As networking transforms to a more software-centric, cloud-native technology, there is now an increased market focus on the intersection of edge computing and network service delivery. With the increasing variety and availability of compute resources for enterprise IT rollouts, network infrastructures need to evolve to meet the challenge of supporting the new generation of digital initiatives.
IT leaders are looking for ways to make rapid and automated changes to their underlying network infrastructures in a similar fashion to how their applications, compute, and storage resources are delivered and managed using software-defined technologies. But conventional networking architectures that rely on monolithic network operating systems simply do not get this done.
What’s wrong with this picture?
Users and things within the enterprise are now connected using a variety of different networks – from the corporate LAN to enterprise Wi-Fi, private LTE/5G to broadband Wi-Fi hotspots, or public LTE 5G networks. Meanwhile, the application layer is becoming increasingly dispersed.
Instead of the traditional on-premises or public cloud options, there is now a continuum of compute locations that range from traditional public clouds to multi-tenant edge clouds, on-premises extensions of public clouds, and centralized or distributed private clouds using virtual machines.
Unfortunately, today’s conventional enterprise network and security architectures have not been designed and built for such a dynamic and diverse structure of infrastructure deployment options. This has resulted in a patchwork of solutions that are as different as the various combinations of network access and compute locations – creating an overly complex infrastructure unable to keep pace with the rapid change driven by digital transformation efforts within the enterprise.
Traditional monolithic network architectures, where discrete network services are individually deployed using custom hardware and software, have become operationally inefficient. In fact, the lack of agility in networks has resulted in efficient and secure connectivity being the number one challenge to the adoption of new applications – especially new automation applications.
The explosion of latency-sensitive business applications and connected smart machines have created the need for a network architecture that connects the devices and applications efficiently and securely in a uniform architecture independent of device or application location - private data centers, public clouds, or the metro edge.
The transformative impact of 5G technology
Ultimately, organizations are looking to build a uniformed, scalable, and highly agile network infrastructure that brings vital network services closer to users and applications to realize significant savings on capital and operational costs – while achieving the highest levels of performance wherever business demands dictate.
Fortunately, the transformative principles of 5G technology have opened the door for enterprise networks to adapt a new "edgeless" enterprise architecture that calls for convergence of network services, enterprise applications, cyber security tools, and WAN optimization on cloud-native, software-defined edge computing platforms that can be deployed anywhere.
Unlike traditional enterprise network technologies, 5G not only delivers a more deterministic means of wireless network access but also mandates the use of a flexible and agile network operating system based on a combination of software-defined principles, cloud-native microservices, and policy enforcement via strict network slicing.
Enter the edgeless enterprise
An edgeless enterprise 5G architecture is based on the premise that there is no one “edge” for an enterprise. Instead, the network needs to automatically shift the delivery route of applications and services based on performance, security requirements, and the real-time health of network paths. This is achieved through a convergence of network and compute and the creation of a cloud-native network service overlay that delivers policy-based routing, QoS, and security segmentation functions for a consistent application experience.
Vital networking services are deployed at one or more edge locations, including on-premises, metro edge clouds, or public/private clouds. In turn, crucial network services can be service chained with other third-party cybersecurity and SD-WAN services to create a unified service edge.
The cloud-native SDN framework in such a model allows the AI-based automation of dynamic deployment and movement of 5G network functions at different application edge locations. Each individual network function is deployed as a microservice with the ability to dynamically allocate resources to them as demands dictate.
Using this model, enterprise 5G networks create a unique architecture not possible in past enterprise wireless networks. While the control plane governs higher-level network management and policy functions such as private spectrum access, mobility management, authentication, and network resource allocation, the data plane performs tunnel encryption, network switching, and dynamic policy-based routing across different device groups and applications.
And instead of IT having to manage both an enterprise LAN/WLAN and a separate private 5G using disparate technologies that require the translation of 4G/5G traffic flow into legacy QoS policy, an edgeless enterprise framework unifies the policy structure throughout the entire enterprise infrastructure.
More importantly, the network itself can now be made as smart as the enterprise apps to automate identifying the apps and self-managing to deliver the performance needed on a per-application basis. Finally, enterprise IT teams can stop saying "stop" to new applications and instead say "bring it on."
The illustration below illustrates how the edgeless enterprise architecture plays out in practical deployment practices.
What's it all mean?
As the critical applications in the enterprise are deployed across a variety of on-premises, private, and public cloud options depending on their compute requirements, the traditional definition of the "Edge" no longer applies.
With an edgeless architectural approach, enterprises can now rapidly deploy essential network and policy services wherever they find enough computing resources. They can migrate their traditional network management tasks for status monitoring towards KPI-based insights about the health of virtualized network resources. It also opens the door to a wide range of business benefits, including:
- Gaining and maintaining a competitive edge by rolling out new automation and digital transformation services faster than ever before,
- Removing friction by streamlining network and security service provisioning,
- Reducing operational cost inefficiencies by leveraging AIOps with a cloud-native network operating system, and
- Decreasing security exposure by unifying the policy framework, independent of network access and compute location options.
The result is an end-to-end enterprise network that, for the first time ever, is cloud-native and highly adaptable to changes in business technology operations in the twenty-first century and beyond.
Rajeev Shah is the Co-Founder and CEO of Celona.
