Guru Ganesan, President and Managing Director, ARM India
According to a report by IDC, by 2018, 50 percent of all networks will transform from having excess capacity available to handle IoT devices to becoming network-constrained, with 10 percent of sites completely overwhelmed with data. To address this, IDC has predicted that by around the same year, 40 percent of IoT data will be stored, processed, analyzed and applied near or at the network management for LAN and WAN applications in general.
Leaders in the networking industry have talked about intelligent networks for years. Today's reality has evolved beyond putting intelligence in individual devices to distributing systemic intelligence in network layers and functions, end-to-end topologies, and services. The ability to create simplified management and orchestration layers that abstract underlying hardware and software complexity is here, bringing great excitement to the idea of programmable networks where services, resources, and capacity can be spun up or down on demand. Distribution of intelligence is opening up the possibilities of moving away from inefficiencies of forced cloud-based virtualization and creating a better technology model that leverages cloud's best practices. In ideal situations, distribution of intelligence will be as abstract as simply selecting the in-network locations for a distributed application based on accessible power, performance, and capability and latency metrics. The industry and supporting ecosystem is beginning to collectively put the intelligent flexible cloud on a trajectory of maximum effectiveness.
The great leaps forward in the definition of network intelligence today are due in large part to three key enabling technology standards or strategies:
These technology standards and architectures are part of the foundation of an intelligent flexible cloud. It is flexible because it can easily and quickly address diverse network requirements. It is intelligent because it leverages business, customer, and network data to enhance existing services and to use as the basis to create a dynamic environment for the enablement of new services that will be highly innovative and competitive.
With relentlessly increasing traffic and diverse applications, existing networking infrastructures are not capable of scaling to handle future requirements. These include the Internet of Things (IoT), which stretches network topologies to encompass numerous different devices. As the network extends to serve applications in sensors providing telemetry, appliances in home-energy monitoring, computers in automobiles providing navigation and component health checks, and many other use cases, complex requirements will have to be satisfied. These include privacy, performance, security, and integration. The advent of 5G wireless standards brings new challenges as performance metrics are redefined. Networks will be expected to deliver much greater throughput, much lower latency, ultra-high reliability, much higher connectivity density, and a broader mobility range. This enhanced performance is expected along with the capability to control a highly heterogeneous environment and ensure security, trust, identity, and privacy. The 5G architecture is expected to evolve to include modular network functions that could be deployed and scaled on demand to accommodate various use cases in an agile and cost-efficient manner. So the birth of such intelligent and flexible clouds is inevitable in many ways.