Open source played a significant role in software development over the past decade from containers to microservices, blockchain and serverless.
Chris Ferris, chief technology officer of Open Technology at IBM, discusses some of the open source trends from the past decade and what to expect in 2020 and beyond.
SEE: Tech Predictions For 2020: More must-read coverage (TechRepublic on Flipboard)
Smaller, faster containers and microservices
The concepts of containers and microservices were merely concepts before 2010, Ferris said. Then Docker launched in 2013, planting the early seeds of the container industry.
At the same time, microservices — and the technologies to make them possible — were borne in open source through the Netflix OSS project.
Docker went on to become one of the most influential technologies of the 2010s, giving rise to a myriad of new open source projects, including Kubernetes, which launched in 2015.
Today, he noted, Kubernetes is the largest open source project on the planet. Companies are using the platform to transform monolithic application architectures, embracing containerized microservices that are supported by service mesh capabilities of projects such as Istio.
“In the next decade, we anticipate that open source projects such as Istio, Kubernetes and OKD will focus on making containers and microservices smaller and faster to serve the needs of cloud-native development and to reduce the container’s attack surface,” Ferris said.
OKD is the open source version of Red Hat’s OpenShift platform. “Keep an eye on unikernels (executable images that contain system libraries, a language runtime, and necessary applications), which may also gain traction thanks to the open source communities around them.”
SEE: Deploying containers: Six critical concepts (TechRepublic)
Instantaneous serverless workloads
AWS Lambda was released in 2014 and put all the PaaS services on notice. Lambda’s release was followed by IBM OpenWhisk (which became Apache OpenWhisk), among others, in 2016. Both open source, distributed serverless platforms execute functions in response to events at any scale, Ferris said.
Kubernetes gained prominence in the latter part of the decade, fueling the desire to extend Kubernetes with capabilities that would enable serverless. This gave rise to Knative in 2018. Now Knative has split into multiple open source projects including Tekton, each with their own set of innovations, he said.
In the next few years, Ferris said we can expect to see containers get smaller, faster. “The potential exists to have an environment that can run containers at very little cost, instantaneously,” pushing the boundaries of serverless platforms, he said.
Trustworthy artificial intelligence
IBM Watson made a huge splash when it appeared on “Jeopardy!” in 2011, bringing artificial intelligence into the mainstream. Now, Ferris noted, AI is part of our everyday lives and we interact with Siri and Alexa daily, talk with customer service chatbots regularly, use facial recognition to unlock our gadgets, and are nearing the advent of fully autonomous self-driving cars.
AI and machine learning have powered these innovations and many of the AI advancements came about thanks to open source projects such as TensorFlow and PyTorch, which launched in 2015 and 2016, respectively.
In the next decade, Ferris stressed the importance of not just making AI smarter and more accessible, but also more trustworthy. This will ensure that AI systems make decisions in a fair manner, aren’t vulnerable to tampering, and can be explained, he said.
Open source is the key for building this trust into AI. Projects like the Adversarial Robustness 360 Toolkit, AI Fairness 360 Open Source Toolkit, and AI Explainability 360 Open Source Toolkit were created to ensure that trust is built into these systems from the beginning, he said.
Expect to see these projects and others from the Linux Foundation AI — such as the ONNX project — drive the significant innovation related to trusted AI in the future. The Linux Foundation AI provides a vendor-neutral interchange format for deep learning and machine learning.
New uses for blockchain’s tracking capabilities
In 2008, the pseudonymous Satoshi Nakamoto published his now famous paper on bitcoin, which introduced the concept of a blockchain network, whose purpose was to be a decentralized cryptocurrency platform.
That innovation made people start to wonder about different ways that the blockchain concepts and technology might be applied in non-cryptocurrency use cases — in asset management, supply chains, healthcare, and identity, among others, Ferris said.
In 2015, IBM contributed its Open Blockchain project to the newly established Hyperledger organization, founded to develop open source blockchain technology for the enterprise. That contribution launched what has arguably become one of the two or three most popular blockchain frameworks: Hyperledger Fabric, he said.
While blockchain’s initial uses were confined to cryptocurrency, open source engagement around Hyperledger and Ethereum has expanded the possibilities for how this technology is used.
In the enterprise, different approaches are being explored not only to enhance privacy but also to build a collection of nodes required to achieve confirmation on a transaction with trust – almost all in open source, he said.
Quantum processors available for developers
There has been lots of buzz around the promise of quantum computing, and although an app with a “quantum advantage” hasn’t been developed yet, the ability for developers to start using quantum processors is growing — and will continue to evolve in the next decade, Ferris said.
IBM’s open source Qiskit software framework, released in 2016, lets developers code in Python on real quantum hardware for systems around research, education, business, and even games.
“The possibilities for how quantum computing will solve problems and interact with today’s technology seem endless … quantum computing could impact a wide range of domains, such as chemistry, finance, artificial intelligence, and others,” he said.
For that to happen will require a “significant hardware environment,” Ferris said.
Open source in the coming years
Open source is the best mechanism to bring about these changes, he maintained. That is what spawned ideas like microsystems, which grew out of the virtualization space, and Knative from Kubernetes.
“That wouldn’t have happened in the closed source space, so it’s a matter of everyone building up on everyone else’s successes and someone coming along and saying, ‘Here’s a better idea,'” he said.
Working together, developers have the power to change entire industries, Ferris believes. “I can’t think of anything that’s been developed exclusively in closed source that didn’t eventually come out in open source.”