Open vs. Closed in the Cloud

The impending announcement of Microsoft's cloud operating system at the company's Professional Developers Conference has me thinking about how the struggle between open source and proprietary software models will play out in the cloud. There's been much chatter about how the relocation of code from one's premises to the virtual skies might threaten, render irrelevant or somehow derail the growth of open-source software by upsetting its natural licensing boundaries and advantages. However, the growth of closed source cloud alternatives will no more push open-source software out of the cloud than open source has pushed proprietary software out of its ancestral on-premises homes...

The impending announcement of Microsoft's cloud operating system at the company's Professional Developers Conference has me thinking about how the struggle between open source and proprietary software models will play out in the cloud.

There's been much chatter about how the relocation of code from one's premises to the virtual skies might threaten, render irrelevant or somehow derail the growth of open-source software by upsetting its natural licensing boundaries and advantages.

First, while most cloud offerings are currently built out of open-source pieces--chief among them the Xen hypervisor and Linux operating system--the fact that the open-source licenses that govern these and most other free components do not require cloud services providers to share their contributions could pose a challenge to open source from within.

However, as projects such as Linux and Xen move forward and accrue improvements, the cost to proprietary extension-minded cloud providers of either forgoing community-driven improvements or of taking on the ever-larger integration load of synchronizing their changes with the mainline will work to keep most cloud implementations open.

More interesting to consider is the way that licensing issues around proprietary software, especially those related to redistribution and usage metering, will fade in importance for end users once the software and hardware resources that host it come to be bundled into common utility services.

For instance, by bundling its software with cloud-based hosting, Microsoft can offer its customers a much simpler licensing picture, in which the tricky pricing-by-projected-usage models now enforced awkwardly by Client Access Licenses and the like can shift to pricing based on actual usage.

I imagine that Microsoft's proprietary cloud offerings will achieve their share of success, and they will play at every level of the cloud computing stack, from slices of space in the company's new data centers, to Hyper-V-powered virtualization, to a Windows Server-based platform, to higher-level stack layers such as SQL Server Data Services.

However, these achievements will no more push open-source software out of the cloud than open source has pushed proprietary software out of its ancestral on-premises homes. In fact, the realities that prompted Amazon to build its cloud offerings out of Xen and Linux rather than, say, VMware and Windows, are as relevant today as they ever were.

Namely, open-source software is available for the owning to whomever wishes to take it up, and it grows in value, rather than declines, with every new interested owner it accrues. For this reason, open source will continue to be a first option for building new platforms and services, whether they be cloud-based or not.