What Is Network Analytics?
Network analytics is all about seeing the whole situation. It’s akin to an airport control tower.
“A good analytics tool will provide visibility across the entire network, from the wireless edge to the data center and across multicloud or SD-WAN environments,” says David Savage, vice president of sales for Extreme Networks.
The key is automation via machine learning. Aruba tracks the network activity of more than 100,000 customers in a data lake.
“We pretty much mirror the market,” says Larry Lunetta, vice president of wireless LAN and security solutions marketing at Aruba. “As a result, we can focus on a customer network, profile it, look across data parameters and see if that network is being optimized.”
EXPLORE: Why ports are turning to IT networks for visibility into cargo data.
Machine learning uses standard network behavior as a baseline to spot and report deviations. That information displays on a dashboard that allows administrators to track everything in real time, and network administrators can customize the information they receive.
“If you don’t want to look at something, you can actually set up the rules for what you want to see,” says Brent Potter, a global government strategist for Cisco.
Cisco’s DNA Center product offers a dashboard that allows network design, the creation of user and device policies, and quick provisioning. It also allows modeling of gear changes. “So, we’re going to add this new piece of equipment into the heart of the infrastructure,” Potter says. “What’s it going to break?”
Network operators can preprovision to avoid potential crashes.
What Is Predictive Network Analytics?
Predictive analytics is the use of Big Data, statistical algorithms and machine learning to identify possible future events based on historical information. It’s as close to a crystal ball as you’ll find in networking technology.
“By taking baseline data and applying it to future scenarios, predictive analytics tools can identify when a scenario is likely, such as a potential entry point for unauthorized network access or when unusually heavy network traffic may start to cause issues for users,” Extreme Networks’ Savage says.
Extreme Networks offers a sandboxing platform that creates a digital twin, an artificial intelligence–driven simulated network that includes data traffic. “IT teams can run potential scenarios in a consequence-free environment and determine what actions they would take ahead of time,” Savage says.
DISCOVER: The current landscape for AI in state and local government.
Bear in mind that like a weather forecast, predictive analytics isn’t always perfect.
“Algorithms are mathematical constructs that run across massive amounts of networking data, looking at many variables in your network, applications and internet paths,” writes JP Vasseur, a machine learning fellow at Cisco, in a blog post. “It’s important to understand that predictions may be incorrect, and they may not account for some scenarios.”
So you need to be realistic. “We should set our expectations that there are limits to predictive analytics, but at the same time, not let that keep us from extracting the full potential value that we can obtain,” Vasseur writes.
