At the macro IT level, many CIOs continue to double down on hybrid, multi-cloud strategies with containers as the dominant foundation for cloud portability. Communications services providers (CommSPs) are aware of these trends for their own, internal IT operations and processes as well as the evolution of network services to support their customers.
CommSPs view cloud native deployment as a natural extension of foundational investments in network functions virtualization (NFV) and enhanced mobile broadband services are growing enterprise and consumer appetites for more advanced, data-centric capabilities. To meet these goals, a cloud native 5G user plane function is absolutely essential.
Ultimately, this architectural foundation will enable CommSPs to pivot from providing connectivity towards providing data-centric, digital services, leveraging their close proximity to their customers via the network edge. The shift is underway as AT&T calls 2020 an “epic year for the edge.” Let’s recap some recent cloud native milestones across the ecosystem and consider the keys to more mature and widespread deployments.
Disaggregation is the North Star
As part of its definition of the term, Cloud Native Computing Foundation (CNCF) identifies containers, service meshes, microservices, immutable infrastructure and declarative APIs as techniques that “empower organizations to build and run scalable applications in modern, dynamic environments such as public, private, and hybrid clouds.”
While containers are an important component to cloud native, the blood, sweat and tears shed in the pursuit of NFV lead to an immutable law of network transformation--disaggregation is our guiding principle or north star. Disaggregation of software and hardware moves us towards a multi-cloud, network environment that supports distributed workloads and open source software innovation.
Cloud Native in 2020: Performance, Scale, Ecosystem Readiness
We believe the three most important considerations for cloud native in 2020 are performance, scalability and ecosystem readiness to rapidly move towards commercialization.
With this in mind, we are working with a broad ecosystem on:
- Disaggregation, end-to-end virtualization and hardware abstraction to achieve cloud scalability.
- Performance improvements that include virtualized data plane performance for network applications in the cloud and new technologies such as AF_XDP. (We are also contributing to improvements in container networking and Kubernetes.)
- Cloud-scale, microservices deployments with robust orchestration solutions built for delivering Quality of Service (QoS).
Virtualized, Disaggregated Data Planes & Scaling Microservices
5G deployments are spawning innovative applications and services, such as video services, Artificial Intelligence (AI) and analytics, that require low latency, high bandwidth and greater disaggregation of the wireless infrastructure. The industry’s acceptance of Control and User Plane Separation (CUPS) set the stage for further disaggregation of 5G User Plane Functions (UPF) and closer proximity to the application, service and data in order to meet these performance and low latency requirements. Intel has made significant optimizations in our platforms for improving 5G UPF and virtualized data plane performance.
Looking ahead, our vision is to pave the path for enhanced packet throughput in the standard Linux infrastructure utilizing AF_XDP enhancements. This allows the network applications further abstraction from hardware within standard Linux infrastructure supporting a true Cloud Native deployment model. AF_XDP is an address family that is optimized for high performance packet processing making it possible for an XDP program to redirect packets to a memory buffer in Linux user space. Intel has worked with the community to drive significant improvements in AF_XDP infrastructure including an AF_XDP poll mode driver in DPDK, but our work has only just started. If you are passionate about the Cloud Native topic, please connect with us. Upcoming events such as the Open Network and Edge Summit (ONES) is a great opportunity to engage with us and learn more about Intel’s initiatives in this key area.
Finally, a key pain point for the industry today is the orchestration, management, security and resiliency of network applications in the virtualized infrastructure to achieve the efficiency of multi-cloud deployments. With an eye on rapidly evolving edge services, and keeping in mind the popularity of containerized applications, we are contributing to enhance both the Kubernetes and Kubernetes networking features. The recent release of Intel® Open Network Edge Service Software (OpenNESS) is one example contribution. This release provides a software toolkit that provides the easy button to securely onboard and deploy the microservices for IOT and enterprise applications at the on-premise and network edge leveraging multi-cloud deployment models. In the coming months, we will continue to drive enhancements in key areas to realize the full potential of 5G and Edge services.
Moving Toward Commercialization
We’re turning the page to a new decade of innovation and change in the network transformation journey. Companies across the industry are investing in open collaboration to break through many of the more difficult barriers to the application of cloud native architectures and concepts in network operations. A few examples include:
- “5G Core Operational Aspects: Running a Software-Defined Network” webinar series by HPE and Intel
- Ericsson, Telstra Container-Based EPC Milestone with Intel architecture
- White Paper: "Multi-Cloud Services on Kubernetes with Cloudify Orchestration and F5 Networks Functions”
- Cloudify webinar: “Running I/O Intensive Workloads on Kubernetes with Cloudify Orchestration”
- Affirmed Networks webinar: “How Cloud Native Core Network Helps Fulfill Demands of 5G Services”
As an industry, we continue to make progress towards commercialization. Stay tuned to hear more from us on cloud native transformation through the course of the year.
Notices & Disclaimers
Intel technologies may require enabled hardware, software or service activation.
No product or component can be absolutely secure.
Your costs and results may vary.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.