IBM Just Made A REALLY Weird Acquisition...
ฝัง
- เผยแพร่เมื่อ 17 พ.ย. 2024
- I never would have guessed Hashicorp and Terraform would end up at IBM, but honestly? Kinda makes sense. Time to rant about why
SOURCES
www.hashicorp....
opentofu.org/b...
newsroom.ibm.c...
Check out my Twitch, Twitter, Discord more at t3.gg
S/O Ph4se0n3 for the awesome edit 🙏
![เจ็บที่ยังฮัก - ม่อน วรวิทย์ [ Official Mv ] จอนนี่มิวสิค](http://i.ytimg.com/vi/KALIGSRSjFs/mqdefault.jpg)
I'm sure this will go great. Just look how good they managed Red Hat after taking over..
🙄
How much has redhat changed since then?
RIP Centos
Bye terraform. Hello opentofu.
Terraform will suffer the same fate as CentOS.
Which is..?
@@iuse9646 which is - becoming a test release for rhel, meaning that it's highly unlikely and risky to use centos stream for production
@@iuse9646 non existence. There's no more CentOS that matches RHEL releases
And alternatives will rise as Rocky Linux and Alma Linux
@@hanes2 Open Source is never dead thanks to the brilliant minds that fight against big corporations like IBM.
You can store the terraform state encrypted, and limit access to it, in any bucket or storage solution. You don't need Vault.
Multicloud is the latest big lie. Having a single tool that can provision and manage services across multiple vendors is NOT what Terraform delivers, because there's no abstraction between the objects being managed and the vendor-specific services behind them. Whether you use Terraform or Cloud Formation, you're using AWS's client libraries and APIs behind the scenes, and your code won't work against Azure or GCP. All you get with Terraform is somewhat similar syntax for some analogous vendor-specific services, and you get that at the cost of stability, support and vendor-specific-feature-lag.
I think libs failed at providing value, as you are right. But i think CF and bespoke apis is the issue. Pulumi / sst i think provides actual value by not shying away from what it is, but idk; we used CF a ton and i hated it. Want to explore pulumi/sst.
@@seannewell397 What do you hate? Honestly, just use the tools the vendors provide you. There's no reason use another tool just because you have a personal dislike for the syntax or whatever.
I wouldn’t agree with that. My company is provisioned across multiple cloud providers running flavors of k8s just fine. Our configurations are pretty simple and they overlay specific modules. Now if you’re using provider specific solutions, then yes, but honestly we’ve found them far more expensive than just hosting in a container. Arguably most of those cloud tools are just abstractions on top of abstractions on top of abstractions hosted on containers anyway. That said, for my current personal project I’m going self hosted. For what I spend on cloud in a month I’ve got enough hardware and redundancy to host almost anything that isn’t Ai heavy. I’ll use a global cdn when I’m ready to release it
@@tc2241 Not enough people are talking about self-hosting. I think it makes so much more sense! It does not take an FTE (for us) to maintain, and our cloud spend would cost us more than an employee so the value-prop just isn't there with cloud.
For sure, anyone using Terraform will have the illusion that it is some kind of abstraction that lets you seamlessly move your cloud infrastructure dispelled very quickly. It's not some kind of abstraction. It's just a universal way to statefully manage things. The nice part of that is it can link/provision different things together, like your CI system and your cloud infrastructure or secret manager, without resorting to piles of glue scripts or manual steps.
Honestly the apis for most major cloud providers are so robust that it’s not like TF holds you ransom. The only issue is that it’s convenient having everything in one place under a unified format, and not having update integrations yourself. However if you’re on a single provider, it’s not really a struggle to have your own tool
RIP hashicorp, you will be missed. Such an innovative, ground breaking company with so many industry standards. Was a fan of them for the longest time, feels like someone's dream that ended in corporate.
If IBM runs at a net less for some of the services it’s deficient in, the goal may also be to transfer from providing 100% cloud service, to cloud “add-ons” that can be easily integrated into an existing deployment at another service provider, without having to partner with a competitor
My prediction: most of this stuff is ending up on somewhere under Red Hat / OpenShift umbrella soon, which means you'll only get untested versions as open source from them.
...edit: Which, admittedly, looks like a better alternative to rest of us than what HashiCorp was doing with the BSL thing before acquisition.
I find the name BSL hilarious. bullshit license! How can anyone create abbreviations starting with BS? 😂
IBM Software Services grew by 5.5% in the last month. This is where Hashicorp will slot in. IBM Software Services will support near enough everything (even from other vendors).
I LOVE Pulumi. It is what we all should have been using from the start. The reason people don't like is because most people working in Infra came from an Ops background and are not very good at programming. Most people coming from a dev background love it though when they actually try it.
Hello OpenTofu.
"Of all companies IBM now owns terraform"........ Of all companies, lolz dude, this happens all the time in the tech, ie. Oracle owns Java
Multi cloud automation should synergise well with ibm's cloud platform as a service b2b enterprise automation toolset
People building companies open source or not need an exit strategy so they can make money from the effort. common strategy is to sell to larger company so not surprising particularly if you are not profitable.
In translation: IBM has many projects / clients relying on TerraForm..........
They will have from now on... everyone who makes extensive use of all those cloud datacenters, they need this kind of thing. They don't manage all those by hand, whether this is virtual servers, specialized serverless (as they call them) functions, gpu time, etc. in the same way that those datacenters don't really manage their prices manually either.
Weird thing about opentofu is due to the nature of the code as infra/service deployment, you don't have to code for effeciency to reach the feature parity with tf
The open source community should react by launching a CLASS LAWSUIT against the likes of Hashicorp, Red Hat...
They clearly misrepresent and MARKET themselves as OPEN SOURCE, whilst they actually are not!
If I were jury member in such a trial, I would clearly rule against these practices!
They are companies who can ultimately do whatever they want. This isn’t some mandated code welfare program
Now you understand why hashimoto was so mad against opentofu.
What did he say?
src?
Did you check who one of the supporters is of openbao?.... Drumroll.... IBM
Yup that will end. In fact maybe IBM will open source again.
In fact IBM encourages use of open tofu internally
@@blazer511 for now
You can use Azure storage account and dynamo db to save tfstate
Yes, this makes total sense if you consider IBM wants people to use their cloud. Probably won’t be good for consumers in the end but in the short term they’d be mad to ruin the products.
We’ve spent millions in the IBM cloud after they acquired SoftLayer. Since then the support has fallen off a cliff and they’ve made tons of mistakes. We are currently racing off of their platform.
As a former IBM employee, I was using the IBM cloud. But as you've said, not because I wanted to, but because I had to. Nevertheless, among many other companies wishing to acquire Hashicorp, IBM was one of the best consumer options.
Multicloud deployments are crazier than hybrid local / cloud (but you could at least make some justifications for the latter) Really, there is no reason to deploy a single application across multiple cloud infrastructures. The only possible reason I can think of is if you wanted to become horrifically vendor locked to cloud specific PaaS offerings. No, the reason IBM likely acquired Hashicorp is merely to own a stake in a technology that enables applications to be moved across infrastructures...Terraform underpins most deployments nowadays, IBM owning the technology essentially makes them a gateway to cloud providers owned by their competitors. Obviously.....I hope they do something good with it tho, devops sucks.
Opentofu released Go support for Terraform, i am pretty sure they will continue making even greater things in future than IBM lmao
As someone that uses Terraform ALOT, I think the most immediate action they will take is divert resources to the IBM TF Provider, because it’s absolutely terrible. The best Terraform provider would be the AWS one.
A bad TF provider for IBM Cloud really hurts its adoption, nobody doing TF wants to touch it. Now the question is, will this result in the AWS provider getting worse over time? Maybe, time will tell. If it does, that will hurt AWS badly.
And OpenTF/Pulumi both rely on the work that Hashicorp puts into the AWS Provider. You might be surprised to find just how much of Pulumi leans on TF to cover its gaps.
Azurerm provider is developed by Microsoft, AWS one is developed by Hashicorp?
@@michaszewczak7392 Hashicorp is the primary contributor to it.
@@michaszewczak7392 Hashicorp is the primary contributor to the AWS provider. Other people contribute as well, but to a much lesser degree.
Re: The whole "IBM Needs Multi-Cloud because it's had trouble getting adoption"
This is nothing new actually. I worked there from 2015 - 2020, and even back then this was the major thing they were pushing, except it was pushed as "Hybrid Cloud", it's just that IBMs business has come to rely less on physical hardware sales, as has the general hosting environment. In 2015 more companies were self hosting as opposed to today, because the cloud has just gotten so much cheaper and easier to develop on, in part thanks to things like Terraform.
The general approach of IBM, at least that I could surmise, was that IBM knew it would never be an industry leader. So instead they've set up strategic partnerships with or acquired companies that are effectively hedges on cloud as an industry. RedHat, Apptio, etc. IBM's vision ultimately is that while your cloud provider might not be IBM Cloud, the stack you use to develop/deploy/monitor your application is.
Personally, I don't see it really taking off. IBM won't ever die, but I think it's never really going to thrive again either.
State management is why crossplane seems like the future
True. Existing state should really be queried directly, not stored some other place which is bound to drift.
What makes anyone think Pulumi can make money either?
opentofu will fight ibm now...good luck! 😅
Indeed, since now LLMs are trained on source code like OpenTofu. And developers at IBM using LLMs like Copilot will now generate GPL licensed code for example in there copyrighted code. Forcing the company to fully open-source all their code, since after some time you don't know anymore which part of your code might contain generated GPL code and which aren't.
This is actually n.e.w.s. not related to JS. Good job.
Terraform is a professional stack. I consullted with about 3 companies that use it heavily. This is a surprisingly good buy by IBM. All of these side pet project companies doing cloud are going to get bought out by the big companies. Most of the small businesses suck at making professional solutions; they always make the user experience really bad (where you can't undo a change (like most nerds do)).
But If you're using terraform in your side business, I'd find something else.
Anyone who keeps an eye on DOD contracts isn't surprised in the least. The US military has a some large contracts out there for cloud solutions.
Also Theo, plz make a video on your development environment / workflow.
I really like the slick DSL of Hashicorp. Terraform is so good imo.
Pulumi just feels like a hack to me for people who can not fathom to use another language for infra.
So sad to see Terraform first get a shitty license and then this.
I feel robbed.
Terraform cloud doesn't make sense when you can use azure storage accounts and azure key vaults to hold your state files and secrets for cents per month. Other clouds have support as well, Terraform isn't a thing that makes money given how easy it is to run it on your own
Hi Theo, I really liked the video and your technical opinion on the acquisition. However, in the last 3 years I've been focused on receiving cloud certificates and haven't used them at all in my day-to-day job and pretty much share your opinion on them. I just wanted to ask what is your reason to not like certificates :)
When cash is no longer cheap to raise round after round many companies will be taken over by big tech.
first Red Hat - Ansible. Now Hashicorp - Terraform. Introducing "Terrible" ....
I knew Hashimoto and the beginning of Hashicorp back in 2012 or 2013... because of Vagrant. Then Docker appeared and I just moved from Vagrant and Hashicorpo... Good to know he made tons of money.
At least it wasn’t Oracle.
If it was oracle the forks of all their products would instantly be more robust and community supported.
At least it wasn't AWS
Don't forget about crossplane and winglang
glad we made the decision to stick with gruntworks, not terraform cloud and use opentofu ....
The will cut costs, fire the devs then after jacking the price up to high heaven. Make it open source with a preference for IBM cloud
You just made a video and summarized our CEO's mail update. Love it. Thanks.
Pulimi is great, was easy to work with when we were doing iac things
hashicorp over hired staff and didn't rely enough on the open source community. Then it would have been profitable which is a shame. I suspect IBM will halve the number of staff inside hashicorp and maybe move back to open source. lots of people use ansible and terraform together so better integration would also help. I bet the Ansible Automation Platform will get terraform added to it. You pay for the Ansible Automation Platorm. -) Selling corporates Redhat/Ansible/Terraform bundle will be good business as they like to pay for support. -)
A solution like palumi is where it's at, you can make your own primitives like a sqs backed lambda and build better blueprints
Pulumi is really great
I also used the IBM cloud when I worked there. It is actually not as bad as people like to pretend. A lot of it is unfamiliarity and lack of momentum compared to AWS and azure.
2 - because they acquired ns1 connect
Is anything sold for less than a billion dollar this days? What a hell! Speaks for true value of dollar (and money in general I guess)...
Ain’t now way I’m using cloud IBM DB2 💀
8:33 IBM cloud? what?
i'm more worried about packer..
RIP Terraform.
As a part of the IBMer it is a Great News, happy to collaborate with Hashi Corp, lets transform the cloud world
Let's be honest, Terraform never hit its stride. Now it's found the perfect graveyard.
AWS CDK
Theo plays roblox confirmed
SST constantly winning
that is very bad news, just look at what they did with redhat
Why are they stating Multi Cloud lol, why would you use anything other than a single cloud provider except with Cloudflare
I'm not knowledgeable enough to answer but I'm assuming he's saying that either the features or pricing for various things makes it desirable to split across different services. That being said, I work at a very large Fortune 500 company and I'm pretty sure EVERYTHING is in AWS, purely out of simplicity.
Rest in peace
you can already tell they are a good fit 🤣🤣🤣
IBM has a cloud? lol
They are going to centos hashicorp
rip terraform
Big L
i dont know why I keep clicking on videos with a talking head reading an online article.
I like to think I’m a less biased source of info then IBM and Hashicorp :)
NOOOOOOO WHYYYY
2 the quantum compute lol
Who the hell would want to provision infra with *Javascript*?!
You can t pay me enough to write JS tbh
Fork it all. 😂
L
This is yuck
Why do you always look like you need to poop in your thumbnails?
Cloud is overrated