AWS Cloud Operations Blog
Enhance your global network performance: A deep dive into Internet Monitor’s new optimization tools
Overview
The Internet Monitor feature of Amazon CloudWatch Network Monitoring now includes enhanced traffic optimization recommendation guidance that you can use to explore how to help optimize your application’s latency by using different AWS Regions or Local Zones, or by using Amazon CloudFront. You can also learn how to reduce latency by routing specific IP prefixes to your endpoint. The Internet Monitor suggested updates feature enables you to explore all of these options for improving latency to support a better experience for your application users.
The latency improvements suggestions you’ll find there are the following:
- Amazon Route 53 routing configuration optimization recommendations: This option helps you learn how you can optimize your Amazon Route 53 DNS routing configurations. Internet Monitor uses the mapping relationship between your client IP prefixes and their DNS resolvers, along with latency comparison data in your monitor, to show you client prefixes that you might choose to route directly to your application endpoints, to reduce client access latency.
- Optimal Region/Local Zones selection, to reduce latency: By analyzing the rich latency data available in AWS, this option provides recommendations for the AWS Regions, AWS Local Zones, or Amazon CloudFront selections that can help reduce latency when clients access your application.
- Client location-based latency optimization recommendations: Using this option, you can see latency optimization recommendations for locations (by country and by city-network) where your application has significant traffic, based on traffic volume ranking.
You can access these Internet Monitor optimization strategies in the Amazon CloudWatch console by navigating to Network Monitoring > Internet monitors, and then choosing the Optimize tab.
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-1.png)
Figure 1: Traffic routing optimization recommendations on the Optimize page in the console
In this blog, we describe and provide examples of how you can use these options in the suggestions feature to explore improving latency for your application.
Route 53 routing configuration recommendations
Amazon Route 53 provides an IP-based routing feature that enables customers to steer sets of clients to specific AWS Regions and optimize performance. Properly configuring it can be difficult because 1) It is hard to visualize the before-and-after effects and 2) It is non-intuitive what IP addresses to enter into it. This is because DNS requests traverse through DNS resolvers (usually operated by the client’s ISP) before they reach the Route 53 Authoritative DNS. Thus, Route 53 never sees the actual client IP addresses.
Route 53 makes steering decisions based on the two addresses that it does see: 1) If the DNS Resolver supports it, it will pass through the anonymized subnet of the client to Route 53 (RFC 7871 otherwise known as EDNS-Client-Subnet or ECS) or 2) If ECS is not supported, Route 53 uses the DNS resolver IP to steer requests for clients behind that resolver.
For example, say that a client has an IP address 203.0.113.10, and it uses the ISP DNS resolver at address 198.51.100.50. If the resolver supports ECS, Route 53 will see 203.0.113.0/24 and use that to steer the end user. However, if the resolver does not support ECS, Route 53 will steer the end user based on the IP address of the resolver that the client uses: 198.51.100.50. Most resolvers do not support ECS, so this can be challenging for a Route 53 customer to configure. Specifically, for this example, if you wanted to steer clients in 203.0.113.0/24 to us-east-1, you would need to add a rule steering that subnet’s resolver, 198.51.100.50, to us-east-1. In other words, you would need to know the IP address of all the ISP resolvers that your customers use, which is not intuitive.
The new routing configuration recommendations simplifies this process. Not only does Internet Monitor visualize the performance of your clients to AWS Regions, it now also provides mapping relationships between client IP addresses and resolver IP addresses so you don’t need to dive deep into ISP DNS resolver details. These mappings provide both geolocations and end-to-end network latency visibility for your application’s clients. Using this capability, you can fine tune your DNS routing strategies, directly load the outcome that you desire into Route 53 IP-based routing, then visualize and monitor the outcome.
To get started, on the Optimize tab, view the information under Reduce latency: Suggested updates to Route 53 routing, as shown in the following figure (Figure 2). There, you’ll find comprehensive client data, including the DNS resolver IP addresses used by clients, the corresponding ISP names and ASNs, and the client IP prefixes that are associated with the resolver IP addresses.
You’ll also see suggestions for improvements, and the potential RTT (Round Trip Time) latency improvements that your clients can see if you implement the suggestions.
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-2.png)
Figure 2: Detailed client DNS information and latency optimization suggestions
To implement improvements for specific client IP ranges showing significant potential benefits, start by reviewing the information of each DNS resolver IP address. Then, you can easily create IP-based routing CIDR collections, so that you can implement recommended routing policies. For more information about CIDR collections, see the Route 53 documentation. After you create the CIDR collection, you can create a routing policy to route those client IP prefixes to a specific Region, as recommended in the suggestions table.
Follow these steps to create a new CIDR collection, or to update an existing collection:
- Select the client prefixes that you want to optimize: In the table of suggested updates, select a DNS resolver IP address.
- Choose Add to CIDR collection.
- Enter values for Collection name and Location name.
- Choose Submit.
These steps are illustrated in the following figure (Figure 3).
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-3.png)
Figure 3: Creating CIDR collections for IP-based routing, based on the suggestions table
After you create a CIDR collection, you can view the newly created collection in the Route 53 console: Under IP-based routing, choose CIDR collections, as shown in the following figure (Figure 4).
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-4.jpg)
Figure 4: List of CIDR collections on Route 53 console
Now, you can use the new CIDR collection to route client traffic to specific IP address prefixes, to improve latency. For more information about Route 53 IP-based routing principles and how to configure routing, see the following blog post: Introducing IP-based routing for Amazon Route 53.
By using this integration with Route 53, you can more easily identify and optimize DNS routing for your application. Using the suggestions described here helps you get to a better solution in the following ways:
- You get visibility into real client performance
- You can see actionable routing optimization recommendations
- You can use this information to make data-driven decisions
Region and Local Zones recommendations based on latency
To help you quickly understand the best options for reducing your client latency, review the top Regions and Local Zones suggestions on the Optimize tab, which are automatically generated for your application. In the table with Suggested options for top Regions (by traffic volume), you can see the top recommendations for configuration updates that can improve latency for your clients, as shown in the following figure (Figure 5).
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-5.jpg)
Figure 5: Regions and Local Zones recommendations for optimal latency
You can also explore recommendations for all Regions that serve your application clients, along with Local Zones and CloudFront, with detailed information for each suggestion broken down by specific client locations. To go to a page where you can explore suggestions and compare latency for all your application Regions, choose Optimization suggestions for all Regions.
Now, on the page that opens, you can select different Regions that you could choose to configure, and see the potential latency differences between them. You can also choose to include CloudFront or Local Zones as comparison options. After you make your selections, compare the time to first byte (TTFB) for each option, as shown in the following figure (Figure 6).
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-6.png)
Figure 6: Selection recommendations and TTFB comparisons across all Regions
Location-based recommendations for latency optimization
You can also view recommended configuration changes based on your top locations, by traffic volume. Then, for each location and recommendation comparison, you can view more detailed information that includes the expected average TTFB for the location, as shown in the following figure (Figure 7). You can also expand locations to view granular details, to see specific ASNs that support traffic in each location.
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-7.png)
Figure 7: Latency optimization recommendations by client location
For even more details at the city level, as you choose each location, see the information that’s displayed below the table. Here, you can see detailed information, including ISP names and other ASN information, as well as historical TTFB trend charts, as shown in the following figure (Figure 8).
![](https://d2908q01vomqb2.cloudfront.net/972a67c48192728a34979d9a35164c1295401b71/2025/02/02/Image-8.png)
Figure 8: Expanded view by city with detailed ASN information and historical TTFB trend charts
Summary
The Internet Monitor feature of Amazon CloudWatch Network Monitoring now includes a set of traffic routing optimization recommendations that can help you reduce latency for your application’s clients in the following ways:
- Intelligent performance optimization – Automatic analysis of application traffic to provide optimization recommendations, including AWS Region selection and CloudFront configuration, to help you improve application performance.
- Comprehensive analysis dimensions – Regional analysis based on traffic ranking and client locations enables you to drill down to specific locations at the city and ASN level, so you can make more targeted optimization decisions.
- Flexible configuration comparison – Performance comparison across multiple Regions and time to first byte (TTFB) analysis helps you evaluate the effectiveness of different deployment options.
- Precise decision support – Demonstrated impact of your potential configuration changes, including detailed latency improvement projections, assists you in making optimal deployment decisions.
- Streamlined user experience – An intuitive console interface combines with real-world business scenario examples, to help make your performance optimization process simpler and more efficient.
By leveraging the real performance data and automated recommendations in the Internet Monitor feature of Amazon CloudWatch Network Monitoring, you can significantly improve your application client experience while reducing operational costs.
About the authors
|
|
|
|