Planning a Cloud Migration with Digital Experience Monitoring Pt. II

When it comes to the cloud, the network makes all the difference.

In our previous installment of this blog series, we covered the whole planning process for migrating your applications, infrastructure, and services to the cloud while ensuring that you’re able to deliver the same consistent service levels your users are used to.

Now that you have a sense of what questions to ask and what metrics to measure, let’s take a lot at how those metrics are measured. When workloads move to the cloud, the network that sits between your end users and the cloud becomes of paramount importance to the service quality you can expect. Slow internet performance means slow response times for your cloud applications.

So how do you monitor how fast your network is? For starters, you need to test application performance across all protocols the application may use to get from the cloud to your users. This includes not just the internet workhorse TCP, but lesser-known protocols like User Datagram Protocol (UDP) and Real-time Transport Protocol (RTP), which reduce latency, making them ideal for applications that require real-time delivery like video, voice, and gaming.

Catchpoint not only monitors the response times of applications from the end-user perspective but also measures latency and identifies bottlenecks from the end-user perspective by supporting the following synthetic monitoring test types:

• Ping – over UDP, TCP, and Internet Control Message Protocol (ICMP). Detects packet loss, network latency, and connectivity failures.
• Traceroute – Border Gateway Protocol (BGP) routes over UDP, TCP, and ICMP. Detects network misconfigurations, peering changes, and route hijackings.
• TCP – measures the time it takes to establish a TCP session. Detects connection failures and poor response times.

Cloud service providers do make the effort to guarantee uptime of their services. One of the most important tools they use to do this is Anycast, a network addressing and routing method that allows multiple cloud hosts or servers to share and announce the same IP address. With Anycast, when a request for a service is made by users in India, London or the US, the same IP address will be advertised for the cloud node, regardless of where that node is located.

The routing infrastructure directs the packets to the nearest instance of the cloud service based on the BGP paths. In this way, Anycast can improve the reliability, load balancing, performance, and availability of cloud services. The biggest drawback to using Anycast is that managing route announcements becomes much more complex. This requires quick detection of any black holes during a failure and identification of customer-impacting events.

This is where Catchpoint’s DNS monitoring comes in. Catchpoint’s various DNS test types and global network of nodes are designed to proactively detect any performance issues which are caused by the Anycast routing mechanism before it impacts the end user.

Staying on top of all the network issues that can impact cloud service delivery can help guide you to a successful cloud migration.

To learn more, stay tuned for the next installment of this blog series.