Twenty-five years ago, during one long Paris evening, I performed an emergency upgrade on a Border Gateway Protocol (BGP) router that isolated half of the France Internet from the USA. I am sure many of us can refer to a similar incident. After all these years, I still remember the agonizing seven minutes while the upgrade procedure lasted. The networking industry was living in the stone age with unmistakable consequences for network engineers when problems arose. Since then, the situation has evolved positively, but there is still a voice in us that confusingly praises the current state of affairs.
There is a latent fear in the network engineers’ community to see jobs eliminated by the advancement of machine learning (ML) and artificial intelligence (AI) capabilities. The question is genuine but lacks precision. Indeed, several tasks we were performing in the past are gone. By the way, who wants to sit in the middle of the night upgrading critical routers? On the other hand, the number of employed IT network engineers has only grown. The role and responsibilities of a network engineers are even more crucial than before, and their assigned tasks much more business-related.
As I look back at the introduction of automation in the networking industry, this innovation created job opportunities and upskilled many engineers. The same will apply to AI/ML. Making the AI more intelligent (so to speak) will be the responsibility of network engineers who will drive the learning process and become data proficient on the job. Network engineers need to broaden their horizons and embrace AI/ML. Currently at Extreme Networks, our system engineers are broadening their AI/ML knowledge and acquiring new skill sets. A good starting point for anyone in the industry is Andrew Ng’s online course: AI for Everyone.
Another elephant in the room to address is the kind of AI we are building for our networks. In a recent blog post, HAL 9000 from the epic film “2001: A Space Odyssey” was part of a discussion about Explainable AI models versus black-box models. In the movie, HAL decides to kill astronauts in order to protect and continue his programmed directives. I can assure you, none of the AI/ML used in ExtremeCloud IQ has even the remote possibility to hurt a network engineer! Instead, we have introduced AI/ML that a network engineer can trust.
Explainable AI (XAI) proposes that any ML model be explainable and offer its decision-making interpretation. XAI also promotes the use of simpler models that are inherently interpretable. We realize that keeping the human in the loop is important when building ML/AI applications, so we decided to build a CoPilot around the concept of XAI for our cloud-based networking management platform, ExtremeCloud IQ.
Our Explainable AI, for the foreseeable future, is a powerful optimizer and an accurate status reporter. Its unique role is to make the network more efficient, always-on, more human in the way it reports metrics. In the not-so-distant future, our Exablianble AI will also have the ability to predict the failure of both hardware and software entities.
Networks are incredibly complex entities encompassing millions of ever-changing states driven by inherently error-prone hardware and software. Network operations have always been frustrating, and I have often felt like I needed a message from a Greek Oracle to provide insights into the operational state of my network. AI/ML might not be a Greek Oracle; however, it can offer network engineers accurate data to support their opinions and plans. Network operations are no longer seen as a cost center but instead as a business planning contributor. AI/ML will offer new tools to IT administrators and, at the same time, bring new exciting work for network engineers.
The way we operate networks is not going to change in a blink of an eye. Adopting AI/ML might start with a single IT project and eventually expand to a company-wide effort to maximize its benefits. If done well, this adoption will strengthen the bond within the company as all teams are engaged in adopting a new paradigm. The route to task optimization through AI/ML forces introspection of our day-to-day job. It is taking each of us out of our routine and makes the journey even more rewarding.
ExtremeCloud IQ customers can immediately embark on this process by taking advantage of the first AI/ML-driven insights and experience how they can extract more value out of their management tools. Every enterprise network needs a copilot. ExtremeCloud™ IQ CoPilot provides explainable ML/AI that builds trust by providing readable output of how insights were derived, enabling you to automate operations, enhance security and enrich user experiences with confidence. To learn more, visit our CoPilot Solution Page!
This blog was originally authored by Benoit Lourdelet, Director of Product Marketing.
