I am very excited to be able to introduce our Intelligent, Flexible Cloud vision to the community today. This is a vision the team here at ARM has been working on for a number of months, and really is an evolution of the ARM ecosystem’s vision and opportunity for networking infrastructure.
The core of this vision is that the network will become a platform - a platform for traditional network functions, as well as a platform for application development. Combining scalable, highly-integrated system-on-chips based on heterogeneous compute capabilities with a common software framework enables the deployment of applications and services at cloud scale, while meeting the very real demands of a diverse network environment. As a platform, the network will be able to scale to meet not only data bandwidth and capacity demands, but also address power efficiency, data diversity handling and data density.
The network is not homogenous. There are various factors, but most notable are the constraints of power, form factor and latency. If a small cell platform can’t fit within a power over Ethernet budget, it can’t be deployed. If it takes too long to process a packet, the packet is dropped and the network becomes highly inefficient. The balance between meeting these real world networking demands and enabling highly configurable intelligence in the network is the key foundational principle of the intelligent, flexible cloud. It is with the right balance that we can truly achieve distributed intelligence from data center through to end device.
The scalable, system-on-chip frameworks ARM provides are the foundation upon which are partners’ innovation and market know-how are added to address the needs from end to end. Whether it is the ability to process L1 line rates in access nodes, or the capability of high throughput storage access in the data center, ARM partners add networking acceleration, IO and storage capabilities for unified silicon systems. With this scalability, the nodes across the network can become more configurable and flexible. Combined with a common software layer, comprised of a mix of embedded and cloud IT technologies, these nodes become accessible intelligence upon which software networking functions and diverse system, business and services applications can be deployed.
The common software layer enabling distributed intelligence for cloud to the edge is an open source software stack based on Linux, Open Data Plane, virtualization and containerization technologies, and management and orchestration technologies. Industry initiatives like the Linux Foundations OPNFV are building and testing real-world platforms and will be significant contributors to consolidating requirements and bringing together standard software platforms.
We view this launch of the intelligent, flexible cloud framework as a start of a discussion and we look forward to continuing the discussion with ARM partners, our software partners and the broad ecosystem of service providers and OEMs. It is going to be an exciting decade ahead for networking. I look forward to the possibilities and opportunities we can enable together, not only for the industry but also for consumers and businesses who will drive new services and applications for IoT, mobile and more.
To dive deeper into the industry trends, I recommend this brief from Moor Insights. For more on the intelligent flexible cloud framework, an in-depth white paper can be found here.
