Configuring (and managing) today’s massive corporate networks has become completely overwhelming for IT staff, but the rise of AI-based networking may soon offer a solution.
“I believe we all recognize that in 2016 we hit an inflection point,” declared Cisco CEO Chuck Robbins during his keynote address at Cisco Live 2017 in June. “The number of new machine-to-machine connections that were added to networks (during 2016 Q3) exceeded the number of phones and tablets added to networks. We’re moving into a world of unbelievably massive expansiveness.”
Driven in large part by the rise of the Internet of Things (IoT), corporate networks — especially large enterprise networks — really are becoming more sprawling and complex. As McKinsey & Co. points out, “Earlier, a large corporate network might have somewhere between 50,000 and 500,000 endpoints; with the IoT, we are talking about millions or tens of millions of endpoints.”
And while an office full of smart devices promises to boost productivity in countless ways, it also complicates the process of corporate network configuration. Crafting a coherent, responsive network capable of supporting a ten- or twenty-fold increase in endpoints is no small task, and all but demands a radical shift in standard practice.
On the plus side, IT administrators have access to all the technology they need to execute such a shift, something that couldn’t have been said as recently as five years ago. Effective network configuration today is defined first and foremost by flexibility — that is, the ability to adapt to fluctuations in network traffic, resource demands, and security protocols.
Virtualization enables IT administrators to build and maintain networks that can be provisioned and secured with a certain degree of dynamism. Once a network expands past a certain point, however, segmentation becomes necessary. For a sprawling network, centralized decision-making (regarding resource provisioning, routing, etc.) is inefficient or even impossible, meaning different network “decisions” must be delegated across the entire network architecture.
But how exactly should IT administrators handle this delegation, both as they configure their networks and as they manage them? Industry insiders like Cisco’s Robbins are advocating for widespread adoption of intent-based networking, a new iteration of software-defined networking that effectively automates a variety of traffic management tasks by pushing low-level rules to switches across vendors and scripting languages. As we’ve discussed previously, however, intent-based networking is not without considerable flaws.
As an alternative, IT administrators should consider an even more cutting-edge network management mechanism: artificial intelligence (AI). A machine learning algorithm (a particularly popular subset of AI) is capable of collecting, organizing, and analyzing network data and, over time, using that data to create a baseline model of the network’s traffic. This model can be used not only to determine which network path is best for any given data packet at any given time, but to execute critical network optimization tasks like deduplication and compression, as well.
That said, AI won’t eliminate the need for highly-skilled enterprise IT teams altogether. “AI can show us new ways of designing faster, more cost-effective networks, but the actual build of the network must still be done by humans,” explains Network Computing contributor David Eden. “Combining human intellect and creativity with the massive computing power of AI will create a situation in which new design and management techniques may be created that humans couldn’t build on their own, but self-improving intelligent algorithms will harness over time.”
As corporate network complexity continues to deepen, the importance of strong networking partnerships will only grow. Moving forward, this will likely include AI-enterprise partnerships, but for the foreseeable future, it will also include partnerships between enterprises and experts in network management.
At Turn-key Technologies, we’ve been helping optimize the speed, reliability, and security of large corporate networks for more than two decades, and we bring industry-leading expertise to every one of our enterprise partnerships. Our wide range of networking certifications enables us to craft customized network infrastructures that are as high-performing as they are secure, something that we recognize is a central component of any modern business.