst might say, is always an inexact science. So folks outside the situation (like myself) can't know what is really going on.
> Based on my experiences with Diane she was an extremely gifted technical talent and an extremely poor saleswoman and leader.
This seems to be a pretty damning statement, as someone who has direct experience with Diane, what did you see that she could have done differently?
They need to offer a better product with more features, support etc. Because adoption within the enterprise isn't driven by the CEO or CTO. It's driven by individual architects and developers.
Most non software firms consider technology as a cost center. Most don’t care about the technical merits of the platform; all they care is cost and support.
This is why IBM cloud is seeing adoption despite not being as full featured as aws or gcp.
That’s not being fired.
Separations can happen for several reasons and they have different terms, I recognize that most people associate "fired" with incompetence or misconduct and that likely wasn't the case here[1].
Generally, with senior leaders, they have a set of objectives that they are tasked with working toward, and every year is a negotiation over whether or not they are meeting their objectives. When they aren't meeting those objectives (or perhaps enough of those objectives) then there is a discussion about next steps. Which is a euphemism for "this isn't working for us, we need to replace you with someone who can meet these objectives."
Now is that getting fired? or laid off? I think it's semantics. At the end of the day you don't have a job with the company any more. Few senior leaders that I've interacted with have ever called out the company for telling them they were going to lose their job. If you have reached that level you should have the maturity to understand that it is a fact of life that this happens. Scott McNealy used to send congratulations to people who were promoted into the senior ranks with "One step up, one step closer to the door." That reflected the reality that there are few alternative positions within a company for someone who is leading a big chunk of the company.
When the separation is the idea of the employee (which is to say they quit), the notification pattern seems proactive on the employee's side. A press release that the employee is moving on and that the company is working on finding a successor, and then sometime later we get the "I'm actually leaving now, and here is my successor." press release. But there is no hard and fast rule.
Bottom line is that the term 'fired' has the connotation of malfeasance and that doesn't seem to be the case here, I should have used "likely involuntary separation" but that seems a bit wordy, although it avoids the baggage of the word fired.
[1] As we can see from documents about Andy Rubin's departure, all of the press at the time read like Andy was just moving on to bigger and better things, when in fact he was being separated from the company because of credible accusations of sexual harassment. My point being that we on the outside can't know why a senior leader is leaving really, the copy in the press releases is carefully crafted regardless of the actual reasons.
Absolutely. But let's be real, the odds are likely that Google tried to bend over backwards to make it work. They want more women in those roles and generally.
"Those the gods wish to destroy, they first make arrogant"
People expect Google to work, as Google trained people to expect its things to just work. Like Google.com, Gmail, Android.
And GCP is complex, it will break down. And then people just get frustrated. Yes, it works from the CLI, but fuck that. I know AWS is slow and expensive, but at least it does the basics. GCP was unable to launch VMs - for fuck's sake - on the UI just a few days ago with mysterious resource not available errors. (Anecdotal first hand experience, I've spent at least 1-1 hour trying to get an instance in a EU zone, to no avail. Even waited half a day between.)
Container output isn't hijacked, it's all logged to files, and those files are tailed by the logging agents. You can run multiple logging agents, for example we use https://logdna.com/ (highly recommended) and it works fine in addition to stackdriver logs.
1. https://groups.google.com/forum/?hl=EN#!forum/gce-discussion
2. https://stackoverflow.com/questions/42380626/how-do-i-disabl...
The logs are still available in the cluster (not hijacked) and you can configure your own export mechanism.
https://www.businessinsider.com/how-the-kurian-twins-won-the...
Unfortunately, there's not much the Cloud team at Google can do. Google as a whole needs to demonstrate long-term support in a superior fashion before I can trust their products for building my products.
Why is this getting downvoted? Provide some input on it, has Google suddenly stopped killing projects that are in their infancy? Why should I entrust my platform to that behavior? Anything instead of just mass downvoting.
HN you're better than this.
But I understand hat some customers require enterprisey products that stay exactly the same for at least 10 years, and a 1 year notice might not be enough to make adjustments.
There's a reason they are 3rd in cloud tech.
Wondering what people think about this.
bings "Google is the new Oracle"
yeah, nobody says that.
To be fair, there was a single tweet in March that used the phrase, so it's been said once in history.
Other companies are huge contracts, and simply won't sign on without that level of dedication. They demand it, and the CEO's give it.
The better question is, "But what type of people will he gravitate towards hiring to develop some products?"
Its funny that Hacker News hates on Oracle, but their database is far ahead of the competition in a highly competitive market. Say what you want about their other products (that are indeed terrible), but that core database gets them billions in revenue for a reason.
Why? A salesman says whatever he’s paid to say. Tomorrow he could move again and be flogging IBM’s joke of a cloud.
But this is happening with increasing frequency even for mid-level leadership positions, and it is showing more and more in the culture.
Source: Ex-googler.
What does "peer cannibalization" mean?
I know of two groups (one of maybe 10, and another maybe 30 engineers) that sat essentially leaderless for several months each before Google hired a manager from outside the company. Neither one of these, in my opinion, have a good handle on Google culture.
There is no real reward for becoming a manager (no more pay, no more level, no more nothing), but a lot of hassle, so very few engineers want to do it.
At least that is what I see from my limited vantage point.
Source: personal experience circa 2013-2015, when I was there and trying to transition from individual contributor and tech lead to first-time people manager. Of course I wasn't considering executive roles like this article is about.
For some areas, I agree. But Google does need some outside influence on how to sell/service the Fortune 500.
They are ceding business to Amazon not because Amazon is great at it, but only because Amazon is less terrible at it.
IBM, CA, etc can sell pure garbage because their sales folks understand how the game is played, who to please, etc. Amazon won't even do simple stuff like relatively minor redline changes to contract terms.
Also, by "good" I mean maximize revenue/sales/renewals. Not specifically delivering something good in the end.
The role is organisation and team psychologist more than API design. If you want to rate his performance, the products are a worse indicator than, say, how well-adjusted and mentally stable his teams were...yeah, you're right. GCP is fucked.
Although I'm not so sure what Google's style is, it's been changing so much in the past few years, and a lot of their slimy stuff has been exposed too.
(To be clear, they are not shutting down GCP, this is just a leadership transition)
I could never see myself making the decision to go with Google over Amazon or Azure, simply because I don't feel confident that the product won't be shutdown with minimal warning at any given time.
Yes you can save money by killing a product, but I would argue the second order effects are more important. People lose the faith to bet on your products.
Having had to scramble - twice - when google cut products under me, I would never make a big bet based on an assumption that they will keep faith after I commit to their products.
Maybe they should read H Edwards Deming e.g. "maintain consistency of purpose".
The problem with these kinds of abstractions is in order to avoid "vendor lock in" they are forced into a "lowest common denominator" deal where you don't get to leverage any vendor-specific awesomeness. For example, in DB land you would avoid certain kinds of database fields (Postgres's structured datatypes like JSON, for example). Instead of leveraging the DB to do certain kinds of JSON queries you now have to write the infrastructure your self "just in case we switch DB's".
The irony is, in my experience to avoid the dreaded vendor lock in "problem", teams and organizations wind up locking themselves into their own homebrew nightmare that does nothing but recreate (often poorly) what they could have gotten by allowing themselves to get locked in to a third party. It goes without saying that in almost all cases, this homebrew nightmare has zero to do with the actual value generating functions of the business. Storing and querying on structured JSON in a database has nothing to do with the core values of 99.999% of all organizations--these things should be locked into third parties.
My point is, most teams should embrace all their cloud providers can offer. Settling for the lowest common denominator shared between cloud providers leaves money on the table and runs a very high risk of what is, in my experience, a worse form of vendor lock in--getting locked into your own stinking pile of rubble.
I could go on and on about the "vendor lock in" boogyman--it is one of the biggest, most annoying anti-patterns I see. It creates huge political and technical hurdles as the business grows. I believe it stems from a poor understanding of actual business risks, not understanding the true lifetime costs of inhouse engineering projects, resume-driven design, and a very healthy dose of premature optimization. (also toss in some Stallmanesque mistrust of proprietary systems for good measure)
Though a lot of the docker tooling, k8s and other options do make it quite a bit easier as an option of you want to self-host. But that has other issues, advantages and disadvantages.
On the other hand, in the past I’ve used Fog to abstract our cloud provider and that was nothing but pain. SQL is well-specced with some vendor-specific extensions, while there really isn’t any actual common denominator for cloud services.
FWIW I'm pretty sure Google projects have at least as good a life expectancy as startups do.
As mentioned above, this is a huge red flag for me.
Source: Ex-Oracle working under his leadership seeing his decisions first hand.
While votes are usually reserved for content quality, HN also finds them acceptable for disagreement in opinion, so most likely people disagree that stock K8S is easier to setup than the managed offerings.