News & Insights for the AI Journey|Sunday, October 13, 2019
  • Subscribe to EnterpriseAI Weekly Updates:  Subscribe by email

Debate: How Many Open Source Platforms Are Enough? 

Representatives of competing open source cloud controllers debated the merits of each and whether multiple open cloud platforms are even needed.

The platform debate often comes down to which application-programming interface (API) is supported along with overall platform stability and ease of use.

One executive speaking during the recent Gigaom Structure conference called the issue of multiple open platforms a false question. "Nobody asks Microsoft or VMware, 'Are there too many proprietary cloud solutions out there?'" argued Sameer Dholakia, general manager at Citrix Systems representing the CloudStack platform. "This is the first time in the cloud era -- in the history of software – where open source is leading. There's not going to be one [open source platform], there will be multiple."

Added Dholakia, "The balance of power in software is going to shift to the open source-based platforms and there will be multiple [platforms] in the same way there have been multiple proprietary ones."

Marten Mickos, CEO of Eucalyptus Systems, replied, "In the future, we won't talk about hybrid clouds. We'll just say, 'I have workloads, I need to run them on different deployments'" in a public or hosted cloud. The bottom line, Mickos continued, is "it must give me the same API."

Mickos likened the situation to open-source Linux providing the same operating system on anything from an Android device to a server.

"Every major IT vendor is supporting OpenStack because it'll give you that interoperability" along with the ability to use a public or hybrid cloud deployment model, countered Nebula founder and OpenStack proponent Chris Kemp, who was at NASA when the space agency teamed up with Rackspace Hosting a few years back to start the OpenStack cloud controller project.

"That would be true if OpenStack had a presence in public cloud, which it doesn't," Mickos shot back, while later conceding that cloud providers like HP and Rackspace represent a small percentage of compute cycles on the public cloud.

"When it gets there, we will support the OpenStack API," Mickos relented.

Dholakia noted that CloudStack, like Eucalyptus and OpenStack, has long maintained a "compatibility layer for the Amazon API precisely because, as business folk, we follow the dollars."

OpenStack came in for the most criticism for issues such cost of deployment and maintenance. "OpenStack is not a product," Kemp responded. "If you are building a large infrastructure, it's more like a tool kit. It gives you a lot of technologies that do take a lot of effort to integrate." The tool kit is used to create a product, Kemp stressed.

An operator like a large bank, for example, would prefer to tailor its deployment and integrate it into their existing infrastructure rather than deploying a turnkey platform, Kemp argued. "That flexibility is a double-edged sword."

CloudStack backer Dholakia acknowledged the distinction among open platforms as either products or "projects." However, he noted that the debate about which open platform or how many is overshadowed by the need to move more enterprises to the cloud.

"How do we get more people doing this, grow the market, grow the pie?" He added: "The number one barrier is their ability to adopt." The industry has so far largely failed to deliver a platform that is stable and easier to use. "We've got to get to maturity, scalability, stability, resiliency," Dholakia said.

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).

One Response to Debate: How Many Open Source Platforms Are Enough?

  1. Tim Wessels

    Well, the better question is how many “cloudstacks” will the market support. For the sake of the argument let’s concede that VMware and Microsoft have commercially viable cloudstacks. Eucalyptus is a niche player but Mr. Mickos is betting heavily that hybrid clouds that are API compatible with AWS will be successful…eventually. Apache CloudStack, which is branded by Citrix as CloudPlatform seems to do well in the public provider space…SoftLayer was built partially on CloudStack. CloudStack may lack the buzz that surrounds OpenStack but getting POs is always better than getting PR. OpenStack has a group of commercial distributions in both public and private cloud environments. HP, RackSpace, IBM and Internap on the public side and Nebula, Piston, Red Hat and Mirantis on the private side. The question is how many of these OpenStack distributions will achieve commercial success? Can the market for public and private OpenStack clouds support a dozen or more OpenStack distributions? Probably not and it is unclear which ones will achieve commercial success and which ones won’t make the cut. It could take another 18-24 months before that picture gets clear(er).

     

Add a Comment

Do NOT follow this link or you will be banned from the site!
Share This