The OpenStack Foundation has launched another independent open source project — the StarlingX project that will focus on edge computing.
StarlingX is not an OpenStack sub-project. It’s a full project on its own. At the same time, it uses a number of OpenStack services to provide the core compute, storage, and networking functionality. StarlingX is based on those core OpenStack services, but then it goes well beyond those to add other kinds of services for delivering cloud at the edge.
Jonathan Bryce, executive director of the OpenStack Foundation, said that about a year ago the group realized its community had expanded beyond just OpenStack deployments in enterprise data centers. “We needed to be able to support growing new technologies like edge, NFV, containers, and machine learning with the existing community we already had,” said Bryce.
To meet this need, in December 2017, the OpenStack Foundation created its first ancillary open source project: Kata Containers, which aims to unite the security advantages of virtual machines (VMs) with the speed and manageability of containers. Like the StarlingX project, the Kata Containers project will be managed by the OpenStack Foundation, but it is an independent project with its own technical governance and contributor base.
Why Another Edge Group?
There are already a bunch of projects working on edge computing. Ildiko Vansca, an ecosystem technical lead with the OpenStack Foundation said she counts about 25 edge projects, including Akraino Edge Stack, Open Edge, Deutsche Telecom’s MobiledgeX, and the Central Office Re-Architected as a Data Center (CORD) projects, which are focusing on the edge, too. There’s also ETSI MEC, which is working on standardization.
So why in the world do we need another edge project? Vansca said, “Edge computing is relevant on all the stack from the hardware to the applications. The point is to make sure we focus on each of the layers, identifying use cases, architectural options, and filling all the gaps in all these layers. Under the OpenStack umbrella, we are still on the Infrastructure-as-a-Service layer. We are not going to the application layer or the hardware layer.”
Bryce said that the scope of edge computing is extremely large. “When people talk about edge, they kind of present this picture that it’s ‘a thing,’ he said. “But really, it’s an entirely new delivery model and infrastructure space as well. It’s really like the web or mobile. It’s that big in terms of the scope of it. StarlingX is really filling one of the spaces in that stack.”
StarlingX is available in its first release today. The project builds on code contributed by Wind River and Intel.
Glenn Seiler, VP of marketing and product management at Wind River, said StarlingX originates from mature, field-proven code because it originated from Wind River’s Titanium Cloud, which is in its fifth generation and deployed in many scenarios. “We actually started Titanium Cloud back in 2013 specifically for NFV,” said Seiler. “We found that it was applicable to much more than NFV.”
The initial StarlingX release is composed five infrastructure services:
- Configuration management
- Service management
- Fault management
- Host management
- Software management
“StarlingX has put into open source these services and builds those into a complete edge stack using other open source services,” said Seiler. “It’s built into an actually-deployable infrastructure stack.”
China Unicom has already run a full validation on StarlingX over the past few months. “It is one of the top strategies of China Unicom to build an open edge platform,” said Dan Chen, senior director of edge computing at the service provider. “StarlingX will play an essential role in China Unicom’s edge strategy.”