Hundreds of companies and more than 10,000 individuals have been part of the OpenStack project, which in its three years of existence has become a widely used software platform for building open-source cloud solutions. But as larger companies are looking for cloud solutions that will transform their businesses, OpenStack is at a crucial stage where it has to convince IT departments that it’s up to the task.
The solution, it seems, is drawing upon the collaborative and open nature of the OpenStack community. By choosing to work together to ensure that different vendor technologies are interoperable, clients can choose OpenStack components based on merit rather than being locked-in to a single vendor. And collaboration is essential in making solutions from different vendors work together seamlessly as a reliable and robust solution.
Ubuntu Server product manager Mark Baker says large customers such as Deutsche Telekom had been asking for assurance that different cloud components would work together and that they could be provided with reliable support.
In an effort to be able to provide this reassurance, Ubuntu started its OpenStack Interoperability Lab, which is essentially an expansion of its existing continuous integration testing lab. The crucial development with OIL is that Ubuntu is now able to collaborate with around a dozen other vendors from across the stack such as Dell, EMC, HP, IBM, LSI, Intel, and VMware.
“There’s a physical lab based in Boston,” Baker says. “We have hardware and software vendors send us their technology.” He says sometimes partners have to install their equipment, but they’re also provided with a technical partner manager who feeds them the results of testing and when problems arise, that manager directs work to whoever is best equipped to solve the issue.
“When you’re looking at interoperability testing it’s something that goes really deep… So, when you encounter problems, you have an entity that is able to triage those problems or work out the source and understand how to fix it,” Baker says.
This wouldn’t be possible without collaborating with the companies behind these technologies.
While the interoperability lab has been generating results since spring 2013, there hasn’t yet been a meeting between all members, although Ubuntu “definitely wants to go down that route” whether it’s in-person or online, Baker says.
Ubuntu is a good meeting place for many companies in the OpenStack ecosystem, he says, partly because it’s estimated that around 60 to 80 percent of OpenStack clouds use Ubuntu, and a vast majority of companies in the space are friends rather than competitors. And while other Linux operating systems haven’t joined OIL, he says they are “certainly welcome”.
In fact, making sure that different operating systems can coexist in OpenStack clouds is important for many companies running a multitude of different operating systems.
OIL is far from the only mixture of technology vendors working together to improve OpenStack. IBM and Mirantis, which are individually among the top five contributors to the latest OpenStack release, joined forces in December to enhance Mirantis’ OpenStack testing and benchmarking tool called “Rally” by allowing it to simulate traditional enterprise workloads on as many as 1,500 IBM servers.
Mirantis co-founder and EVP Boris Renski says being able to test OpenStack clouds before they’re actually deployed is essential in getting more cautious businesses to understand how an OpenStack environment would run their particular application workloads.
“The more traditional enterprises need guarantees in regards to how a particular configuration of OpenStack would function with respect to…various application workloads,” Renski says.
According to him, OpenStack’s quick growth has largely focused on adding features and functionality. Things likes tools for validating that configurations work against SLAs haven’t been a major focus, but as OpenStack matures they need to be addressed in order to gain the trust of larger enterprises that already have processes and procedures in place.
“It’s still a very new platform. It’s at a point now where enterprises can definitely derive value, but very traditional enterprises have a set of processes and policies with respect to how they deal with new technologies.
“We’ve come across this several times, in large organizations,” Renski says. “For anything that is production-bound, they have a specific list of certified technologies that they are allowed to use, and anything that is not on that list you cannot use, and to get on that list, the technology has to be at least four years old. If it’s less than four years old it’s not even considered.”
Renski says that many of these companies’ processes don’t keep up with the pace of technology, and this rigidity excludes organizations from being able to take advantage of new technologies. And while some enterprises cannot budge from their established procedures, instilling confidence in OpenStack solutions through extensive testing is one way to quell an enterprises’ anxiety over new technologies.
In December, Mirantis was able to launch a preliminary version of Rally, which was able to codify the workloads of companies such as GAP, Intel, PayPal, and Expedia. These use cases were run on different configurations of OpenStack on top of IBM’s 1,500 bare metal nodes to explore how exactly OpenStack performs in real-life environments and where, if at all, are there any bottlenecks that need to be addressed.
Even fierce competitors and vendors accused of locking-in customers seem to be coming around to the idea of cooperating on making the OpenStack platform enterprise ready – something that’s virtually impossible to do on their own. A truly open and reliable platform for cloud solutions would be a benefit for customers, and undoubtedly open up new opportunities for vendors who are willing to work together.