Google, Red Hat, Ubuntu, Parallels bet on virtualization's surprising successor Docker has spun off a key open source component of its Linux Containerization tech, making it possible for Google, Red Hat, Ubuntu, and Parallels to collaborate on its development and make Linux Containerization the successor to traditional hypervisor-based virtualization. The company announcement on Tuesday at Dockercon in …
This isn't my area of expertise, but this looks more like a modernization of Big Iron implementations from before the rise of the Wintel server than a "burgeoning technology."
I've always blamed Microsoft for this whole virtualization thing. As I remember, it went something like this:
Here, use our new Windows Servers. They can do everything those big machines do for a fraction of the cost!
Whoops! One bad application can bluescreen a server, taking all applications on it down. We better only put one application on a server to protect ourselves.
Wow, our servers are severely underutilized!
Hey, here's an idea: How about we create "virtual machines," so each application thinks it's alone, but it's actually sharing resources. We'll need a new OS to handle all that.
So, we end up with a much more complex datacenter environment than we would have had if Microsoft had paid attention to the big boys (at the time) in the first place.
Oh, and Microsoft gets to sell more server licenses.
While I'm sure Microsoft's PR department is happy that you're giving them credit for virtual appliances (note: virtual machines run multiple applications that can bluescreen, or they would if they were really in evidence in 2001), but I'm pretty sure this is a Linux-only thing for now.
Good argument - that Windows instability propelled virtualisation. Hypervisors are quite an old technology, go back to 70s mainframes IIRC.
Virtualization has indeed escalated complexity in the datacentre. You still need all of the old skills, because they have been virtualised instead of being superceded. But you need all the virtualisation skills on top of that, across every area - networking, storage, compute.
One clarification on my OP. I never meant to credit Microsoft with the rise of hypervisors in the modern datacenter. All of the statements in my little timeline after the opening MS fanfare were meant to come from the mouths of IT managers. I think it's Microsoft's fault we needed them but, as mentioned elsewhere, they were pretty late to the party with an offering.
Is it like the venerable Solaris Containers or BSD jails?
An article that explains how it is different to virtualisation and/or jails would be nice.
Oh bugger it, I know how to use Google.
It appears that it is like an old fashioned container with a bunch of automation tools included. Or something. Doesn't sound too revolutionary but it is interesting none-the-less. Might have to fire up a server and try it out.
Sounds very much like FreeBSD jails, to be honest. Something which OpenVZ could do on Linux 10 years ago, too. Then came the para and HVM virtualisation Xen hype (okay hype is a strong word, as Xen is still very successful). And now we're back to containers. I hope this time around containers will be easier to manage on a bigger scale. Maybe docker can do that. I shall give it a spin.
To those who credited virtualisation or containers to MS above: wrong. They were very late to the game, and it hasn't even been working properly for a very long time. They were under pressure to come up with an answer to linux based virtualisation which was able to run Windows Server in instances.
EDIT: Solaris was first with containers, IMHO by a very large margin.
"This drastically reduces the number of moving parts, and insulates Docker from the side-effects introduced across versions and distributions of LXC. In fact, libcontainer delivered such a boost to stability that we decided to make it the default."
So Docker is the amazing thing, but libcontainer is the amazing thing, Docker does good stuff because of libcontainer. And libcontainer is included to reduce something else and make it better, Hang on. Docker wsnt stable and had side-effects, so we wrote some code to sort this out, which was Docker, but is now spun out of Docker, but is still in Docker. But is called libcontainer.
And sticking it out there will make everything work together. It was there because of the changes the lxc team kept making, so we'll let them change libcontainer. And that will work.
I wish I could write press releases.
Reddit here:
http://stackoverflow.com/questions/16047306/how-is-docker-io-different-from-a-normal-virtual-machine
For Sun people:
LXC is the zone/chroot bit. AUFS is the file system for inheriting read-only parts of the file system. Docker is the above plus a bit of cloning/snapshotting tech (often performed by ZFS in Solaris).
Cool. Guess the big surprise is that nobody implemented this till now.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
