The Paradigm Shift for Telcos: Enabling an Infrastructure for Zero-Latency Computing
By Nitin Serro
The telecommunications industry is at a critical crossroads. Facing tough competition from cloud providers; demand from customers who want faster connectivity to on-demand services and advanced technologies like IoT, virtual reality and robotics, service providers – with their legacy infrastructures, current cloud implementations and existing bandwidth – are not meeting market demands.
The answer needs to be more than “upgrade equipment, add bandwidth and implement software-defined networking technologies against a legacy infrastructure.” They must also build network functions virtualization infrastructures (NFVi); embrace new concepts such as fog computing to enable architectures that support zero-latency computing; and automate service life-cycle orchestration so they have the flexibility to move application resources – network, compute or storage – wherever they are needed.
Oh, and to maximize its potential, zero-latency computing also requires micro services-based applications.
It’s a tall order, but by incorporating these strategies, telcos will come closer to giving customers what they want. Partners need to ask their telco partners about plans and prospects.
Virtualization Next Steps
Let’s look at various terms.
- NFVi provides the framework for compute, storage and networking resources on which virtual networks are built. NFV management and orchestration is the integrated system of components used to operate the NFVi. Virtualized network functions (VNFs) are not part of management and orchestration; they are the components that process production data in the NFVi. VNF is a software function that is decoupled from the underlying physical hardware.
- Fog computing (think in meteorological terms: cloud, closer to the ground) is a concept that allows cloud resources to be logically and efficiently positioned anywhere on the continuum from the data consumer to the data center.
- Zero-latency computing is when no time is lost during the exchange of information between application interfaces; therefore, the system is able to respond instantly to an application-layer request. Moreover, low latency (ideally, zero latency) is essential for applications such as medical robots. A fast, dependable network is critical in a scenario, for instance, when a surgeon conducts remote surgery using robotics. Currently, many medical robots have a two-second lag, a delay that could be problematic when conducting high-stakes surgeries. Many IoT, virtual reality and other next-gen applications also require a strong network that delivers services in near real-time.
What will it take to get there? For one, telcos will have to eliminate the silos that still exist in their infrastructures and organizations so they can become more agile. In the past, it was acceptable for a service change to take a few days or even weeks, but now these have to happen on demand.
Much of the equipment in underlying infrastructures today is legacy gear, with capability limitations. Furthermore, today’s cloud infrastructures are topology-specific, and thus still create bottlenecks when applications traverse multiple services. This reality goes against NFVi’s ultimate aim of orchestrating network assets and resources and optimizing how services are delivered.
There are open-source technologies under development that address this. For example, Service Function Chaining, a capability of the Open Daylight SDN framework, provides the ability to stitch together a predefined list of network services (firewalls, load balancers) to create a comprehensive service chain.
Service providers will for some time yet rely on legacy infrastructure to deliver mission-critical functions and existing services. New NFV concepts will have to find ways to seamlessly work alongside; hence, orchestration and management of all the combined elements through flexible, adaptable and reusable software development for automation will be paramount.
For next-generation applications, orchestration will have to be more robust and adaptive, and be able to scale in time, not just to scale.
In order to provide better services, faster, it is critical for service providers to be attentive to where the industry is headed. They need to start thinking strategically about infrastructure design. By upgrading individual components of their networks and automating them via NFVi, they can transform their overall services offering. To do so, they’ll need partners who grasp the enormous potential of fog computing and concepts of zero-latency computing.
NFVi in conjunction with fog computing is the future, and in order for service providers to keep pace with or leapfrog their competition, they need to reimagine what is possible in terms of how services are delivered. Yes, getting there will require a drastic change in business models, design approach and operations. There are a lot of technical and cultural challenges to overcome. But the earlier telcos start, the better for customers and partners alike.
Nitin Serro is the founder and CEO of Serro. He began a career with Juniper Networks in which he facilitated large-scale proof of concepts for service providers. Nitin worked as an architect for Google and other enterprise service providers as the industry started to incorporate software-defined solutions utilizing open source technologies.