Four tips to ensure optimal SD-WAN performance
While the adoption of SD-WAN brings significant cost-savings and flexibility to the enterprise, it is also a significant operational transformation—one that requires greater attention to performance monitoring. The good news is that active monitoring solutions give network operators the intelligence they need to ensure network performance—before, during, and after SD-WAN adoption.
SD-WAN ADOPTION BRINGS NEW COMPLEXITY
When you switch from using fully-managed MPLS services to ISP circuits, you also give up your provider’s network monitoring and troubleshooting services. Unless you acquire SD-WAN from a managed services provider, you will need to take on the responsibility for troubleshooting network performance issues. This includes making sure the solution can deliver the experience your users expect, as well as continually monitoring performance to avoid major disruptions.
With software-defined networking your infrastructure is software based, which enables new options for speeding up delivery of application and services. One example is the combination or ‘bonding’ of different transport options to reduce overall latency and deliver applications more quickly to the end-user. Because the transport technologies and the bonding are all dictated and handled by software, it will not obvious to operations center staff where to look for the root cause of performance issues.
Another example of complexity is when administrators take advantage of software flexibility to specify different delivery priorities for traffic from different applications. Users benefit when their most critical applications have precedence over other services, but the resulting variability in traffic management makes is harder for NOC engineers to rollout fixes and upgrades across the enterprise.
In both cases, being able to break overall performance down into hop-by-hop components makes it easier to identify problems and validate a resolution. An active monitoring platform is able to do just that. Here are four specific ways an active monitoring platform, like Ixia Hawkeye can help.
ESTABLISH A BASELINE
As with any big network transformation, you want to maintain and possibly improve your users’ experience over time. To do this, you first need to establish a performance baseline using your existing MPLS technology. A monitoring platform can help you define and measure key performance indicators such as response time, packet loss, jitter, and mean opinion score for voice services. Development engineers can then use these baseline metrics to evaluate potential vendors and technology options in the design phase.
RUN PRE-DEPLOYMENT TESTS
Once a final design is selected, you need to live test your circuits before rollout. An active monitoring platform lets you generate test traffic that is very similar to the actual traffic that flows through your network. Field testing helps you uncover any issues that arise during live operation of your architecture. Results that meet or exceed your baseline reduce the risk that you will waste time and resources deploying an inadequate solution.
MONITOR POST-DEPLOYMENT OPERATIONS CONTINUALLY
Once your SD-WAN architecture is in production, you want to make sure user experience does not degrade. You can use an active monitoring platform to proactively inject traffic into the network to verify all routes. You can program your monitoring platform to continuously and automatically record key performance indicators on a daily (or more frequent) basis. The platform’s dashboard displays results over time and makes it easy for you to spot trends and anomalies.
Continuous monitoring is particularly useful for monitoring the performance of software-as-a-service that is accessed and used over the internet. More companies are choosing SaaS for business-critical applications and need a way to validate the quality of service their providers are actually delivering.
MONITOR THE EDGE TO ASSESS USER EXPERIENCE
Once SDN is deployed across your network, you will also want to monitor performance from the perspective of end users. Traditional WAN performance indicators may be sufficient for network operations, but they can be blind when it comes to end user quality-of-experience. Without end-to-end visibility and end user application metrics, it will be difficult for you to correlate events and identify problems.
If you are contemplating your own move to SD-WAN, read about the best practices for performance monitoring in the white paper “Ensure optimal SD-WAN Network Performance and Service QoE.” You can also listen to two recent Ixia webinars: “SD-WAN: The next phase of SDN” and “SD-WAN: Achieving a high-quality user experience with Ixia Hawkeye."