The Growing Necessity of Real-Time Networking

Thursday May 25th 2017 by Arthur Cole
Share:

Streaming analytics, the Internet of Things and hybrid cloud computing are driving the need for faster network performance.  

The pace of business is accelerating in the digital services era, which means enterprise networks must be not only fast, but lightning fast. In fact, real-time performance is quickly becoming the norm, and the fact that data volumes are orders of magnitude greater than they were just a few years ago is irrelevant.

Compounding the problem, network topologies no longer adhere to traditional north-south or client-server patterns. Instead, they are gravitating toward flexible fabrics where data is expected to flow east-west across multiple endpoints both within the data center and over the wide area.

One of the ways this is playing out is through new networking protocols that bring real-time performance to legacy environments. The Open BGP Monitoring Protocol (OpenBMP), for example, is designed to monitor Border Gateway Protocol (BGP) sessions that govern autonomous routing on the Internet. The Linux Foundation recently adopted the framework for its Streaming Network Analytics System (SNAS) where it is expected to provide a mechanism for identifying route views for real-time analytics and other applications. As Cisco’s David Ward explained to ENP, the goal is to allow DevOps teams to understand the dynamics of route topologies so high-speed, high-volume workloads don’t overload available networking resources.

Real-time performance is also expected of emerging hybrid architectures where the smooth flow of information between public and private clouds is required by key applications. A company called FusionLayer recently launched a new version of its Infinity network management solution, which provides for real-time management of resource assignments in multivendor environments. The system centralizes and automates virtually the entire network management process, leveraging common security and compliance functions in conjunction with providers’ virtual private cloud offerings. The system enables real-time visibility and auditing of networks that are activated within public cloud infrastructure as well.

The emerging Internet of Things (IoT) is being purpose-built around real-time networking, particularly on the edge where delayed responses for devices like autonomous cars can be life-threatening. Storage systems developer X-IO is out with a new edge platform designed to meet the real-time requirements of high-volume streaming environments for key industry verticals like finance and cybersecurity. The Axellio platform leverages the company’s NVMe FabricXpress technology to provide sustained throughput across end-to-end networks at up to 480 Gbps. Edge modules contain dual-port NvMe SSDs and dual-CPU servers to provide up to 88 cores, 2 TB of RAM and 1 PB of flash storage, all within a 2U form factor.

And on the industrial side of the IoT (IIOT), designers are quickly implementing time sensitive networking (TSN) on their Ethernet platforms. This is expected to provide high determinism, low jitter and other benefits to real-time manufacturing and industrial control applications. As explained by Belden’s Rene Hummen, TSN brings uniformity to the current crop of real-time Ethernet protocols, such as EtherCAT and Sercos III, while also enabling higher bandwidth connectivity and improved transparency between industrial and enterprise networks. A key element of TSN is the Time-Aware Scheduler that prioritizes data Ethernet frames based on transmission time. This is crucial for the highly automated processes that are infiltrating robotics-driven product assembly lines.

At first glance it may seem that real-time networking is being driven by the enterprise’s need to accommodate an increasingly data-hungry, and impatient, user base. But the fact is that high-speed communications is more a factor of emerging machine-to-machine (M2M) environments than those that require human input.

As automation takes over more facets of our daily lives like driving, bill-paying and shopping, split-second delays in data delivery will become increasingly intolerable.

Data process and storage infrastructure has long had the ability to manage bits in real time, so it is now up to the network to eliminate the lag that comes from moving data from place to place.

Arthur Cole is a freelance journalist with more than 25 years’ experience covering enterprise IT, telecommunications and other high-tech industries.

Share:
Home
Mobile Site | Full Site
Copyright 2017 © QuinStreet Inc. All Rights Reserved