The City of Lights hosted the largest in-person KubeCon ever, with over 12,000 attendees gathering in Paris for KubeCon + CloudNativeCon Europe 2024. This record-breaking gathering celebrated Kubernetes’ 10th anniversary and spotlighted the maturation of cloud-native technologies.
Compared to the KubeCon North America 2023 event in Chicago, the Paris event, organized by the Cloud Native Computing Foundation (CNCF, part of the Linux Foundation), showcased a more mature and diverse cloud native landscape, with a markedly greater emphasis on AI/GenAI, sustainability, observability, platform engineering and emerging technologies like WebAssembly (WASM).
A decade of Kubernetes: catalyst for cloud-native innovation
The tenth anniversary of Kubernetes (K8S) served as a poignant reminder of the project’s monumental impact on cloud computing.
From its inception to powering the world’s most influential brands, Kubernetes has become synonymous with cloud-native, enabling advancements in AI, security and multi-cloud environments. This celebration underscored Kubernetes’ role as a foundational platform for years to come.
AI and cloud-native: synergy and caution
One of the most prominent themes was the intersection of artificial intelligence (AI) and cloud-native technologies. Many sessions and keynotes focused on how K8S and cloud-native ecosystems are becoming the backbone for modern AI workloads. Companies like NVIDIA, Google and Bloomberg shared their experiences in optimizing K8S for AI/ML applications, leveraging techniques such as GPU sharing and dynamic resource allocation (DRA), and projects like Ray and KubeFlow.
CNCF was eager to showcase the close relationship between cloud-native and GenAI, bringing GenAI leaders like Paige Bailey, lead product manager (Generative and Models), from Google DeepMind; Timothée Lacroix, cofounder at Mistral AI; and Jeffrey Morgan, founder of Ollama out on both the keynote and breakout stages.
While many speakers emphasized the affinity between cloud-native and AI, I align with the viewpoint of Arun Gupta from Intel. During a media and analyst panel, Arun cautioned that we should continue to ask ourselves whether K8S and other cloud-native projects are the optimal tools for building large ML training or inference clusters.
K8S might be the correct answer today, but the answer could be different tomorrow.
Furthermore, there were discussions on the emergence of a new AI Engineer role (like the evolution of DevOps and DevSecOps) and security and governance. In particular, concerning AI risk, large language models (LLMs) are expected to boost developer productivity.
Still, it’s crucial to establish standards and guardrails to ensure the safe and responsible development of AI-powered applications. I don’t know if LF and CNCF are appropriate homes for establishing responsible AI practices and standards. Nevertheless, the CNCF can and should encourage platforms, technologies and projects that can facilitate responsible AI and enable the effective placement of guardrails within AI/ML systems.
The rise of WebAssembly and edge computing
WebAssembly (WASM) was given more prominence on center stage than in Chicago. WASM was spotlighted as a critical enabler of lightweight, portable and secure cloud-native applications, especially for edge and on-premises workloads. The release of SpinKube (supporting Spin on K8S), the ongoing maturation of WasmCloud and the release of WebAssembly System Interface (WASI) 0.2 are all indicators of WASM’s growing momentum.
On center stage, ZEISS Group, using Fermyon’s Spin solution, highlighted how WASM reduced resource needs, cut costs by 60%, and achieved a 43% performance gain in their enterprise workloads. Meanwhile, telecom operator Orange discussed their use of WASM as a highly-portable run-anywhere solution (including at the edge) using wasmCloud (of which Cosmonic, a startup, is a significant contributor).
Orange and other telcos, Vodafone, Etisalat, and NBN Co will showcase their efforts at TM Forum’s DTW Ignite in Copenhagen in June.
Sustainability takes center stage
Another key focus area was environmental sustainability and the role of cloud-native technologies in reducing computing’s carbon footprint. The emphasis on sustainability reflects the growing awareness and commitment within the cloud-native community to address technology’s environmental impact.
Deutsche Bahn’s keynote on “Building IT Green” showcased how they are leveraging K8S and tools like Kepler to monitor and optimize energy consumption. The event also featured discussions on code audits, intelligent autoscaling and other techniques to improve cloud-native deployments’ efficiency.
eBPF: the powerhouse under the hood
eBPF, a technology allowing safe and efficient code injection into the Linux kernel, was omnipresent at the conference. It, along with the Cilium, a project that leverages eBPF to provide networking, security and observability capabilities for cloud-native environments, had gained prominence prior, even in Chicago. Cisco’s recent acquisition of Isovalent, a significant contributor to the Cilium networking project, drove even more interest.
The widespread adoption of eBPF (including by vendors like Tigera that supports an alternative to Cilium named Calico) signals its critical role in enhancing the performance and functionality of cloud-native infrastructure.
Security: baked in, not sprinkled on
Security continues to be a top priority in the cloud-native world. I expect every KubeCon to continue emphasizing the importance of software supply chain security (with SBOM), zero-trust models, and embedding security early in the development lifecycle.
CNCF projects like in-toto, TUF, and sigstore were highlighted across multiple presentations at the conference, hoping to drive awareness with developers and DevOps to build security in as early as possible in the process and to promulgate security best practices.
Multi-cloud, multi-cluster, multi-tenant: growing pains
The increasing complexity of cloud-native deployments necessitates managing applications across multiple clusters and clouds. Sessions and project updates at KubeCon focused on tools and best practices for multi-cluster and multi-cloud Kubernetes management. This trend highlights the need for abstraction layers and solutions that simplify the management of these complex environments.
Platform engineering takes flight
The rise of platform engineering was evident in the dedicated Platform Engineering Day, co-located BackstageCon (focused on the Backstage open-source project that makes building developer portals easy), and numerous sessions focused on building internal developer platforms.
Organizations increasingly leverage cloud-native technologies to create tailored platforms that empower developers and streamline development processes. A recurring theme throughout the event was the importance of empowering developers and providing a convenient, self-service, scalable platform for application developers to efficiently build, debug, deploy and scale their enterprise applications — numerous panelists issued exhortations to platform engineers to “empathize” with developers and treat them as customers.
Looking ahead: the future of cloud native
KubeCon + CloudNativeCon Europe 2024 served as a catalyst for reflection and forward-looking discussions. The cloud-native community feels vibrant and innovative, poised to tackle emerging challenges while capitalizing on the opportunities presented by AI, edge computing, and sustainability.
As Kubernetes enters its next decade, the focus will shift towards optimizing developer experiences, enhancing security and observability, and embracing sustainable computing practices. For an analyst with one foot in the cloud world and another in the telco world, it’s refreshing to be at a conference where the focus is on “do” versus “say” and where there’s more action than talk — we have much to learn from the cloud natives.
Roy Chua is the founder and principal at analyst firm AvidThink.
Industry Voices are opinion columns written by outside contributors — often industry experts or analysts — who are invited to the conversation by Silverlinings' editors. They do not represent the opinions of Silverlinings.