Open source software has been a key underpinning of enterprise IT for years, so it’s no surprise that it’s helping to drive the infrastructure part of the equation forward just as much as application development.
Some projects are much more influential than others, and here are five that are doing the most to help enterprise infrastructure keep pace with the demands of an ever-more sophisticated operating environment.
OpenStack is notable in part for being an open source competitor to the most important proprietary virtualisation software on the market—VMware’s VSphere. For the basic task of virtualising servers into a flexible pool of computing resources, the difference appears to be ease of use—it’s simpler to use VMware when there isn’t a lot of in-house virtualisation or private-cloud expertise.
OpenStack is important in networking for one main reason: the telecom sector and network function virtualisation (NFV), which uses enterprise virtualisation technology to perform networking tasks previously allocated to dedicated hardware tied to proprietary software.
Telecom providers love this idea because it lets them replace expensive, proprietary products with general-purpose switches and servers. Also, software used for NFV, like OpenStack, lets them dynamically provision workloads and deploy new capabilities more flexibly.
Arpit Joshipura, general manager of networking and orchestration at the Linux Foundation, said that OpenStack and other NFV-enabling projects have quickly become central to telecom operations.
“Telecom was all proprietary, everything from the [radio access network] to the edge to the core,” he said. “In the last five years, telecom networks have become totally open source dependent.”
Originally developed by Red Hat, Ansible is an open source IT-automation and configuration-management tool that offers an alternative to configuring hardware manually. The idea is that the IT team writes a script that describes the network and what it’s supposed to do and then Ansible automatically configures the relevant devices. It doesn’t use agent software, instead pushing “Ansible modules” directly to the devices via SSH for easy deployment.
“Ansible’s important because you need to be able to orchestrate your machines once you have a lot of them,” said Joseph. “You can manage a server or two or 10 on your own, but it’s much easier to deploy them and manage them automatically.”
Red Hat offers an array of paid add-ons for Ansible, as well, including improved security, role-based access control, and job scheduling. Ansible offers a method of network configuration that lets IT workers set configurations once on a single controller and automatically pushes them to devices on their networks.
Software can also be pushed to all devices on the network or just a relevant subset by editing to the main playbook. Changes can be tracked and identified in Git or some other versioning system.
Akraino launched in 2019 and is the product of the Linux Foundation’s LF Edge program, which is aimed at creating open frameworks for edge computing deployments. Akraino is a collection of configuration blueprints designed to offer a freely available, off-the-shelf recipe for network and hardware configurations for specific use cases.
Akraino currently includes 11 blueprint families grouped by general use area and 27 specific blueprints. One example is StarlingX Far Edge Distributed Cloud, which specifies a hardware setup, containerisation providers, and an orchestration framework to enable applications to run in high-density locations like airports, sports stadiums, and malls. Other blueprints focus on AR/VR infrastructure, telecom radio deployment, and various types of IoT.
The idea behind StarlingX is to offer vendors and sophisticated end-users a way to streamline the configuration of the common elements of edge deployments. A company with a new application for a particular vertical market—say, providing real-time monitoring to connected factories—can focus on that without having to design the underlying computing infrastructure.
Kubernetes is a containerisation platform for all kinds of enterprise workloads, originally the product of Google engineers, but released as open source in 2014. It’s since become an industry standard, accounting for 71% of containerisation use in the enterprise, according to a study from 451 Research.
Enterprises like containerisation in general, and Kubernetes in particular, because it’s an effective simplification vs. monolithic models of service deployment. Instead of a single application offering a range of services and requiring a specialised infrastructure, Kubernetes breaks each process used by the application into its own container and virtualises it.
What that means is that the containerised workloads can run wherever—on premises, public cloud, private cloud, or various combinations thereof simultaneously—and work just as they would if they were bundled in a single application running on dedicated hardware. Consequently, developers can create a file that outlines how the services are supposed to work, and Kubernetes automates everything from provisioning to failover to updates.
Kubernetes was released as open source with the goal of simplifying the underlying infrastructure while leaving vendors and users the option of creating modifications to address a particular market or a particular enterprise need, according to Elizabeth K. Joseph, a developer advocate and open source expert at IBM.
“A lot of huge companies got together to build these things that give a skeleton or core of how to function,” she said. “A small company can run it themselves, but that’s actually kind of hard. The reason these companies invested in [these projects] is because they know they can sell things on top of that skeleton, letting them skip having to write the basic, boring stuff that it’ll have to do anyway.”
Any listing of open source projects that are important to the enterprise networking must include the Linux kernel. Linux fundamentally underpins huge amounts of modern enterprise networks, including all of the other projects listed here. By extension that means it’s also the basic operating system behind 90 per cent of the public cloud, according to a 2019 survey by Red Hat.
Even by itself, the operating system includes robust networking features that make it easy to deploy on white-box hardware. As the tasks of deploying and managing networks become more and more software-based, Linux skills are increasingly critical to just about every network IT professional out there. “I think people take it for granted,” said Joseph.