It's interesting that you never saw something like this (landscape/hellscape) with OpenStack. I think it really drives home that Kubernetes isn't a platform in the same way, you use it to build a platform.
I'm founder of a startup in the space (robusta.dev) and in my opinion Kubernetes standardizes and formalizes the complexity not creates it.
If you want to develop a cross cloud app with autoscaling, self healing, health checks, service discovery, network policies, etc then you're going to have complexity. The question is whether you want to invent it yourself or use a platform like k8s
Anyways, I know people like to hate on the cloud native landscape but the reality of the situation is that there is a large healthy competitive ecosystem out there and a lot of options.
At best it's pointless. Everything is apparently cloud native, so why even bother?
At its worst the cloud native landscape is complexity nightmare. Why are there 12 API gateways? Are they all equally good? They certain don't do the same, well mostly they do, have fun figurering out which one to use. Almost none of the product are named all that well, so that's no help.
Right now the cloud native landscape is just a collection of pretty icons that look kinda cool. It doesn't really mean anything. Best I can tell: Graduated doesn't even mean that you're guaranteed that these product will work together... They might.
I get that others look at the cloud native landscape and see endless possibilities, a smorgasbord of options, and that's wonderful. I just see complexity, confusing, and shameless self-promotion in an attempt to ride the Kubernetes/container wave and make a name for yourself and maybe a few bucks.
Don't get me wrong, many of these tools are fantastic, but that some people can look at that massive graphic and think that the industry is moving in the right direction it troubling. Half of these product are dead-ends. We just don't know which yet. The room for mistakes are enormous and some poor soul will pick the wrong API gateway and be stuck with it for the next ten years.
How do you measure or understand the end user usability of these landscapes?
Because I actually like the content in them but am just totally baffled and annoyed at trying to use it. A text table I could slice and dice with a few dimensions would be infinitely more usable.
These landscape diagrams are fantastic, available in multiple formats, open source, and look great. I wish every market landscape was formatted the same way. Any plans to make a "landscape generator" tool of some sort available for non-coders?
I built out my first ISP in 2001, worked in FAANG by 2006, haven't the slightest comprehension of what this is for. Is it some kind of certified partner directory?
This is a list of "things" in the cloud native space. Some of them are from CNCF members and some aren't. Some are paid products and services while other things are entirely free and open source.
Many (most?) of the things are tied to organizations and they try to show information on those organizations.
The CNCF doesn't have certified partners. It is a sub-foundation of the Linux Foundation and a non-profit. The CNCF does have members. Some of them are vendors and some are end users. Things listed in the landscape don't need to be from a member.
You can reasonably argue about what should be included but I have a feeling if you started trying to draw a line about what databases are properly "cloud native" and which aren't, you probably end up making a lot of somewhat arbitrary calls.
I'm not sure I disagree with you but I also understand why they might be there.
It seems like Kubernetes has reinvented JCL from the IBM Mainframe
From the Wikipedia pager for JCL:
"Job Control Language (JCL) is a name for scripting languages used on IBM mainframe operating systems to instruct the system on how to run a batch job or start a subsystem.[1]
More specifically, the purpose of JCL is to say which programs to run, using which files or devices [2] for input or output, and at times to also indicate under what conditions to skip a step."
Once upon a time, the landscape in hadoop and bigdata had something like this. Projects, companies and startups with weird animal names. I miss those days where you couldn't tell a difference it was name of a bigdata project or pokemon.
28 comments
[ 3.6 ms ] story [ 78.0 ms ] threadIf you want to develop a cross cloud app with autoscaling, self healing, health checks, service discovery, network policies, etc then you're going to have complexity. The question is whether you want to invent it yourself or use a platform like k8s
https://landscapes.dev
There are also two sub landscapes in the CNCF landscape in case you're interested:
https://landscape.cncf.io/serverless https://landscape.cncf.io/wasm (still in beta)
There's also a secret feature that shows who is raising funding etc in the space: https://landscape.cncf.io/funding.html
Anyways, I know people like to hate on the cloud native landscape but the reality of the situation is that there is a large healthy competitive ecosystem out there and a lot of options.
*Disclaimer: Mostly happy user of many of these projects
So basically an accurate representation of the market. :)
At its worst the cloud native landscape is complexity nightmare. Why are there 12 API gateways? Are they all equally good? They certain don't do the same, well mostly they do, have fun figurering out which one to use. Almost none of the product are named all that well, so that's no help.
Right now the cloud native landscape is just a collection of pretty icons that look kinda cool. It doesn't really mean anything. Best I can tell: Graduated doesn't even mean that you're guaranteed that these product will work together... They might.
I get that others look at the cloud native landscape and see endless possibilities, a smorgasbord of options, and that's wonderful. I just see complexity, confusing, and shameless self-promotion in an attempt to ride the Kubernetes/container wave and make a name for yourself and maybe a few bucks.
Don't get me wrong, many of these tools are fantastic, but that some people can look at that massive graphic and think that the industry is moving in the right direction it troubling. Half of these product are dead-ends. We just don't know which yet. The room for mistakes are enormous and some poor soul will pick the wrong API gateway and be stuck with it for the next ten years.
Because I actually like the content in them but am just totally baffled and annoyed at trying to use it. A text table I could slice and dice with a few dimensions would be infinitely more usable.
Many (most?) of the things are tied to organizations and they try to show information on those organizations.
The CNCF doesn't have certified partners. It is a sub-foundation of the Linux Foundation and a non-profit. The CNCF does have members. Some of them are vendors and some are end users. Things listed in the landscape don't need to be from a member.
FYI it's also an open source project that others can use and build upon! https://github.com/cncf/landscapeapp
I'm not sure I disagree with you but I also understand why they might be there.
From the Wikipedia pager for JCL:
"Job Control Language (JCL) is a name for scripting languages used on IBM mainframe operating systems to instruct the system on how to run a batch job or start a subsystem.[1]
More specifically, the purpose of JCL is to say which programs to run, using which files or devices [2] for input or output, and at times to also indicate under what conditions to skip a step."
That’s scary.
Seems like this isn’t really saying anything. But maybe useful to find things to look at further?