Google IoT Core will be discontinued on Aug. 16, 2023 | Hacker News Reader

Google IoT Core will be discontinued on Aug. 16, 2023

215 pointsborner791159 comments

Your current IoT Core Services will remain available through August 15, 2023. Start your migration to alternative solutions. Hello [NAME],

We’re writing to let you know that Google Cloud’s IoT Core Service will be discontinued on August 16, 2023 at which point your access to the IoT Core Device Manager APIs will no longer be available. As of that date, devices will be unable to connect to the Google Cloud IoT Core MQTT and HTTP bridges and existing connections will be shut down.

Your current IoT Core Services will remain available through August 15, 2023, unless you terminate your usage of IoT Core at an earlier date.

What do I need to do? We recommend that you take action early to migrate from IoT Core to an alternative service. As an initial step, connect with your Google Cloud account manager if you have questions about your migration plans. Your account manager can also help you learn more about Google Cloud partners that offer alternative IoT technology or implementation services that meet your business requirements.

Over the next year, we will continue to reach out with additional information to support you during your migration.

—The Google Cloud IoT Core Product Team

View on HN

If you haven't, please read the classic "Dear Google Cloud: Your Deprecation Policy is Killing You" [1] by the always fantastic Steve Yegge (who used to work at Google). As he so succinctly put it, the email follows the standard Google GCP template:

Dear RECIPIENT,

Fuck yooooouuuuuuuu. Fuck you, fuck you, Fuck You. Drop whatever you are doing because it’s not important. What is important is OUR time. It’s costing us time and money to support our shit, and we’re tired of it, so we’re not going to support it anymore. So drop your fucking plans and go start digging through our shitty documentation, begging for scraps on forums, and oh by the way, our new shit is COMPLETELY different from the old shit, because well, we fucked that design up pretty bad, heh, but hey, that’s YOUR problem, not our problem.

We remain committed as always to ensuring everything you write will be unusable within 1 year.

Please go fuck yourself,

Google Cloud Platform

[1] https://steve-yegge.medium.com/dear-google-cloud-your-deprec...

  [..] in the Google world, deprecation means: “We are breaking our commitments to you.” It really does.
  That’s what it ultimately means. It means they are going to force you to do some work, possibly a large 
  amount of rework, on a regular basis, as punishment for doing what they told you to do originally — as 
  punishment for listening to their glossy marketing on their website: Better software. Faster! You do
  everything they tell you to do, and you launch your application or service, and then, bang, a year or
  two later it breaks down.

I didn't know Steve was a Kubernetes admin!

3 years ago Google killed off Android Things[1] and told everyone to migrate to IoT Core. Now IoT Core is dead. But don't worry about all that, just connect with your Google Cloud account manager! They'll tell you exactly which Google platform you should migrate to next.

[1]: https://android-developers.googleblog.com/2019/02/an-update-...

This reminds me of a David Letterman skit where Letterman was working at a Taco Bell drivethru. He told a customer that they were out of several drinks in a row and then said something like “ma’am, I need to know what drink you want so that I can tell you we’re out of it” in response to the lady asking what drinks did they have.

Reminds me of an old Soviet joke. A guy walks into a shop, sees how empty it is and asks "You're out of beef?" The clerk explains, "No, the butcher's shop across the street is the one who is out of beef. At this store, we specialize in being out of fish".

Like the good old Monty Python cheese shop :) https://youtu.be/Hz1JWzyvv8A

Yeah IoT definitely seems to be following Google's messaging strategy - if the platform isn't a runaway success, delete it and start again. Maybe people will love the new platform!

Are they too spoiled by the early days of the internet when that would happen regularly because they were simply first? Does it ever still happen?

"spoiled" in the sense of having gone bad, maybe. It does appear this is a deeply ingrained cultural assumption they aren't getting rid of (no clue if they try). The usual meme is of course "nobody gets promoted for maintaining a not-so-successful service"...

Why I only target Android with the NDK, or Web apps?

Because while they are natively supported on the device, they aren't tied to Android's future.

Google should just go ahead and shut down their entire Cloud business. I mean, nobody in their right mind would build a business or anything substantial on top of any Google provided service, and the hobbyist market can't be that big. Just go ahead and drop out Google, and quit fucking wasting everybody's time.

Hobbyist market is also not using gcp and more likely using digital ocean or similar

I just read an article that Discord is heavily relying on Google Cloud [0]

[0] https://news.ycombinator.com/item?id=32474093

Luckily, it sounds as if they’re relying mostly on lower level primitives vs. these higher level abstractions that are more likely to be discontinued.

Yes. They already angered the game dev community, and few believe Stadia will stay around. This angers the industrial internet-of-things community. You can get Siemens industrial equipment tied to Google IoT Core. Hopefully you don't already have it installed. Industrial controls people think in terms of decades, not months.

A Coursera class on Google IoT Core for industrial control started today. 77,000 people signed up.

> This angers the industrial internet-of-things community.

Hopefully they'll learn that if it's locked down to being tied to one person's cloud it's likely to become a paperweight in a few years.

I still think it's generally terrible style to fully depend on external services (and should probably legislatively disallowed), but there are differences. You can probably get a contract over a decade from some vendors.

>few believe Stadia will stay around

I think the people that still think Stadia is gonna shut down any time soon are just the gamer blogs/communities that have been running shutdown rumor pieces every month or so for the last 3 years. (Most recently the killedbygoogle guy's "my friend heard on facebook that Stadia's shutting down..." post that went viral.)

Stadia's actively signing hardware deals with Samsung/LG and other smart TV manufacturers, and about to expand availability to even more countries. AFAICT most people that actually play on Stadia don't think it'll shut down any time soon.

Twitter users who apparently play Stadia tell quite a different story in every encounter I've had with them.

I play Stadia. Couples of weeks ago, ToS change. Couple of days ago, an ad on my LG TV that if i sign up for Stadia i get 3 months free, and the Stadia app is very prominently shown in the Applications menu. Same on my ISP's Android TV box. Each month ~10 games get added to the Stadia Pro tier.

Stadia isn't going anywhere soon, or the people working in the division haven't been told and have continued spending money left and right with partnership deals, which would be a weird waste of money.

'Stadia isn't going anywhere soon, or the people working in the division haven't been told and have continued spending money left and right with partnership deals, which would be a weird waste of money.'

LOL, that is Google 101.

Allow me to introduce you to the Google Graveyard[1]

[1] https://killedbygoogle.com/

Very few, if any, of those killed by Google things, are products one paid for, let alone had paid partnerships with third parties. The situation really isn't comparable.

They just killed IoT Core.

Well said, I had no words- this is literally a thread about them doing just that.

Also not just killing but making worse like the Nvidia Shield.

How are there still googlw apologists?

Interestingly enough, if you look at threads like this, the comments almost always boil down to "pro-Stadia" versus "anti-Google" sentiments.

Not to make more generalizations, but I think you'll find that people that like Stadia like it because they actually like the Stadia product (surprise!) -- and people that hate on Stadia often do so because they have other issues with the company behind it.

Sure, Google might shut down Stadia someday. That'd be unfortunate but not too unexpected. In the meantime, I'm pretty sure enjoying a product that's been around for years and shows no indication that it'll go away any time soon doesn't make you an apologist for the megacorp behind that product.

I have no problem with people liking stadia. I'm against cloud gaming because of its own issues. I am against google because of its owm issues.

I'm sure one could be pro stadia/cloud gaming and still be cognizant of the garbage google pulls.

I am glad you are enjoying stadia, and I truly hope it'll be one of the services that survive, maybe it'll improve cloud gaming issues to the point I'd be on board.

But the op was in here touting google rarely closes paid offerings, ignoring they dont have many paid offerings and those are every bit at risk of being 'googled' like all their other services.

Google's own track record is every reason for me to put zero trust in them.

OP said nothing of the sort. All they did was list some (substantial) differences between the comparisons. They certainly might feel that way, but there's not enough in their post to conclude that.

> How are there still googlw apologists?

It's not Google apologizism to state the obvious that Stadia is getting money and time thrown at it to this very day, so it's unlikely that it will get killed soon, and pointing out the difference with the majority of the killed products which were usually on maintenance mode for prolonged periods of time before being killed (IoT Core included), not to mention in 99% of cases, free.

Stadia has the best UX by far of all cloud gaming platforms. It needs Google to show long term commitment by signing a few more big third party deals, and improving hardware and thus stream quality. Without that it will languish as a niche, with it it could get successful.

Which hasn't received updates since 2019. Unlike Stadia, which not only gets updates every month, has recent ads in third parties.

Google are still investing money and in Stadia. It's unlikely they'll shut it down soon. In most of the things they killed, there had been no investment for some time before the official announcement.

This year GDC Google announced a new support libraries for migration into Stadia, as very few studios want to mess with Linux and Vulkan, and then be told Stadia is no more.

> Coursera class … started today… 77000 signed up

That’s actually just a Coursera marketing tactic. Most of their courses are not real-time and are self-paced content.

So, they act like they are starting the course “Today” every day, and the 77k enrolled is the # of enrolled users forever since the course has been offered. Totally misleading!

And tensorflow

Google Cloud (VMs, blob storage, databases) bring in multiple billions of revenue a year from enterprise. The company invested billions in new data centers and transferred large parts of intenral technical infrastructure into cloud to build these products over a decade.

If Google Cloud (VMs, blob storage, databases) shut down, not only will large # of enterprises howl (data gravity is huge), Google will absolutely have no future in any enterprise product ever again.

I do expect them to aggressively trim small products from the cloud lineup that aren't worth the engineering investment. IoT clouds were a popular thing for several months and then people realized there wasn't a huge market or tons of profit.

Large revenue numbers don’t guarantee large profits. Sometimes you’re just losing more money, faster.

AWS doesn't do this. AWS keeps around old deprecated services that they aren't developing new features for, in perpetuity (e.g. SimpleDB). They might not launch that service into new regions, but they will keep it around in existing regions to not break customers who depend on it.

This is a reflection of AWS being "customer obsessed" where Google is not.

If they were "customer obsessed" even though they are not doing development of a service, they would lauch it anyway in all regions.

By not doing this, they make it hard for their customers to be multi-regional.

It is easy to spin "customer obsession" any way you like.

That's the least generous, and least logical, possible take, and I'm the opposite of a fan of Amazon or AWS.

If a customer is not multi-region already, and if Amazon hasn't promised multi-region availability, then the customer has simply chosen the wrong product. It's not in anyone's interest for Amazon to waste resources rolling out a product that they can't sell, just because maybe someone wants to expand their existing use.

Exactly. AWS never promises every product in every DC. If you build on it before it's available in a region you need it in, it's a choice and risk you are making

There's a lot of money in IoT; it's just not rapid growth. It's a business line that will require years of enterprise sales followed by the integration work. This video of from 2019 shows Volkswagon starting a rollout of AWS's IoT cloud for 122 factories world wide. I don't think Google is good at running a business with multiyear sales cycle followed by multiyear scale-ups to starting making money.

https://aws.amazon.com/solutions/case-studies/volkswagen-vid...

As a former head of engineering for a mid-sized IoT company with many tens of thousands of devices deployed in the field using a competing product (no, not the AWS one), seeing this made my stomach churn.

This was my second biggest fear after waking up to a ransomware attack.

It's hard to imagine anyone trusting Google for IoT again. I will certainly put them at the bottom of my list for any other infrastructure I develop against in the future, and ensure that we have a documented exit strategy should it come to pass.

The idea of having just one year to develop against a new IoT core, test it, update all deployed devices, and then coordinate logistics and budget to do truck rolls when things invariably go wrong is really grinding my gears.

I feel for all of the startups having to deal with this. To the folks who are invariably scrambling, I really hope you either got advanced notice, or you're getting an extension far beyond what is publicized. Edit: The more I think about this, the more I want to believe there must be contracts in place for certain customers that extend the lifetime of this product beyond what is public. There must be.

IoT is not an easy business. Designing and programming hardware is hard. Supply chains are hard. Maintaining working inventory is hard. Building logistics networks for installation and maintenance is hard. Courting and explaining to investors why you don't have the profit margins of a pure SaaS business is hard. Relying on your cloud provider to give you more than 1 year notice should be the easy part.

Had I started an IoT company, I would not have depended on any cloud IoT product. VMs, blob storage, containers, load balancers, hosted databases. But not anything labelled "IoT".

I have worked in an IoT company, and sometime time-to-market makes uncomfortable dependancies necessary. Hopefully people making those compromises have escape plans, but I doubt all of them do. I expect to read several "Our incredible journey" product end-of-life posts as Google kill off startups without a way to recover from this shutdown...

I worked at an IoT company, and I saw a few demos of IoT cloud solutions. We never found them to provide a lot of value. They were solving the easy problems and ignoring the hard ones.

> Relying on your cloud provider to give you more than 1 year notice should be the easy part

Isn't that exactly what contracts are for? Can you build an entire company on the premise that some random cloud service will always be available?

I fully agree that having more than just one year to migrate it's a bit short. Especially within the lifecycle of the devices that counts in decades.

Changing the ingress endpoint shouldn't be a big deal via OTA or configuration change. The lack of such configuration in the first place would be concerning for any new devices.

Ransomware attack it's a real threat.

> It's hard to imagine anyone trusting Google for IoT again.

I mean, presumably they decided the IoT vertical just wasn't for them altogether.

If Google intended some new service to take this one's place, they would have 1. launched it before deprecating this one; and 2. built a backend shim to route data sent to the old API into the new backend, so that people's code wouldn't have to change. (Like they did for Firebase, and for Stackdriver, and for anything else they actually cared about keeping the business of the customers of.)

This move, meanwhile, clearly sends the message of "we don't want to be in this business; stop trying to buy this kind of service from us; just go away." It's the feel of being on the receiving end of the "fire your [bad] customers" advice — just applied to "firing" an entire (bad?) market.

There’s no doubt that they have invested a lot of money but that doesn’t guarantee returns and data centers have a fair amount of cost to keep running.

The big thing I’d like is to have Google Cloud broken out as a separate business. While it’s true that shutting it down would harm them, I haven’t gotten the impression that their management are especially concerned. As an enterprise customer of both, AWS seems way more motivated not just to develop their service but also just to do things like show up. Trying to get GCP people to sell things like Anthos was surprisingly hard, like they thought it was 2005 and people would buy just because of their name.

I reward AWS with my business and tolerate GCP when necessary.

What's funny is there was a core of engineers back in 2009 who understood everything necessary to get the technical side of Cloud to be competitive with AWS, and mostly succeeded, but by then, it was clear that Google simply didn't know how to market to enterprises. The whole thing was a squandered opportunity and I simply did not understand at the time (2009-2011) just how unprepared Google's leadership was to expand beyond ads.

Strong agreement - I really think ads’ strength cost them multiple markets just because nobody was really worried about not being profitable. I’ve heard mixed things about working at AWS but everyone I know who’s worked there was keenly aware of whether customers liked what they worked on.

I simply did not understand at the time (2009-2011) just how unprepared Google's leadership was to expand beyond ads.

That's a succinct and on-point observation.

My snide remark is Google the hive mind doesn't understand what a customer is. As in someone who in return for you promising to do something on an ongoing basis will give you money. Googlers react to that as if a street hustler is trying to get them to play three-card monte.

More likely they have the same problem Intel had. Intel had a lucrative impossible to replicate business in processors. As a result they regularly abandoned one market segment after another because they couldn't get the gross margin they expected. Googles doing the same exact thing.

Other problem is googles business is backwards in that they set the requirements out for what they'll do not the supposed see above customer. All these other business ventures have customers with strict demands that they want met.

How is IoT a hobbyist business?

I deployed a large restaurant equipment project on Azure backends and couldn't be happier.

I've made a policy of not touching Google for anything embedded-related and it continues to pay off.

I guess, yeah, anything non-AdWords related is a hobby. For Google.

I think he's saying sane businesses will stay away from GCP, leaving the hobbyists, and there aren't enough of them to sustain GCP.

Not just GCP.

I think the parent comment was saying that Google cloud is for hobbyist because no one in their right mind will build a business that depends on Google cloud offerings.

They didn't say IoT is a hobbyist business. They said no serious company in their right mind would use a Google IoT product, so the only market they have is the hobbyists who aren't deterred by the fact that Google only keeps these services around for a couple of years before shutting down.

Yeah, I misread that. Thanks.

Yes, correct. I probably could have phrased that more clearly. Sorry. :-(

Are you aware Snapchat has a two billion dollar contract with GCP? GCP isn't some rinky dink hobby computer company.

well then maybe Google should stop acting as one. IoT core was used by serious companies.

That was reported in 2017 and was supposedly a five-year contract.

2017 + 5 = 2022 :-P

https://www.vox.com/2017/3/1/14661126/snap-snapchat-ipo-spen...

GCP isn't some rinky dink hobby computer company.

So they suckered one big company into signing a deal with them. Hell, maybe they even suckered a handful into doing so. But there are good reasons that Google is barely (if that) playing on the same level as AWS or Azure. Last I saw they had about 10% of the cloud computing market, compared to well over 50% for AWS and Azure combined. And I don't think it's any kind of stretch to suggest that Google's reputation for lacking long-term commitment to its products is one of those reasons.

And make no mistake... I take no glee or joy in saying this. I was a google fanboy (like many people my age) for quite a few years after Google first came along. And I'm still a fan on many levels. I don't want Google to fail, I just think they are going to fail (at selling to the enterprise at least) if they keep displaying the behaviors they keep displaying. Now the Search and Adwords businesses might sustain them indefinitely, so I don't expect them to go out of business. But as far as selling SaaS services, especially ones targeted at enterprise users, it's clear that Google just does not "get it." And I doubt they ever will. So unless they're going to find a way to change that state of affairs, I suggest they just exit this line of business. It would be easier for everybody, IMO.

We've moved off GCP for exactly this reason recently. It's enough. They're probably not going to shut down the lower level infrastructure, but they simply can't be trusted.

Unfortunately they have this weird dichotomy where they have some of the most outstanding and ahead of the curve cloud services that make it worth it. This IoT Core thing is a shit show and I imagine they will introduce a partner who will do it instead of them.

But there is nothing like BigQuery on the market elsewhere, Cloud Run and PubSub are standout services that others have tried to emulate and haven't got near it, plus everything on GCP doesn't feel like either an afterthought (AWS) or a cobbling together of mismanaged and old-version open source services (AWS & Azure).

Not every product is in the experimental phase.

Google has a long history of experimenting with a product and then shutting down once they've learned enough. Maybe they learn it's not as lucrative, maybe they learn how to build the tech, maybe they learn they're too late, whatever. But it's experimental and the experiment doesn't prove to be the "next billion users" so it wanes and then dies.

Not every product is like that. Cloud isn't just some product on a spreadsheet next to IoT Core. The heavy players like Cloud, YouTube, Ads, etc, have massive investment, have dedicated CEOs, and are in it for the long haul.

Google's deprecation policy sucks, but it doesn't apply to every product uniformly.

I'm in the process of migrating PrintNanny.ai's remote command/event system off Cloud IoT Core. I've been running on IoT Core for 1.5 years. Here's my breakdown of the costs...

- $236.99 in usage, approx 1% of project's total revenue

- ~20 hours to implement pub/sub applications running on a mix of Raspberry Pi & GCP VMs. Implementations were in Rust and Python. It would have taken much, much longer to stand up a managed MQTT broker and identity/key management that I felt comfortable using in my own home, let alone providing to customers.

- Hundreds of hours implementing and debugging glue between GCP's Pub/Sub product, websocket-based subscribers, and MQTT subscribers/publishers.

I don't regret my decision (wouldn't have shipped otherwise), but I'm looking forward to the next phase. Here's what I'm migrating towards:

- NATs message broker. NATS supports connections via MQTT and Websocket protocols, besides NATS own protocol.

- django-nats-nkeys for org, identity, and JWT management (not production-ready, don't use this until I've been eating my own dog food for a few months) [1]

- AsyncAPI schemas [2] for core message APIs, including schemas for 3rd-party printer software events (OctoPrint, Moonraker, Repetier, etc). This will underpin PrintNanny's plugin system.

[1] https://github.com/bitsy-ai/django-nats-nkeys

[2] https://www.asyncapi.com/

Don't hesitate to join the NATS slack group and reach out if you need any help, we're a friendly bunch

Oh very cool, see you in there! Thanks for the work you've put in.

Good to know, perhaps over a Martini?

NATS doesn't support MQTTv5 so for serious MQTT usage it is still useless unfortunately.

We are waiting on demand for it. We have a bunch of folks using MQTT on NATS and NGS but no requests yet for MQTTv5 so far.

Once we see legit demand we will make sure it's supported.

I think its a vicious cycle - No one supports it so services don't implement it, and because no services implement it people use hacky ways to do what MQTTv5 does out of the box or use services they normally wouldn't because the ones they want only support MQTT 3.1.1.

We use NATS for pretty much all of our MQ / Service Bus type stuff, we were using GCP Iot, AWS IoT and recently moved to EMQX as it has native MQTTv5 support.

NATS would have been a much better choice for us but the support wasn't there.

Build it and they will come :)

Would your company be willing to fund the development? Or would you be willing to post a PR? We would be happy to take a look.

Funding possibly - PR not likely as we dont work in Go. Will socialise and come back to you

hahaa; wow. So, we were doing this sensor project; and I picked "boring" things like raspberry-pi, python scripts, wiring the GPIO with a screw-down terminal or using SDR for other off-the shelf sensors to build the network. And in one of the demos showing our very low-budget type project a reviewer said: "you should look at IoT managers like the G offering, we're using it". So, they declined to use our methods and built their own around G-IoT. It's important to "own" what you can in your stack otherwise this vendor-driven-churn is forced on you and is outside a schedule you control. Sure, 365 days is a lot of time to migrate -- which, IME, leads to "we can fix this later" which then leads to "oh crap!"

I think it’s important to “own” what makes you different.

Running your own servers isn’t that important and probably isn’t what makes you different. EC2, Azure VMs or whatever short term project Google is running for compute are all extremely comparable, and you loose very little by using them.

But if your business is owning and operating an IoT platform for your customers, you should invest in a high quality solution, not just buying the off-the-shelf tool. It may be that a provider’s offering is better than what you can do, but it better be a lot better if you use it.

The point is to invest heavily in what actually makes a difference, vertically integrate what matters, externalize what doesn’t.

Slightly disagree. The important thing is to externalise things that multiple providers support. So externalise hosting Postgres, because 20 companies exist that do that. Don't externalise a one of a kind service you can't build against using open standards. Insource that.

I think it is important to mostly be in charge of your own destiny too.

When my doorbell (Doorbird) goes off at home the following happens:

* A real chime sounds (battery backed, mains powered, no internet required) * Native app goes ape shit and twitters (often about two hours later) * Home Assistant app kicks in on queue and does what I tell it to: Speakers speak, SMS sent etc

A slight variation I prefer, and strongly advocate is ‘own (or control) what makes you successful’. It may not be unique, and doesn’t imply always building it yourself, but you should be in control of those things.

I hate the current trend of smearing your business logic across dozens of third-party providers with various business goals that rarely align with yours.

I'm also not confident that the constant churn of keeping up with API changes (if not outright deprecations like this) and costs of the third-party services end up costing you more than just doing it yourself.

Finally, what we've seen with Okta, Twilio, most recently MailChimp (which was used to attempt to attack DigitalOcean customers among others) clearly shows that these companies aren't magic and may not actually be any better than doing it yourself when it comes to security.

> Okta, Twilio, most recently MailChimp

One issue is that the bigger you become, the more interesting you are as a target to break. So if you're a small company using a big company as a vendor, you've probably introduced a weaker link.

But if you're still small as a SaaS, trust is an issue to acquiring customers. I'm wondering where the sweetspot is.

>I hate the current trend of smearing your business logic across dozens of third-party providers with various business goals that rarely align with yours. I'm also not confident that the constant churn of keeping up with API changes (if not outright deprecations like this) and costs of the third-party services end up costing you more than just doing it yourself.

I suppose the problem is that it's hard to evalute the total cost of ownership. Migration to a 3rd party and operations costs can be reasonably forecasted, but how do you evaluate business or product continuity risk? Without being able to evaluate that risk, the 3rd party service always seems artifically "cheap."

Can you buy insurance against vendors dropping their services? I've licensed some tech before that included statements that if the company were to go out of business, the code would be open sourced so as to not leave people high and dry.

I don't know why anyone would build anything that relies on a google offering these days. What are they going to discontinue next?

That's why I prefer to invest time and effort on AWS. I know I won't have to rewrite everything from scratch whenever Google feels like abandoning their projects.

AWS has never discontinued one of their services?

Not trolling, seriously asking (I don't use hyperscalers)

To the best of my knowledge, no. At least none of the AWS services I've used. They've deprecated support for some CAs and stuff, but that seems reasonable.

Not a full service but partial features like S3 Bittorrent support was removed.

Not that I'm aware of. They'd be fully justified in shutting down simpleDB since it hasn't had a feature update in over a decade, but they still keep it running. New accounts can't sign up for it and it only exists in the oldest regions.

They were going to retire old s3 path styles, but they changed their mind after customer feedback.

Ec2 classic networking is going away this month.

But that's ancient technology at this point.

Migration paths and assistance have been repeatedly sent out though - you’d have to literally not care about the system to have this impact you.

As in "the second AWS product launched", and has been deprecated since 2013, which is around the time GCP became more than just App Engine.

In fact the last ec2 classic instances are due to be retired... by today! Though I'd bet no one waited until the very last day to migrate :)

There have been rare changes, like mandatory API version upgrades or some features being turned off. I do not recall anything being removed without a (better) replacement available beforehand, let alone entire simply services being turned off. See simpleDB as the canonical example of long term deprecation but continued availability.

Disclosure: long time Principal at AWS. The above is my own observation and opinion

Yes: Amazon Sumerian.

There was an ML service before sagemaker that you can't start using anymore but I believe it still works for existing customers

They just retired EC2-Classic, one of the largest services they ever had

EC2-Classic was not its own service. It was a specific configuration of EC2. They retired that configuration by providing a full migration path into a new system that could be configured to perfectly replicate the old system if you wanted it to. Having gone through that migration myself with a bunch of services a long time ago, this isn't really a good example of a thing that AWS has killed off.

That's exactly what you will do - invest time, and energy, a lot of both because AWS is a victim of its own legacy. Everything is an after-thought, everything is messy and its genuinely painful to use on anything except some basic UI driven functionality.

EC2 and S3 and some basic hosting stuff is fine but newer services such as AWS Glue, Athena etc are a horrendous mess for anything other than 101 type stuff.

Supposedly the funding for Google Cloud will go bye-bye if they don't beat Azure by 2023. They won't axe all of Google Cloud, but I bet they'll start killing services that don't make money. (Remember: Google is an advertising company; loss-leading cloud services don't help them sell ads)

Forgive my ignorance, but from a business perspective, is Google Docs, Drive, etc. considered part of Google Cloud or is GSuite distinctly different?

G Suite / Workspace is not typically considered part of Google Cloud Platform (GCP), which is Google's equivalent/competitor to AWS and Azure.

They are pretty separate. It is like Azure (Google Cloud Platform) and Office365 (GSuite) for Microsoft.

From a business perspective it is part of Google Cloud overall. Seems like leadership doesn't stay long though. https://9to5google.com/2022/06/29/google-workspace-javier-so... https://www.linkedin.com/in/greg-tomb?trk=public_post_share-...

>Supposedly the funding for Google Cloud will go bye-bye if they don't beat Azure by 2023.

Source on this?

Also, Amazon was just an eCommerce store before AWS.

This article/rumor from 2018-ish:

https://news.ycombinator.com/item?id=21815260

I wouldn't put too much stock into something from 2018, which was a 3rd party report on a meeting in 2016.

The article is overblown but the point remains: if GCP doesn't stop shedding money, the business is not going to let it burn cash indefinitely on a fool's errand. There will be pull-back and we don't know what services (or regions) will be affected. This may be the first of a trend.

"Google Cloud still operating at a loss despite revenue, client wins" (https://www.ciodive.com/news/google-cloud-revenue-Q2-2022/62...). Their revenue increased but so did their losses by about the same percentage. They are still 14% behind Microsoft's marketshare. They are leaning into a niche of "data, analytics, security", which is so broad that it won't take much for their competitors with much larger market share to eat their lunch.

A year ago they said they'd be profitable by this time. A year before that they said the same thing.

It's kind of sad. All that technology and nobody wants them, like a nerd without a date.

nah the article was really overblown. 'we've approved budget for five years. and our goal is to beat azure.' more or less, but that doesnt mean in five years there wouldn't be more budget approved, and it didnt seem like an ultimatum connecting the two. businesses dont plan infinite years out.

while google cloud as a whole i dont see going away, individual products within it(like iot core) are gonna always be up for the chopping block

Given that it is a two horses race for most businesses, I guess they should be cleaning up their CVs.

I see Google is getting a spanking in this thread and I wouldn’t suggest for a moment that they don’t deserve it.

However, I hadn’t seen anyone mention the specific Enterprise API / product designation they rolled out over a year ago to deal with this kind of thing.

If I understood things correctly when they launched it [1] the plan was to start by saying which parts of the GCP platform you could confidently rely on with the implication that the other parts you should understand you’re using products and services that haven’t proven their long term value inside of Google and as a result things like this can happen and you should plan accordingly.

As far as I know this is the first thing to go since making that announcement but I’d be happy to be corrected as well.

[1] https://cloud.google.com/blog/topics/inside-google-cloud/new...

Ah yes, the "It's still in Beta" defense.

I've been a defender of GCP in the past, and I've been quick to point out that while Google has a reputation for killing products, GCP has been fairly stable. This, however, changes my opinion.

I've been successfully using Cloud IoT for a few years. Now I need to find an alternative. There's a vendor named ClearBlade that announced today a direct migration path, but at this point I'd rather roll my own.

We did this, EMQX + Kubernetes across multiple AZ's has worked very nicely so far and its cloud portable / on-prem if necessary :)

If you are looking to find a way to migrate from Google, Qubitro offers you a warm welcome.

The offer for six months of free credits as well as free technical support and more is here: https://blog.qubitro.com/migrate-from-iot-core-to-qubitro/

This is starting to feel like the Windows platform post MFC/Win32.

Every couple years, it was a new initiative to get developers and the middle management deciders revved up. A sort of a “we’re building it so that you’ll come, right?” thing. And then a year or so later, a new initiative replaced it, the offering was dramatically watered down, or just altogether sunset’ed.

In their defence at least MFC/Win32 is still patched and supported even if it's not getting new features.

Although it's probably not the same because they don't have to run any physical hardware to keep it going like IoT Cloud needs.

You can still get new project templates on VS 2022, bugs get fixed, and little things like HDPI support added.

A very big difference with Google's offerings.

So Google is basically breaking their commitments to their own users again, they don't give a fuck about back-compatibility as usual, they feel like the whole world should care of what their internal product teams want to do, and they feel entitled to set whatever arbitrary deadline for users to migrate (usually less than a year), and if you don't migrate then your money-making software will just stop working. Oh, and of course they provide no customer support whatsoever nor any way to provide feedback - because, of course, it's Google, and it needs to be dumbly unhuman and faceless to the core.

Google has become an abomination and a complete denial of what software is supposed to be. It deserves to die in a ball of fire, and none of its shitty products should be spared.

Wow one year to move off seems pretty fast for embedded customers. I'm going to guess they have very little real-world usage and customers.

I'd think IoT-related services (especially an entire gateway) would have years-long deprecation schedules, not a single year. If you coupled your product to cloud-based updates only (no companion app stuff), then this means your unsold inventory right now has a timer to be sold until it's dead and requires shipping back for manual updating.

Also means it's price should get discounted for that swapping out, and for "essential" stuff, like the things you stock spares for, your customers will have to go through two cycles of swapping.

What devices are there in the field are reliant on the Google IoT Core Service platform?

Are there ‘smart’ devices that are just going to stop working unless people rewrite the firmware for a different system like AWS or Azure IoT?

Exactly, everybody who bet on Google in IoT has just got a giant RMA bill.

Same has happened when Azure changed its root cert, which was likelly hardcoded into 99% of deployed devices using it.

Do IoT devices not have a USB port to allow for updating of the firmware?

A very large amount of customers are not technically competent enough to successfully update the firmware over USB. Also, from the business side, normally you would do OTA updates for deployed devices; now you need to build and deploy a customer-facing USB update tool for them to run on their laptops. And provide tech support for it, etc.

For industrial or B2B IOT you would in many cases need to send out your own field techs to do the update, which costs $$$.

So overall, even if you're not literally doing a RMA, it is very very expensive.

some do but you need to actually take every device - which you may have sold to customers - and update. Industrial stuff has multi decade contracts often in remote places.

Devices we worked with for a project (thankfully not on GCP) are welded shut once built. They are designed to run for years on an internal battery then need to be recycled. They would all be paperweights in this scenario!

IoT devices can be all kinds of things, from vending machines, HVAC systems, industrial machinery, vehicles, and so on. Some of it might not be updateable over the air, and can be located in all kinds of inconvenient places. I wouldn't want to be the one that has to manually deploy an update to sewer flow monitoring devices, or heavy mining machinery that's in the middle of nowhere.

Google was probably first when it came to managing 100s of 1000s of servers. They had the know how, they could have eclipsed AWS before they could even start. They weren’t too late to the game either, Azure started late and ramped up fast.

The scale at which Google managed to fuck up their cloud business says a lot about their DNA.

I was so excited about AlloyDB, but the documentation is crap. After two days of setting things up and dealing with their complex network configuration I gave up. Why would they make it so complex for a new dev to try their shiny new DB? Do cloud googlers seriously not think about new user experience?

Totally ridiculous. Sunsetting a major service like this in 2022 just proves their platform to be unreliable and unfit for enterprise-grade projects.

If there is anyone trying to find a way to migrate from Google, Qubitro offers you a warm welcome.

The offer for six months of free credits as well as free technical support and more is here: https://blog.qubitro.com/migrate-from-iot-core-to-qubitro/

Honesty, this is "just" MQTT/HTTPS ingress.

Further communication to Pub/Sub wouldn't change in any way.

IoT Core as a service has some design choices that make it attractive such as JWT token auth, and complex to optimize, such as communication pattern details.

If any of you would be interested in migrating to a fully compatible solution, give me a shout at rwarz[at]softserveinc.com since we are building one ;)

Use JWT over HTTPS to RealtimeDB. Has the same presence detection as MQTT. A lot cheaper too. I actually like it better than using MQTT. All of our devices attach to multiple RealtimeDB and then we sweep the data into Firestore for human interaction. Front end for RealtimeDB is open source so you can port it.

At this point it might be useful to think of what Google projects are unlikely to be deprecated soon, or rather, has the potential of surviving deprecation. For instance, would Flutter being an open-source framework be able to save it from Google discontinuing it? Or is having to update it to mimic iOS behavior across its Cupertino widgets with each update too much?

Well that's hilarious. And here I was always told that Google Cloud was supposed to be the one that didn't kill services.

Wait what? That’s like the opposite of what everyone here says. Like, Google has a bad reputation for doing this.

Sure, Google at large has a reputation for randomly axing products, but a frequent defense I see in such threads is something like "sure it might kill consumer products, but GCP is a different thing aimed at Enterprise that wouldn't do that". (I say that I see this argued, not that I hold the position myself.)

Pretty wild for IOT. Short of extensive OTA update infra being in place that likely bricks devices in the field

"Google Cloud IoT Core is being retired on August 16, 2023. Contact your Google Cloud account team for more information." -https://cloud.google.com/iot-core

I wonder what it’s like spending the best years of your life building vaporware for Google.

This doesnt surprise me, I wrote about this not long ago and how the Cloud Providers are only paying basic lip service to MQTT and IoT. There is not one single Cloud Provider that implements MQTTv5 properly (Azure implements a subset of it), and considering it came out in 2019 after being in "close to final draft" status for several years before that, the clouds have been very slow to pick this up.

I refuse to believe that anyone is actually using IoT Core on AWS or GCP for modern MQTT workloads. Pulling data in from a few "things" - sure, but industrial level capability across multiple systems, I really don't see it.

How could the costs of keeping a managed MQTT service running for a few more years possibly outweigh the negative press from this move? I just don't understand.

Looking at the release notes (https://cloud.google.com/iot/docs/release-notes) it seems like you could see it coming as there was no new features since 2019. This doesn't help GCP credibility compared to other cloud providers, AWS still supports legacy EC2-Classic from the early 2010s for god sake.

EC2-Classic went away literally today :)

I ignore every Google product / service since I know better. Well I do use search and youtube but that is universal I think

Google Graveyard is the second most consistent thing they’ve ever contributed to society.

Are we talking about this? https://cloud.google.com/solutions/iot

Yep, see the discontinuation banner at the top of https://cloud.google.com/iot-core

> Google Cloud IoT Core is being retired on August 16, 2023. Contact your Google Cloud account team for more information.

Gotta love the "retired" nomenclature -- as though it has had a long, productive career and now is the normal/usual time to end.

Perhaps one should just migrate to AWS IoT Core.

It is unfortunately the least garbage of most of the cloud providers, but until it supports MQTTv5 out of the box then it is useless for most modern MQTT deployments

If there is anyone trying to find a way to migrate from Google, Qubitro offers you a warm welcome.

The offer for six months of free credits as well as free technical support and more is here: https://blog.qubitro.com/migrate-from-iot-core-to-qubitro/

I don't work for Google or really care, but the GCP bashing dogpile every time they cancel a product is just getting pretty unimaginative. They offer hundreds to thousands of services, and some bets just don't work out. Glad to see many other people have perfectly functioning crystal balls.

They are free to shut down their services whenever they want - but good luck trying to win sales if they keep doing this. Especially given their reputation.

Deprecation as a Service?

Google regrets to inform you that its Google Graveyard service deprecation service (aka Project Pallbearer) has now reached the end of its life and is deprecated.

-This message generated by Google Graveyard service deprecation service

How are they going to keep the “google home” ecosystem alive without IoT Core ?

They said they were going full on the matter protocol but I don’t see how this will replace it fully yet.

There are so many different teams and organizations and acquisitions and such that I suspect Google Home has nothing at all to do with Google IoT. The IoT thing looks like it was focused more on being device management presumably to help funnel industrial and embedded customers in to using GCP and other Google cloud services. The Google Home stuff is probably all from Nest and such and entirely separate.

This is really going to make 5 people angry.

Reply on news.ycombinator.com