The industry is in the midst of a perfect storm. We must support billions of devices and things that generate an incredible volume of data, prepare for the imminent arrival of 5G, and continue the multi-year, network transformation process by cloud economics across the network to the intelligent edge. At Intel, we view this wave of innovation as data centric computing with a central goal of unleashing the power of data to fuel new business opportunity and societal good. Data centric computing demands underlying infrastructure that supports high demand workloads, such as AI, analytics, visual cloud and media delivery and high performant network functions, and delivers them at location of highest return across traditional data centers through the network and to the intelligent edge.
While mobile data gave way to the digital service economy, 5G and the Internet of Things (IOT) will now help lead to data centric computing era and an equal disruption of ecosystem innovation. This will deliver exciting 5G and edge services that provide the necessary ROI for service providers and enterprises to place data center-grade network, compute and storage at the edge.
Now imagine that you are a developer, who has created one such promising application that has the potential to drive tremendous business value and revenue to a service provider or enterprise. Further imagine that this application can run anywhere in the cloud, or on any edge location, or in the public or private domain, or on any cloud framework, or in a network managed by any service provider. Now imagine even further that you are this developer who has been given an easy button to abstract away the complexity of the network. Seriously, who wants to deal with the esoteric complexities of the network, such as multiple access terminations, LTE or 5G, or any of our other favorite networking alphabet soup technologies? Providing this easy button to abstract away the complexity of the network can help unleash hundreds of thousands of cloud developers on the edge. It can help create the plethora of applications and services that can drive the ROI for our service providers and enterprises.
Defining the edge: On-Premise and Network
The edge can be defined primarily as the on-premise edge and the network edge. The on-premise or enterprise edge has compute nodes or edge platforms that are located on-site with the customer or enterprise premises. The network edge is located off the customer’s premises (hosted in a network node) and is primarily comprised of edge platforms or nodes in the wireless access, wireline or cable access, and the next generation central office. Any of these edge locations can host network and IT-based services, in addition to the virtual network functions. These locations can have additional edge data centers or servers called multi-access edge computing (MEC) platforms to host these services.
Intel is extending its data centric leadership to the edge in three primary ways:
- Intel is developing the right data centric infrastructure with optimized performance that our customers can use to develop strong edge solutions that run a wide variety applications.
- Intel is enabling multiple workloads via networking innovations and software optimizations, including contributions to open source in areas such as data plane processing, orchestration, telemetry, service assurance, closed loop automation, computer vision and inferencing workloads. These optimizations provide both the flexibility and scalability needed to build highly capable on-premise and network edge platform solutions.
- Intel continues to partner with our broad and mature ecosystem to ensure our hardware and software platforms can be easily consumed and in a standard way.
Universal Customer Premises Equipment (uCPE) is a great example of an on-premise edge platform that incorporates virtualization to operate multiple virtual network functions, such as firewalls, routing, intrusion prevention systems, Wi-Fi, and SD-WAN, on a single server. Many CommSPs see uCPE as the ideal platform to deliver managed enterprise services with enhanced data analytics, AI and high-speed processing.
uCPE may support services across any industry, but let’s look at retail and industrial examples. Smart retail stores can generate terabytes of data a day from many IoT devices and require on-premise, edge nodes that collect, store and analyze the large amount of data for use across a number of back-end, e-commerce or point-of-sale systems. In order to truly enable smart factories, we need to support world-class, on-site data processing, network and cloud capabilities to bring intelligence to the machine but also connect machines in new flexible ways. At MWC, don’t forget to check out the visual cloud super demo in the Intel booth to see an example of this technology innovation at work.
Intel® Select Solutions for uCPE makes deployment easy with benchmarked hardware and software configurations. At MWC, Intel will also demonstrate Dell EMC Virtual Edge Platforms featuring VMWare ESXi* and NSX-T* along with Intel Atom® and Intel® Xeon® D processors with Intel® Ethernet, Intel® QuickAssist Technology (Intel QAT), and DPDK running converged uCPE workloads, like SD-WAN.
The network edge includes the wireless access or radio access nodes (RAN) and base stations, wirelines and fixed cable access and the next-generation central office (NGCO).
OpenNESS: Easy button for developers
Now let’s get back to how we are making easy for our developers to create the exciting 5G and edge services that are the real driving force behind edge transformation. The industry needs a strong developer ecosystem for the on-premise and network edge.
That is why Intel is developing and making available an open reference software toolkit – called “Open Network Edge Services Software (OpenNESS)” that abstracts the complexity of the network and provides an easy button to “network-as-a-service” capabilities to the developer.
OpenNESS offers cloud and IOT developers an easy-to-use toolkit to develop and deploy applications at the network edge or on-premise edge locations. By abstracting out complex networking technology, OpenNESS exposes standards-based APIs from 3GPP and ETSI Multi-access Edge Computing (MEC) industry group to application developers. Using this software toolkit, applications can steer data traffic intended for the edge at 5G latencies. This is a very critical capability as it not only abstracts the various network access terminations from the developers, but enables seamless movement of data and traffic from the devices to the edge and to the cloud and back through an intelligent, scalable, and flexible network. Other features include a microservices based architecture enabling easy integration of new APIs, support for heterogeneous platforms (FPGA, AI Accelerators), and support for scalable container and VM orchestration.
Future use cases are a combination of analytics, media and visual cloud capabilities and networking. Hence it is important for OpenNESS to integrate with analytics and inferencing frameworks, such as OpenVINO, Media SDK, and visual cloud software optimizations, to deliver various types of media workloads in the most efficient and effective manner. We believe it is important for OpenNESS to be cloud-agnostic, so that framework will provide plugins to various cloud service provider frameworks.
OpenNESS offers CommSPs the opportunity to deliver strategic, business critical services to their customers and therefore generate new revenue opportunities. TEMs and OEMs can use the OpenNESS edge toolkit and APIs to expand their reach into an ecosystem of proven, application developers. These companies may offer a range of server platforms, based on Intel architecture, that operate with a common, open source edge framework. Enterprise IT and development teams will be able to develop and deploy end-to-end edge services in a lab, field trial or proof of concepts. Once tested, developers may load the applications into the OpenNESS environment to support the seamless portability to any edge or cloud-enabled, production environment.
The developer community will be able to build software applications that can run on any OpenNESS-supported platform. They can then port existing cloud applications to the on-premise or network edge locations on standard, high-volume servers. The community will then develop using the OpenNESS toolkit the potential integration into ecosystem of TEM, OEMs and the rapidly expanding universe of 5G services those edge platforms will support.
Is OpenNESS real? Come visit the Intel booth at MWC 2019 to see OpenNESS in action in six of our demos including the Industrial, Retail and Media demos with applications orchestrated on Intel Edge Platforms over the 5G network. Other demos are the Adaptive Edge demo, the Alibaba Road Side Unit demo, and the ZTE media demo with China Unicom.
The Adaptive Edge demo is a multi-edge location demo across vRAN, MEC, NGCO and cloud. It is adaptive with an ability to orchestrate and manage applications across different edge locations in a scalable, flexible and dynamic manner. It also demonstrates architectural consistency across multiple edge platforms for the ecosystem who can build an application for cloud and run it on any edge location. It supports an open ecosystem built on open source components versus proprietary closed software. It also leverages the Intel Rack Scale Design which supports dynamic resources (compute, network & storage) management in the platform.
The ZTE demo will demonstrate its 5G-ready capabilities from access to core network equipment, all based on Intel technology. Running on top of a ZTE edge server and MEC platform, the demonstration will show how the system can host multiple applications, including real time collection and distribution of 8K video streams. China Unicom will use this for the upcoming Beijing 2020 sporting event.
Visit the OpenNESS web site at https://www.open-ness.org to find details about the OpenNESS program. Source code and documentation will be available in an open source repository in 2019. An early access version is available directly from Intel. You can access source code and documentation from the Intel Network Builders website.
No matter how you define edge computing, there’s no disputing the vast opportunities for technological and business innovation it creates. OpenNESS begins the process of establishing an open community of developers and ecosystem that have operated in different worlds. With the convergence of cloud and networking, it’s time to remove any artificial barriers. It’s time to provide an easy “network” button for the developers to unleash new edge services and applications. Edge is the new cloud!!