With the help of artificial intelligence (AI) and machine learning (ML), predictive network technology alerts administrators to possible network issues at an early stage and provides potential solutions.
Bob Hersch, principal and head of U.S. platforms and infrastructure at Deloitte Consulting, said AI and ML algorithms for predicting network technologies have become critical. “Predictive network technology leverages artificial neural networks and leverages models to analyze data, learn patterns and make predictions,” he said. “AI and ML significantly enhance observability, application visibility, and the ability to respond to network and other issues.”
While predictive networking technology has made impressive progress over the past few years, many developers and observers believe the best is yet to come. “Tools and systems are available now, but like most important evolutions in technology, early adopters are at risk because development is still in flight on how to even measure the effectiveness of the transition,” said David Lessing, director of technology research and technology. Consulting firm ISG.
Predictive analytics is no longer just for predicting network outages and proactively dealing with bandwidth and application performance issues, said Yaakov Shapiro, chief technology officer at telecommunications software and services provider Tangoe. “Predictive analytics are now being applied to solve problems around the network and help address the shortcomings of SD-WAN, most notably vendor sprawl and the need for broader operator service management and telecom cost optimization, “He said. “These have become even bigger issues in an era where MPLS (single-carrier services and dual-carrier services) are exchanged for broadband services that may include hundreds of Internet service providers.”
Artificial intelligence is driving predictive networks forward.
The latest developments in artificial intelligence are the most important developments in predictive network technology. “Cloud-based AI technology can improve the quality and speed of information provided to network technicians while giving them a valuable tool to investigate outages and other issues,” said Juniper Networks researcher Patrick MeLampy. “AI can detect anomalies faster than humans and can even analyze the root cause of anomalies, helping guide technicians to understand and fix issues faster than before.”
Integrating artificial intelligence tools into predictive network technology also has the potential to be an economic game-changer. “With proven AI and ML tools, service providers and organizations alike can reduce the cost of identifying and resolving issues,” MeLampy said. In addition to bottom-line economic benefits, AI can help streamline management within an enterprise or across a service provider portfolio. “Mean time to repair is reduced and end-user satisfaction is improved,” he said.
Bryan Woodworth, chief solutions strategist at multi-cloud network technology company Aviatrix, said network technology is predicted to develop rapidly in the next few years. It has helped resolve network issues quickly and efficiently. “AI can correlate alarm and error conditions across many different systems, discovering relevant patterns in minutes or even seconds that would take humans hours or days,” he said.
Woodworth said predictive network technology can also significantly reduce the number of false positives in logs and error analysis, resulting in more intelligent and useful alerts. “You can’t heal from something you don’t discover,” he said. “For example, you have to know where the problem is before you can change the network to work around it.” Self-healing networks based on AI and ML provide better recommendations on how to recover from errors and avoid outages.
Predictive modeling works best in the data center.
Network Behavioral Analysis examines network data such as ports, protocols, performance, and geographic IP data to provide alerts when there are significant changes in network behavior that may indicate a threat. “In the future, this data can be fed into an artificial intelligence model to help confirm whether the threat is real and then make recommendations on how to address the problem by making changes to the network,” Woodworth said. “This kind of predictive modeling works best in private networks such as data centers, because [that is] where humans have complete control over all network components and the data they generate.”
For public networks, including those connected to the Internet, the task becomes more challenging. Learning models must be designed to compensate for systems that are not under direct control or that are provided with incomplete data sets. That means the learning model’s predictive accuracy will be reduced, Woodworth said, and manual adjustments may be needed to account for missing data.
To be fully effective, advanced AI and machine learning models should be run at production levels and scale for bug fixes, Smith said. “Decision makers need to trust the modeling results and technology sponsors need to execute effectively,” he said.
At the same time, continued advances in cloud technology and graphics processing units (GPUs) are taking modeling to the next level. “Open source and commercial frameworks are helping organizations deploy ML operations quickly and at scale while reducing the risks associated with the time and complexity required to configure cloud and open source systems for AI,” said Maggie Smith, managing director of applied intelligence at the consulting firm. Accenture Federal Services.
Smith said several major cloud providers have implemented AI model optimization and management capabilities. This technology can be found in tools such as Amazon SageMaker, Google AI Platform, and Azure Machine Learning Studio. “Open source frameworks like TensorRT and Hugging Face provide greater opportunities for retraining model monitoring and efficiency,” Smith said.
Artificial intelligence is driving predictive networks forward.
The latest developments in artificial intelligence are the most important developments in predictive network technology. “Cloud-based AI technology can improve the quality and speed of information provided to network technicians while giving them a valuable tool to investigate outages and other issues,” said Juniper Networks researcher Patrick MeLampy. “AI can detect anomalies faster than humans and can even analyze the root cause of anomalies, helping guide technicians to understand and fix issues faster than before.”
Integrating artificial intelligence tools into predictive network technology also has the potential to be an economic game-changer. “With proven AI and ML tools, service providers and organizations alike can reduce the cost of identifying and resolving issues,” MeLampy said. In addition to bottom-line economic benefits, AI can help streamline management within an enterprise or across a service provider portfolio. “Mean time to repair is reduced and end-user satisfaction is improved,” he said.
Bryan Woodworth, chief solutions strategist at multi-cloud network technology company Aviatrix, said network technology is predicted to develop rapidly in the next few years. It has helped resolve network issues quickly and efficiently. “AI can correlate alarm and error conditions across many different systems, discovering relevant patterns in minutes or even seconds that would take humans hours or days,” he said.
Woodworth said predictive network technology can also significantly reduce the number of false positives in logs and error analysis, resulting in more intelligent and useful alerts. “You can’t heal from something you don’t discover,” he said. “For example, you have to know where the problem is before you can change the network to work around it.” Self-healing networks based on AI and ML provide better recommendations on how to recover from errors and avoid outages.
Predictive modeling works best in the data center.
Network Behavioral Analysis examines network data such as ports, protocols, performance, and geographic IP data to provide alerts when there are significant changes in network behavior that may indicate a threat. “In the future, this data can be fed into an artificial intelligence model to help confirm whether the threat is real and then make recommendations on how to address the problem by making changes to the network,” Woodworth said. “This kind of predictive modeling works best in private networks such as data centers, because [that is] where humans have complete control over all network components and the data they generate.”
For public networks, including those connected to the Internet, the task becomes more challenging. Learning models must be designed to compensate for systems that are not under direct control or that are provided with incomplete data sets. That means the learning model’s predictive accuracy will be reduced, Woodworth said, and manual adjustments may be needed to account for missing data.
To be fully effective, advanced AI and machine learning models should be run at production levels and scale for bug fixes, Smith said. “Decision makers need to trust the modeling results and technology sponsors need to execute effectively,” he said.
At the same time, continued advances in cloud technology and graphics processing units (GPUs) are taking modeling to the next level. “Open source and commercial frameworks are helping organizations deploy ML operations quickly and at scale while reducing the risks associated with the time and complexity required to configure cloud and open source systems for AI,” said Maggie Smith, managing director of applied intelligence at the consulting firm. Accenture Federal Services.
Smith said several major cloud providers have implemented AI model optimization and management capabilities. This technology can be found in tools such as Amazon SageMaker, Google AI Platform, and Azure Machine Learning Studio. “Open source frameworks like TensorRT and Hugging Face provide greater opportunities for retraining model monitoring and efficiency,” Smith said.