Marten Mickos Says the Cloud Won't Kill Open Source

here is a fervent debate going on in the open source community about cloud computing. Will the cloud kill open source? [2] Will multi-tenancy services make cloud providers need to hide the source, as a recent Forbes article [3] suggests?

I recently went to the best expert I could find on the matter to ask: Marten Mickos [4]. Mickos is the CEO of "private cloud" software (aka virtualization) maker Eucalyptus Systems [5]. He earned himself Open Source Hall of Fame status as the charismatic former CEO of MySQL. Mickos was adamant that open source not only won't kill the cloud, but that it must be used to power it. He says he has staked his career on that belief.

"Cloud computing is the biggest shift in IT that we will ever see. It will permeate everything," he says. "We see it already fueling the mobile Internet. It affects how hardware is designed and it is disrupting some old players. I wanted to join it --I like to be involved in a big shift." Mickos (pictured right) says he approached Eucalyptus and asked to be a part of the company. He didn't want to move into cloud computing while abandoning his open source roots.

Eucalyptus is among the most mature vendors to straddle between the open source and cloud worlds. Eucalyptus was originally developed [6] as an open source project as part of an NSF-funded academic research project at the University of California, Santa Barbara. It is built on Linux. The company is known for its partnership with Canonical and inclusion in Ubuntu's server edition [7]. (Aside: Ubuntu also supports KVM and libvirt for virtualization.)

But let me backtrack. After all the progress that open source has made to win the minds of business users, cloud services can wipe that right out. The cloud delivers software as a service, making the source code irrelevant for the end user and, maybe, for the IT professional, too.

For instance, Richard Parrish, chief technology officer for Wood County, Texas, has only begun to dabble with open source. He has recently deployed Astaro [8], a Linux-based cloud app that archives Exchange e-mail. Astaro is based on open source code, namely Linux, a fact which figures prominently in Astaro's marketing materials but matters not at all to Parrish. While he cares about customization -- for instance, he wants the ability to add personal messages to screens that users see -- he's less concerned with the innards of the service itself.

"I haven’t done anything with the source code -- as far as I know I can’t access the source code [so] I can’t modify the source code," Parrish explains. " I login to an interface. I think when they say it is open source they mean it is Linux based … but it doesn't really matter to me."

Cloud computing services differ from the "private clouds" that Eucalyptus delivers. Private clouds refer to a virtualized infrastructure hosted on-premises. In the case of the private cloud, users have, and care more about, access to the source code. But Mickos says even with software delivered fully as a service like the type Parrish uses, many users still care if the underlying code is open or not -- as a matter of principle, if not practicality.

"The users do care. There will always be people who don't care, but ... awareness about open source is greater than it was in the 1980's and 1990's. We have gotten used to everything being open, and we are demanding that it be so even in adjacent areas," he says.

His vision for Eucalyptus is that clouds must be big and broad enough to support the whole planet. Users worldwide will eventually be networked, billions of Internet devices, all of it powered by software, he says. "It will have to follow a global standard and support many sorts of hardware. It will have to provide a unified interface."

Mickos naturally believes that Eucalyptus has the software to do the job. But his bigger point is that, as public cloud services grow to envelope the world, so, too, will "private clouds" grow. And for enterprises, the lines will blur. Most will use "hybrid" clouds with some infrastructure for each application placed on premises and some delivered as a service, he says.

While that certainly aligns with Microsoft's software-plus-services vision [9], in Mickos' future all of it is open source. Users will care because they will have access to their piece of the cloud and will need it to function with their service provider's piece. They may also simply care that their service provider isn't just taking from the open source world and building their engine on it, but is giving back. (Throw in obligatory punch at Apple [10] here.)

All that said, Mickos doesn't believe that every piece of code written by a software vendor should necessarily be open. For my next post, I'll write about his defense of the controversial open core [11] concept.

For those who are “resource intensive” application users, I think the Cloud concept will not be fast enough. I am speaking from experience using database publishing software. (also CAD software)

For those who have “security concerns”, I am not sure we should trust the Cloud. For that reason I think I will be using a Desktop PC for a long time to come. (and I built my first PC in 1982)

I am also a strong proponent of Open Source software, have been a Linux user for over 10 years. There are only a few applications in the Windows environment that I cannot find a corresponding Linux app for.

Of course it won’t. I’m working on a project along these lines and we’re planning to release the code as open source. However, as they said in the Forbes article, we’re not making the ‘multi-tenant’ (<-—dumb buzzphrase...I mean, weren’t we all ‘multi-tenant’ on our Geocities accounts in the 90’s??? ) version of the application, which is quite different in many ways, open source. That’s something that is going to be making us money while we give the open source code away for free.

Wordpress is open source. I’m sure the code they run on wordpress.com is quite different from their open source offering and I don’t believe that’s open source. (I’d hope so, anyway...)

The mere existence of closed source software even from open source companies does not ‘kill’ open source. Someone is free to take the open application that we wrote and adapt it to a SaaS model if they so choose. And if they want to make that code open source, good on them. All they need is the $50k of equipment we’re using to run it on once they’re done with that and they’ll have a business. In the meantime, anyone can use the open source version on their $10/yr ‘multi-tenant’ shared hosting account.

They’ll have to pry my Linux desktop from my cold dead hands. I don’t like using ‘the cloud’ for applications myself, though I do use Evernote and RTM. But as a substitute for local apps? Not any time soon for me.

However, the things that are going on with virtualization and ‘cloud servers’ in the hosting arena are really exciting even if the buzz words are annoying.

For the average user the Cloud will appear just like local apps, but more friendly.

For a “power user” it will not.

For a person concerned with data security it will not.

I am with you on prying my local PC from my cold dead hands.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.