ar130405 - Fotolia

From open clouds to open infrastructure: OpenStack's evolution continues

OpenStack is embarking on a push to position itself as a lot more than just an open source cloud software provider, with forays into containers and continuous integration tools, as enterprises continue to demand their infrastructure do more

The range of use cases that OpenStack’s technology can be applied to has broadened significantly in recent years, beyond simply providing organisations with a means of standing-up their own open-source-based private and public cloud environments.

The output of the open source community that supports OpenStack has paved the way for the telecommunications industry, for example, to embrace the concept of Network Function Virtualisation (NFV), and provide it with a means of building edge computing environments.

Its contributors have also laid the groundwork for the Foundation to offer greater support for application developers with its forays into containers and continuous integration too.

In line with these developments, the Foundation has now sought to reposition itself in a way that fully encapsulates everything that OpenStack has to offer enterprises now – which is access to Open Infrastructure.

The Foundation has, by its own admission, previously struggled with how best to communicate to enterprises what exactly its technology does and how it stands to benefit them, but Open Infrastructure is succinct cover-all, said OpenStack Foundation chair Alan Clark.

Not only in terms of the technologies that come under the OpenStack umbrella, but also in the Foundation’s attitude towards working with adjacent open source communities, continues Clark.

“The Open Infrastructure tagline is also because we recognised the need to not just support OpenStack, but all those other technologies, and you also want to make sure your infrastructure is viable, not just for today but tomorrow as well,” he said.

Read more about OpenStack

  • The OpenStack Foundation opens up about the work it is doing to address the open source cloud platform's identity crisis and integration issues, as part of its continuing pursuit to improve the technology's enterprise-readiness.
  • The OpenStack Summit revealed details of how the seven-year-old open source platform is overcoming its growing pains and positioning itself for even greater enterprise adoption.

“We know there will be new technologies [emerging], so you have to make sure you have the infrastructure in place for these new ideas and technologies, and the vision going forward is that we’re the open infrastructure to be built on.”

During the opening keynote at the OpenStack Summit in Vancouver, the Foundation’s chief operating officer, Mark Collier, said the Open Infrastructure concept is also reflective of the growing pressure IT operators are under to build software stacks that meeting a wide variety of use cases.

“One of the most interesting developments in infrastructure and cloud in general [at the moment] is that our operators are being asked to do more for their businesses and end users,” said Collier

“People expect their infrastructure to handle artificial intelligence, machine learning, [and] containers are really a given these days at various levels of the stack because of how powerful they can be, and people are starting to experiment with serverless.

“This is the world the operators live in right now – more pressure on cost and compliance, and more pressure to deliver additional functionality in their clouds, and on top of the functionality piece they are also being asked to do it in more places.”

A new era of openness for OpenStack

While the Vancouver Summit essentially marks the start of the “Open Infrastructure” era at OpenStack, the Foundation has been laying the groundwork for its repositioning at its previous meetups, with Clark describing the Boston conference in May 2017 as being a pivotal moment.

It was here, Clark explains, a number of key decisions about the future direction of OpenStack were hammered out, including how to forge closer, collaborative ties with other open source initiatives, while clearing up the confusion about what OpenStack is all about.

A major contributor to this confusion was the introduction of the Big Tent governance model in 2015, and the resulting overhaul in how OpenStack project are defined.

Whereas contributors previously had to petition to have their projects included in an integrated OpenStack release before they could start work on them, under the Big Tent approach, they were given the green light to start working on them provided they adhered to certain OpenStack community guidelines.

“We had people really confused, and one of the things that came out of the strategy session [in Boston] is that we still had people asking what is OpenStack?” he said.

“We’d introduced this notion of Big Tent and it caused confusion to users about what was really OpenStack, but we still needed a mechanism to enable innovation and new ideas.”

Open integration push

The Boston summit laid the groundwork for the Foundation to announce a multi-year commitment, at its Sydney Summit in November 2017, to addressing the integration challenges enterprises commonly come up against when trying to build heterogeneous, open source-based infrastructure stacks.

Several months later, in February 2018, a whitepaper followed that saw the Foundation make a case for the creation of a cross-industry coalition to address the stumbling blocks that may serve to hinder the adoption of edge computing in the years to come.

“We’re seeing the fruits of [those initiatives] all delivering dramatically,” said Clark.

This has seen the Foundation develop closer working relationships with open source platform-as-a-service Cloud Foundry, and with the Cloud Native Computing Foundation (CNCF), who look after the container orchestration engine, Kubernetes.

“Each community is a little different, so how we [forge ties with them] is very different. Some don’t need a lot of interaction – they just need to know what our interfaces are like. Others have been much more directed by us,” he said

“A good example of that is with the Kubernetes community. We have a special interest group that is focused on Kuberetes integration, who have come up with code to improve the integration [with OpenStack].”

Commitment to Open Infrastructure

Another show of the Foundation’s commitment to the Open Infrastructure cause can be seen in its decisions to spin out a couple of projects that started life within OpenStack to ensure they reach as wide an audience as possible, added Clark.

These include the open source continuous integration tool, Zuul, which allows OpenStack users to automate large parts of their software development cycles, and is now managed as an independent project by the Foundation.

The Foundation also used the Vancouver to debut the first release of its hardware agnostic container management software, Kata Containers, which boasts compatibility with similar offerings from the Open Container Initiative and Kubernetes.

The latter is designed to address user concerns around container security, continued Clark, but both offerings should be viewed as OpenStack practicing what it preaches about the importance of ensuring open source communities from adjacent communities play nicely together.

Clearing up the cloud confusion

One of the biggest criticisms levelled at OpenStack during the Big Tent era is that it made it difficult for users to differentiate between the core and periphinery pieces of its stack, and – in turn – what parts were absolutely critical to standing up private clouds in their datacentres.

As alluded to by Clark, the Foundation has made a concerted effort over the last couple of years to bring some clarity to the situation by culling under-performing projects. But Mark Shuttleworth, co-founder of Ubuntu OpenStack distribution maker Canonical, claims there is scope to take these efforts even further.

“In the past I’ve been critical of the Foundation for not being clear enough about what you needed to stand up an Openstack cloud,” he told Computer Weekly at the Summit.

“I would still like them to say these seven pieces of code are OpenStack, and if you have those seven pieces of code, that do a great job of running a cloud, you’re good, and I think it would be in their best interests to.”

Given his support of the Foundation’s past efforts to streamline the number of projects running under OpenStack, what does he make of the Open Infrastructure concept, and its implied messaging that there is much more to what does than pure cloud.

“They started that process of simplifying the definition of OpenStack, but then they said we’re not just OpenStack anymore. Can they manage that dance? Let’s give them the benefit of the doubt for now,” he said.

“Some of the new things they’ve embraced aren’t in their cloud of stuff around OpenStack, like Kata Containers and Zuul - they’re really different, so maybe there is some argument for saying there are other classes of infrastructure.”

Read more on Managing servers and operating systems

Data Center
Data Management