Edge networks are evolving towards intelligence and computing enhancement

"Always on, always connected" has become a deeply rooted lifestyle, and mobile phones play a vital role in it. They allow us to obtain data anytime and anywhere, and stay in touch with others through a variety of communication tools in real time. This way of obtaining information has fundamentally changed the way we make decisions and further reshaped our behavior.

Data from Cisco shows that by 2022, the global mobile network will have more than 12 billion mobile devices and IoT connections. And these mobile devices will support a variety of functions. Today, mobile phones have replaced the functions of various gadgets in life. For example, mobile phones can provide WeChat Pay, Alipay, Apple Pay, Google Pay or other electronic payment functions, so there is no need to carry a wallet. When the phone can unlock and start the car or open the garage door, there is no need to carry the car keys. Mobile applications currently also include real-time streaming services, support VR/AR experience and real-time sharing. Although future mobile services and applications seem to be full of infinite possibilities, they need the support of the next generation of data infrastructure to truly land.

The Need for Intelligence at the Edge

With the development and popularity of new data-intensive applications, network connections and traffic continue to grow, all of which put higher demands on network bandwidth and give rise to the development of a more intelligent infrastructure. This infrastructure can identify specific application and infrastructure needs through intelligent technology and perform processing operations on the edge network when needed. With the development of Multi-Gigabit Ethernet and 400GE backbone network connections, network speeds continue to increase, but the latest 5G and Wi-Fi available bandwidth still has bottlenecks in backhaul. At the edge network end, enhanced processing helps alleviate the pressure of moving large amounts of data across the network. This higher level of network intelligence enables the network to provide complex software-defined infrastructure management without user intervention, as well as management reasoning engines and application policies. Most importantly, active application functions can be provided. By providing a near real-time interactive platform based on a low-latency, highly reliable and secure infrastructure, the user experience is effectively improved.

With bandwidth demand growing so fast, how do we address this problem at scale? If we look at it from a parallel perspective with cloud data centers, we believe that increasing processing power at the edge of the network is a good way to effectively scale and handle the increasing bandwidth and number of nodes. In the data center, this can be achieved by using smart NICs to offload complex processing tasks from the server, including packet processing, security, and virtualization. Carrier networks take a similar approach by deploying AN/uCPE/vCPE devices at the edge of the network to perform intelligent processing while reducing connection costs. However, this approach has problems when dealing with enterprise networks where endpoints are diverse and bandwidth demand first occurs at the network access layer.

Smart use of artificial intelligence (AI)

Another challenge arises when services are deployed in enterprise networks using traditional approaches, such as centralized firewalls and authentication servers. With the expected increase in devices accessing the network and the bandwidth required for each device, the inherent limitations of these traditional approaches can cause bottlenecks. To address these issues, we must look to the edge of the network to move processing closer to where it is needed and make it smarter. Network OEMs, information technology (IT) infrastructure owners, and service providers can all address the problem by leveraging the next generation of artificial intelligence (AI) and network function offload at the access layer of enterprise networks.

How to survive in the edge network era

This is the first in a series of articles that describe the essential technologies needed for the growing borderless campus as mobile and cloud applications proliferate and network functions move from the core to the edge. In this article, we discussed the trend toward network intelligence. In this second article, we will explore the performance levels required and provide more insights and responses on how to survive in the edge network era.