Regcast training : Hyper-V 3.0, VM high availability and disaster recovery

Thanks to projects like OpenStack and the mighty operation that is Amazon’s EC2, open source and Linux are quickly becoming the building blocks of “cloud” computing.

OpenStack, which started life in 2010, releases compute, storage, networking and other components under an Apache licence, and it is being adopted by huge companies such as telecom giant NTT in Japan and IT behemoth Hewlett-Packard in its fledgling cloud.

Open-source clouds - and we’re talking platform and infrastructure-as-a-service rather than hosted email or collaboration – currently have closed-source efforts such as Microsoft’s Windows Azure encircled and outnumbered.

Interestingly, HP went with OpenStack rather than Azure, despite having first signed on the dotted line with Azure in 2010, with the promises of appliances and services that - to our knowledge – still have not appeared to this day.

Now, with the middleware APIs and the Linux kernel well-established, the attention should now move to new areas – management frameworks, storage tools and software defined networking, according to the Linux Foundation.

Foundation executive director Jim Zemlin told The Reg these are the next obvious areas for change.

Typically such areas would be considered less sexy than, say, working on the OpenStack compute core or Linux kernel, and so wouldn’t attract too much interest.

However, Zemlin reckons that such projects will thrive now - partly because cloud is being deployed on a large scale and partly because engineers are being paid for by big computing companies.

“The difference now is they are not just obviously tinkering around with how to make a software defined network or block storage file format,” Zemlin said. “These are broad-scale industrial initiatives that are financed by the largest computer companies in the world to create the comments they need to make commercial products.

“Linux went from hobbyist project to commercial and industrial scale project over time; this time all the projects being used to power the cloud are starting out as industrial-scale projects.

“The computing industry at large has decided the days of hiring the best programmers, locking them in a room and not letting them come out until they write some brilliant code is over. The cloud‘s software in general is being written though collaboration – that’s the rule not the exception.”

Zemlin was speaking as his organisation held its European conference in Barcelona, Spain, and used the occasion to publish a joint IDC report and survey on the uptake and use of open source in the enterprise cloud.

The survey, taken in August and covering nearly 300 users, found 86 per cent said they would increase their amount of spending on open source in the cloud in the next year.

Collaboration and the importance of ecosystem received near-universal support – about 50 per cent called these “very important” factors for the open-source cloud while about 45 per cent said these were "somewhat important".

Asked to explain their view of “what is an open cloud”, respondents put the ability to port and access their data and option to run the cloud platform on premises rated as first and second in order of importance, followed by - yes - "community".

Open APIs also scored highly but it seems people are more flexible in terms of what they expect from their cloud service suppliers. Asked how important it was for a service provider to use open-source software, open standards or open APIs, and people appear more willing to compromise.

Twenty per cent said use of open-source software, standards and APIs would be important, around 50 per cent called it “a key factor, but not a requirement” while a surprisingly strong 20 per cent said this was “nice to have, but not a priority.”

It seems that at a certain point, people are concerned less with the principles of the software and services they’re using and more with whether the service works, and works well, at the right price. ®

Agentless Backup is Not a Myth