Drive business success through distributed edge computing
Edge computing is becoming mainstream. For future business success, service providers need to deliver high-quality customer experiences and monetize their network infrastructures through simplified access for partners and application and content developers. They also must make edge resources such as computing and storage capacity easily consumable for new vertically targeted services. Creating a platform for low latency can help achieve three important goals:
To prepare for 5G, service providers need to update their infrastructure and embrace a distributed cloud model that includes multi-access edge computing (MEC). These improvements need to support:
The changing dynamics of edge computing lead to:
A significant 5G differentiator will be a better user experience from the reduced latency that results from edge offload. Consider this question: would you pay more for a faster data plane or for a high-fidelity, real-time gaming experience that has advanced augmented reality (AR) functionality?
Network complexity leads to high costs and a slow time-to-market. Simplifying your operations through full-lifecycle automation and a reduced number of touch points becomes critical. The problem is compounded by highly distributed cloud workloads, which can have a direct impact on total cost of ownership. A consistent automation tooling for integration, validation, and operation ensures a faster service rollout.
It’s important to expose network resources to global reach partners and application and content developers using the same set of edge APIs. These APIs use standard cloud consumption models, which industrialize the underlying infrastructure.
A ‘Push button’ deployment of 4000 edge sites co-located at “Group Centers” (owned by Nippon Telegraph and Telephone – NTT) and 50,000 Radio sites in Rakuten.
72 hours from ‘Site delivery to first call’ for virtualized EPC rollout.
Zero-touch OpenStack software release upgrade and patching in less than 24 hours.
Today, the typical round-trip time (RTT) latency between an end-user and cloud or content distribution center (CDN) services is around 150 ms. This time includes:
New cloud applications such as interactive gaming and immersive augmented and virtual reality suffer from higher latency times.
To meet customer application demands, you need to determine where the service edge needs to be and where the low-latency services edge exists in your network. You should start with the initial service requirements, which are typically 10–30 ms. Examining service needs is better than a bottom-up-approach which could inadvertently result in thousands of difficult to manage edge cloudlets put as close as possible to cell sites, whether or not there’s a real need for them.
A properly designed transport network supports low latency through well-known quality of service mechanisms, which reduce queuing delay. It also supports new technologies such as low-latency segment routing, which allows the concentration of functions at central office locations. This approach provides an optimal balance between app requirements and economic efficiency.
Together with MobiledgeX, Cisco has built the first public mobile edge network. It provides sub-30 ms latency anywhere in Germany. The network is already live in six sweet spot locations and will be extended over time. It will open up lower latency windows down to 1 ms as the new 5G capabilities are deployed into the network.
The network functions virtualization infrastructure (NFVI) cloudlets in this foundation are enabled by Cisco NFVI solution that is powered with Cisco Virtual Infrastructure Manager (CVIM). The embedded function-as-a-service (FaaS) system, dynamically places application backends as close to the requesting mobile applications as possible. It uses device- and platform-independent software development kits that connect users automatically to the nearest edge location.
Application developers can use edge APIs to provide a better user experience:
The ingredients for a successful low-latency setup start with a distributed telco cloud that is fully automated. It should feature a loosely coupled, modular architecture where the integration between layers is driven by open APIs and data models. The modularity of this architecture allows compliance to any disaggregation requirement and the ability to scale from hundreds to thousands of locations.
The distributed telco cloud is composed of two components:
To provide a fast vSwitch data-plane, the carrier-class cloud solution is built on a standards-based architecture with various open source components such as OpenStack, KVM, Linux, Docker, Kubernetes, and fd.io. The platform is engineered to enable optimal footprints that range from more than 100 servers down to three servers in physically constrained locations like cell-site cloudlets.
Key capabilities include vGPU support for offloading AR/VR processing for rendering and encoding and future cloud-native support for hosting container workload on bare metal servers.
The 5G-ready transport infrastructure foundation is a lightweight software defined networking (SDN) implementation that is based on innovative segment routing technology (SRv6, FlexAlgo). It provides deterministic end-to-end latency paths from Internet peering down to host-level at the farthest edge.
All network infrastructure and cloud components are managed and controlled by a common end-to-end automation and orchestration layer that provides a single view for service and function management.
With the proper setup, you can reach partners and provide access to a global application developer audience and community. The main objective is to establish a normalization layer with the same set of APIs across different infrastructures. A global edge cloud SaaS portal allows operators to visualize application delivery performance and developers to deploy their application container. This approach gives application and device makers a good experience so they can get applications into production quickly.
Low-latency offers a better customer experience and is becoming the new growth engine for network service providers. With expertise, innovation and service excellence from Cisco and our partners, you can monetize your network, save money through innovation, and lower network transformation risks with proven methodologies and expertise you can trust.