Debian Virtualization: LXC Network Isolation

I have a lot of fun seeing all these different ways people are using to do the same basic job, "virtualization".

It reminds me of the much ink spent bemoaning the "unnecessary waste" of "competition" back at the start of the 20th century. In the name of efficiency, competition was eliminated by creating cartels.

Let's choose one full-system virtualization, umm, just do VirtualBox. No one needs VMware.

And for isolation, oh, I don't know, let's make everyone use sandboxes.

etc.

This wasteful competition between virtualization ideas is much more fun, and interesting, and while I expect there will be a narrowing down of projects over time, new projects will arize as someone gets a "great idea" of how to do it better.

What fun!

Virtualization is important enough that I'm happy to see as much competition as possible.

The "wasteful" argument seems superficially enticing until you get down to the question of "which one shall we concentrate on".

Then someone goes and forks it and goes in their own direction.

Besides, the whole waste argument is predicated on the idea that everyone is working to a single goal within one organisation for one purpose. Apart from that just not being the case, I'm sure that none of those people feel that they are wasting their time.

It's like suggesting that model makers are all wasting their time. Let;s just get one person to make the model, then everyone else can move on and do something else. Sometimes, the journey is the point.

It was a phrase used around the turn of the 20th Century. Two different drives were going on, one to prosecute cartels because of "monopoly practices", and the other to impose cartels to eliminate "wasteful competition".

As is obvious, it was a very confused time that was exemplified by Woodrow Wilson's near complete nationalization of industry under the auspices of WW1.

As Skelband says, "Sometimes, the journey is the point." Innovation cannot be forced, only fostered through the liberty of experimentation and play.

We have more than 20 window/desktop managers, hundreds of distributions, lots of text editors, mp3 players etc. Why would virtualization be different? I would say it is about choice, nothing to do with the journey.

i don't think people create competing products because "there needs to be choice" but rather because: "i think i know a better way to solve this problem or the existing solutions are not good enough." this is more about the journey. if i were not interested in walking the journey i would not work on that product.

greetings, eMBee.

@jazz: There are at least 3 perspectives, and they are not necessarily related at all:

1) From the user's perspective, it is nice to have choice.

2) From the developer's perspective it is because they want to do what they want to do.

3) From society's perspective, it is about multiple-evolutionary paths, and the avoidance of a mono-culture, which we should avoid at all costs. The bridge that saves us is emerging standards. Standards are a mechanism whereby we broadly agree to do some fundamental things in the same way for which there is no reason not to. Standards should emerge from necessity, not be imposed.

@skelband: I was talking from an user perspective, and choice came to mind. From a developer perspective, it could be a journey.

The field is too new to be standardized. We have VMware with 60% of the market, Microsoft Hyper-V with about 25%, and among the rest we have Xen and KVM. I don't think Xen or KVM are capable of becoming standards. They've been trying for ages to gain market share, and are still working on it.

There is still room for something completely different: LXC, OpenVZ, VirtualBox, etc.

Some may see "wasteful competition." Big deal. Eliminating the "wasteful competition" is the tactic of Gates and Jobs.

Remember what "free" in "Free Software" really means.

I'm with gus3 on this...

Quoting: Some may see "wasteful competition." Big deal. Eliminating the "wasteful competition" is the tactic of Gates and Jobs.

Remember what "free" in "Free Software" really means.

{Applause}

(plus honorable mentions to eMBee, skelband)

> VirtualBox, etc.

Given Oracle's history, I wouldn't count it being VirtualBox.

But, going by Oracle's record on FOSS projects, how long will Oracle hang onto VirtualBox?

> ...how long will Oracle hang onto VirtualBox?

I doubt they would be willing to sell it for anything resembling a reasonable price. So "hang on,? Probably forever. Developing and supporting it is another matter, and I don't see that continuing much longer. Fortunately, there is the OSE version which someone could pick up and continue developing. But we can thank Sun for that, not Oracle.

> we can thank Sun for that, not Oracle.

Indeed, Sun was a much nicer neighbor in the F/OSS camp.

> Remember what "free" in "Free Software" really means.

I expected we'd all be in agreement on the benefits of a diverse ecology.

I hope no one (Gus?) was confused by my parodying the arguments of the control freaks.

> Sun was a much nicer neighbor in the F/OSS camp.

I (mostly) disagree. The whole point of the CDDL was to be incompatible with the GPL and BSD licenses.

The biggest way Sun was a nicer neighbor was that they at least smiled while they put the binds on their users.