More than even before, companies rely on WANs that enable their employees to use business applications, collaborate, and serve customers. The primary cause of this increased network dependency is that applications are no longer running on the user’s desktop, but in the cloud. As result, the network is a critical element of the end-user experience; not surprisingly, WAN technologies and network monitoring are going through a major transformation.
To successfully monitor a modern WAN, centralized network monitoring based on SNMP is not enough because it doesn’t verify connectivity from remote locations, and doesn’t generate a 24/7 baseline that can be used to proactively detect performance degradation issues.
Network administrators need to complement traditional SNMP-based tools with active monitoring.
Evolution Of The WAN
From a connectivity perspective, remote offices have evolved from hub-and-spoke topology, mostly based on MPLS or other dedicated circuits, to split-tunnel, where a virtual private network (VPN) tunnel is established over a regular Internet connection from the remote office back to the headquarters. And software-defined WANs (SD-WANs) are an emerging technology just starting to be deployed.
In a hub-and-spoke network, each remote site, or spoke, is connected to a central site—the hub—with a point-to-point connection. All traffic from any spoke to any other spoke or external destination passes through the hub. Hub-and-spoke WANs rely on dedicated circuits (e.g. MPLS) that can assure quality of service for real-time applications and guarantee bandwidth. Not surprisingly, private circuits are more expensive than regular Internet connections.
In this environment, it is important to verify from each spoke site that connectivity is available to other spokes as well as to the hub. It is also important to verify that the service provider is honoring its SLA for bandwidth, quality of service, and uptime.
An active monitoring solution such as NetBeez generates SLA reports and constantly monitors end-to-end connectivity and performance for each site.
In a split tunnel configuration, each remote site is connected with an Internet broadband connection. To reach internal corporate networks, a VPN tunnel is established between the branch router and the VPN concentrator located at the company’s headquarters or data center. Because Internet traffic is routed with best effort, quality of service cannot be assured for real-time traffic (e.g. VoIP). On the other hand, each remote site has direct access to the Internet and can use more bandwidth.
A centralized SNMP server cannot detect connectivity issues at a remote site nor detect application performance degradation that the remote users at that site experience.
Deploying active monitoring agents at remote locations enables network administrators to perform tests on the network and application layer to verify services availability, correct functioning of the ISP’s DNS, and measure application performance.
In an SD-WAN, a branch office router can use multiple lines in parallel to dynamically route traffic over one line or another based on its profile. For example, real-time traffic can be routed through MPLS while applications that consume a lot bandwidth can use an Internet connection. An SD-WAN solution is capable of abstracting the complexity of such a configuration thanks to its integrated intelligence.
SD-WAN vendors generally include passive application performance monitoring in their appliances to identify applications traversing the WAN, prioritize mission-critical data, and then optimize routing decisions. So far, we have seen limited active monitoring capabilities integrated into such appliances. Besides, it’s not their core functionality, and an active performance monitoring solution is still needed to get a comprehensive view of the network.
How NetBeez Implements Active Monitoring
NetBeez is a complete active monitoring solution that uses wired, wireless, or software agents deployed at remote locations. It also supports virtual and external/cloud agents to extend its monitoring capabilities beyond the boundaries of an enterprise network.
Each agent runs tests such as ping, DNS, HTTP, and traceroute; each constantly tests network connectivity to Internet and intranet services and applications, and tests application performance.
All real-time data is sent back to a central server, which acts as both the agents’ controller and analytics processor, generating dynamic performance baselines and alerts when performance degradation occurs. When that happens, the network administrator can use the dashboard for triage and determine if the issue affects only one user, a remote office, the whole WAN, or the entire user base of the application or cloud service.
Beyond Networking And Monitoring
It’s an exciting time for networking and network monitoring. SD-WAN is slowly but surely taking its place in our networks. And that means more automation and intelligence is taking over our networks and more applications are interacting with the IP fabric to deliver real distributed applications.
In this panorama, network monitoring is a critical element of the network infrastructure to automatically tune and optimize its configuration and assure the availability and performance of services and applications.