Viewing the ‘Highway’ of Network Performance Management
We are often asked how our technologies very from traditional network management tools. While most other tools rely on simple network management protocol (SNMP), we utilize our own technology to perform ‘active’ network monitoring infrastructure.
Around here, we compare SNMP to a car’s dashboard. You are able to see how much gas is in the tank, what the engine RPMs are, and how hot the engine is running. But if you’re in a hurry to catch a flight out of Logan Airport, what you really want to know about is the traffic on the highway. How long will it take me to get through Boston? How many other cars are on the road? How many lanes does the Massachusetts Turnpike have? Is there a detour because of the Big Dig? You need to be able to see roadblocks that hinder performance (and your timely departure!) on the way to your application’s destination.
SNMP is configurable on most routers, managed switches and other network devices. The metrics gathered using this protocol are unique, and critical to network performance: You can see utilization of memory or CPU, and the health of that devices hardware components. This is all very valuable information that no network engineer should be without, but even a network that is fully instrumented with SNMP cannot answer the fundamental question of how well your network is doing its job, connecting users with information. What happens when the packets leave that device; traverse the network; go through another device, and so on? Do the application packets, traveling from the service inception point to the service consumption point, face any problems reaching their destination on time?
With all of the packet shaping technologies that are enacted across WAN connections, there is a risk that the performance of a network application can be hindered unintentionally. Traffic shaping and QoS is often set in a vacuum, with settings determined by the intuition of the engineer who may not be aware of all critical applications that are using a WAN connection. To really understand what circumstances these application packets face while traversing the WAN, you need visibility through each and every device the packets encounter to see the cumulative effect of the packet shaping.
PathView Cloud sends ‘network analysis’ packets from the microAppliance to the service consumption end point. On the way to the destination, any deviation or interruption in the route, such as a point of congestion or point of data loss is revealed and reported back to the cloud reporting server. All of these metrics are delivered from each hop in the network path, even from devices you don’t manage. This process fundamentally differs from contemporary SNMP monitoring; Instead of having a management information base (MIB) telling us what’s going on, PathView Cloud is able to find out on its own.
At the end of the day, the network admin needs to weigh the importance between status of devices and the status of the application path. When it comes down to it, there is not a right or wrong answer, both are needed. But one thing that is true, end users don’t care if it is the device, the application, or even themselves that is causing application performance degradation, they just want their applications to work!