Network performance and analytics tools provide several key benefits for networking managers, including network...
performance visibility and assistance for related areas, such as security and troubleshooting for root cause analysis. Network analytics software is an established product category, with both commercial and open source tools.
But, as network traffic and the number of mobile devices continue to rise, network management has become more complex, and existing oversight tools are facing limitations.
IT managers need a new generation of network analytics software, as enterprises need to understand how the entire network is operating. Relying on a device-by-device view of switches, servers or client devices is not holistic and doesn't scale. Add the adoption of cloud services to the mix, and it becomes readily apparent why enterprises need new ways to view systems and empower their networking staff to work more effectively.
To meet these new demands, a new generation of network analytics software is emerging. These products offer a blend of machine learning, artificial intelligence and cloud-based data processing to help enterprises monitor their networks and troubleshoot problems. Some even learn how a network is supposed to operate and can flag managers for potential problems and offer ways to solve any issues that may occur.
Let's take a look at products from three vendors -- ExtraHop Networks, Nyansa Inc. and Savvius Inc. -- and examine how they can help monitor and analyze network performance.
Different flavors of network analytics software
ExtraHop's primary focus is the analysis of wire data. In 2017, it added a cloud-based machine learning engine, dubbed Addy, which combs through wire data to report on network anomalies. Over time, Addy understands normal network behavior, thus giving ExtraHop customers an early warning system of potential problems.
Addy, along with ExtraHop's appliances, can help managers evaluate problems they may not have considered. I consider Addy an automated assistant. In other words, the technology is like a rescue dog for a search team.
Nyansa, which made its debut two years ago, focuses on client performance through big data analytics. It uses a combination of deep packet inspection and cloud-based analytics, probing WAN links and client devices, and then it displays performance data in an easy-to-understand format.
The company's Voyance platform is tailored to today's environment, where cloud and mobile applications dominate. To that end, I consider Nyansa a cloud, end-user and application-centric view of network management -- like a new set of eyeglasses, perhaps.
Savvius, which was recently bought by LiveAction, specializes in products that capture all network packets for later analysis, like a DVR that constantly records television programs. Products include Spotlight, which offers a quick visualization of traffic based on application or traffic segments. Results are displayed in a dashboard, where data can be reviewed and addressed as needed. While it provides application performance analysis, its viewpoint is centered foremost on the network.
I consider Savvius a tool for network managers who understand their jobs well and need something that can provide deeper diagnosis capabilities. Think of Savvius as a better dashboard for airline pilots.
Which approach is best for network analytics?
All these products can help network managers pinpoint network and application performance issues. They can also help managers charged with monitoring software-as-a-service applications, as well as traffic flowing through their own data centers.
Used effectively, these network analytics tools can improve security, speed problem-solving, lower operating expenses and increase agility. And despite the fact they aren't considered traditional security applications, these analytical platforms can help secure networks by identifying problem areas and fixing them.
As you evaluate network analytics software, is one approach better than another? Not really, as this depends on each enterprise's requirements. But consider these tradeoffs: Complete automation can boost novices' ability, but it can create complacency and blind spots. On the other hand, manual analytics software may intimidate beginners who don't know where to start. A novel approach may look intriguing, but it may not fit smoothly into existing IT workflows. What matters depends on who you are.
Similar capabilities are available from vendors large and small, but the examples cited represent some different approaches to tackling traditional and modern problems. Existing tools may be reaching their limits, so I recommend enterprises work with their existing providers to see how they can assist with their new network performance and analytics requirements.