Monitoring the performance of your infrastructure is a worthy and necessary part of any IT environment. Whether you have one office or thousands, knowing the status of your equipment is important. However, many of the applications we depend on today are delivered over networks owned by carriers. Accessing your email, for example, likely routes through a few hops in the LAN before hitting the wider internet and eventually the data center hosting Gmail or Office 365.
Traditional network monitoring tools were based entirely on legacy network infrastructures, where SaaS and cloud services weren’t in the picture. This kind of monitoring relies on collecting information from devices owned by the company looking to monitor. Now, that seems quaint. Traditional monitoring uses SNMP, and it’s blind to the majority of the application delivery path that traffic now takes. That’s because SNMP was born out of a different necessity—managing and configuring large in-house networks of owned devices. It was approved as a standard way back in 1990, before distributed cloud- and SaaS-based services existed in any form.
AppNeta’s approach to monitoring was born out of the need to refocus monitoring on the performance of applications from the perspective of their end users. AppNeta is not a device-centric monitoring company. SNMP has value for troubleshooting, but there are multiple solutions, both free and paid, that solve this problem very well. Here’s how our modern, post-SNMP approach works, and what that means for our users.
1. Active End-to-End Performance Data
End-user experience is paramount in a cloud-based world. Sites no longer go down with any frequency. Slow is the new down. To detect slow applications, network and application monitoring must take a holistic view of performance and isolate performance issues to the LAN, WAN or provider environment to narrow the scope of troubleshooting problems. SNMP data only shows metrics on the devices you own, and still limits metrics to collecting device-centric data. Users traverse any number of devices on the network path that connects them to applications, but LAN hops are typically limited to three or four at most. The real challenge is that you can’t see all of this with SNMP. If you’re using MPLS or a VPN, you may have certain SLAs with your provider, but likely no real visibility. If your users are connecting to SaaS apps, you’ll have no insight behind the vendor’s firewall unless you start using an active testing methodology.
AppNeta uses patented technology to send traffic (typically 30-50 packets) over the network every minute. This allows AppNeta to continuously verify network performance between an end-user location and an application without the high overhead of network flooding techniques. This methodology allows AppNeta to automatically escalate the frequency of testing once an issue is detected in order to confirm any anomaly. Once confirmed, AppNeta will trigger an alert and run a diagnostic test that will look at the network path hop-by-hop to return specific metrics like available capacity, latency, round-trip time (RTT), quality of service (QoS), jitter and loss. This process means that when you’re alerted of an issue, AppNeta is already actively gathering more information. SNMP cannot provide this type of detailed and actionable information due to its passive nature. But for good measure we collect the SNMP state of each device you own at the time of the diagnostic, just in case it is a hardware issue.
2. Application Context for Network Issues
While some tools may attempt to infer application performance from device status, it is no substitute for seeing traffic on the wire. To identify the applications in trouble, our Usage monitoring analyzes all packets on the wire to identify all applications in use and the latency and retransmit rate on a per-user basis. This provides a quick survey of which users are having a poor experience, and with which applications. When deeper analysis is needed, this same architecture provides 100% Wireshark-compatible packet captures to remotely record traffic for secure, centralized analysis.
The best way to proactively measure web application experience is to execute complete synthetic transactions, mimicking the behavior of real users with the latest web browsers. AppNeta Monitoring Points do this concurrently with our patented network performance monitoring, so we can definitively identify if a poor web experience is caused by the network or the application.
3. Troubleshooting With the Right Data
SNMP information is needed to properly manage any enterprise-scale network, but one major challenge is identifying transient performance issues which may come and go in a few seconds, when you are collecting summarized SNMP information every 5 to 15 minutes. AppNeta ensures that you get this info when you need it by collecting SNMP data only when there is an issue. That way, you get the context you need without the need to store SNMP summary data for all devices at all times.
4. Using SNMP Traps
Even in 100% cloud-based companies, AppNeta will be used alongside other best-of-breed technologies for infrastructure monitoring as well as incident management. To support the workflows of enterprise customers, all AppNeta alerts can be sent via SNMP trap, email, and Rest API to a wide variety of systems to integrate with your desired workflow. AppNeta hardware Monitoring Points can even be polled by SNMP device systems to monitor the status of our hardware, ensuring no gaps in visibility.
SNMP still has an important place in IT today. Its type of infrastructure monitoring is certainly still required–but it’s limited. For the kinds of challenges IT is facing today, like seeing into SaaS-based apps and cloud provider networks, AppNeta is a lot more relevant.