Enterprise IT strategies have changed across the board in the past decade. Perhaps the greatest change is the adoption use of cloud computing. The cloud has been the logical response to an increasingly distributed environment for personal devices, the IoT, mobile workforces. 

We’ve seen the emergence of public cloud computing, a model that has developed to the point where many enterprises are choosing to spread workloads across multiple providers of public cloud services according to suitability. This shift to the cloud has reduced the burden and, in some cases, the costs for IT managers and enabled companies to provide an improved experience for customers and employees. But, that’s not the whole story.

Public cloud  

Powering operations through a public cloud provider adds new considerations.  For example, organizations need to ensure data governance. Additionally, cloud operations can complicate long-term budget forecasting. Findings from ACG Research indicate that applications with high data transfer requirements and a demanding compute overhead may be more expensive to run in a public cloud versus a private cloud built on data center infrastructure. 

With this said, if an organization continues to operate data centers, digital business needs will require IT teams to get things done faster and for a lower cost. Fortunately, advancements in data center architectures and operations tools are available to let IT teams adopt public cloud practices within their on-premises facilities to meet these requirements.

By bringing together a large number of common servers and connecting them uniformly with a high-speed fabric network, any software can be run on any server or set of servers. This fungibility of resources offers flexibility to keep up with shifting workload demand. By adding closed-loop network automation powered by AI, the resulting consistency of operations ensures reliability and enables speed without tradeoffs.

Closed-loop delivery

Closed-loop automation means the network understands how it should design, build, deploy, and operate. This type of network can self-diagnose problems as they happen and has the intelligence to aid rapid remediation so that when problems occur, the network can quickly return to its normal operation. 

The latest closed-loop automation technology is designed to understand not only a set of commands but also the business intent. In this way, operators need only to specify the desired services, and the intent-based management system takes care of all the details to set up the network. Operators no longer need to configure the network device by device, with the inherent risk of human errors or overlooked details. The same intent-based management system then validates the network is working as expected and monitors for any problems, including early warning brownouts and grey failures. During problems, the most powerful of these systems can utilize underlying technology known as a knowledge graph for relationship tracking to quickly identify the root cause of an issue across many possibilities and then automatically takes steps to resolve them. 

In this way, the system delivers automation for each stage of the network’s life cycle — from network design, often known as Day 0, to deployment, commonly referred to as Day 1, to the operations of running the network, Day 2, and beyond. By tying together the ongoing operations of the network with its design intent, closed-loop automation results in fast remediation and high overall uptime. Through automated workflows, the same framework can also support fast, consistent, and self-documenting change management for new services delivery.

Finding Balance

Thinking beyond the traditional data center model opens up options to find the right balance between on-premises and public cloud services. No matter which combination of options, leveraging automation to simplify operations is key. Everything from the initial level of design and deployment to everyday operations and assurance can be automated to help manage the increased complexity.