Skip to main content Skip to search

Application Performance

Application performance is a challenge, especially at large scale. Often, application performance meets KPI's for most users - say 95% of them - Building a high-performance website or application requires investing in scalable architecture, optimizing static resources, and leveraging distributed infrastructure to serve users near their location.

In this page you will learn about next-generation DNS and application performance:

  • What is application performance?
  • Elements of the page load and how they are typically optimized
  • How DNS can be a silver bullet for performance
  • 5 performance challenges solved by next-gen DNS
  • Which page load elements can DNS accelerate?
  • Measuring the impact of smart DNS traffic routing

Is Next-Generation DNS the Silver Bullet?

There’s a “silver bullet” you can leverage, without major investment, that will immediately improve application performance. That silver bullet is next-generation DNS, and is made much more effective when coupled with Real User Measurement (RUM) data for intelligent traffic management, which can route users to the distributed resource that delivers the best experience using real-time data.


Web 43 compressed

What is NS1?

NS1 delivers best in class DNS response time. But DNS response time is only part of the application performance picture. NS1 also improves on the key elements that impact application performance. It does this by intelligently routing users to the best available resource, taking into account network latency, congestion, server load and capacity and much more.

What is Performance?

Application performance is growing in importance as Internet users become accustomed to near-instant responses from websites and web applications. According to data from KissMetrics, a 1-second delay in page response can result in a 7% reduction in conversion, and 40% of web users abandon a website that takes more than 3 seconds to load.


Website performance refers to the speed by which web users can access content and other resources online. A core metric for website performance is page load time -- the time it takes for a web page to fully load on a user’s browser. Within page load time, an important first segment is Time to First Byte, the time it takes the user to initially connect to a site.

Elements of Page Load Time and How to Optimize Them

According to studies from Google and MachMetrics, average page load time and related metrics are quite similar on mobile and desktop, across multiple industries and geographies.

Average Page Size

1.2 - 2.5 Megabytes

Average Page Load

7 - 12 Seconds

Average Page Resources

70 - 180 per Page

Which Elements Make Up Page Load Time?

The diagram on the right illustrates what happens from the moment a user performs a search or clicks a link, until the content fully loads on their browser..

Application Performance Diagram compressed

How Important is Bandwidth vs. Latency?

A seminal study by Mike Belshe, one of the creators of the SPDY protocol which laid the ground for HTTP/2, showed the comparative impact of bandwidth vs. latency on the performance experienced by end users. Belshe performed two tests: holding latency fixed and increasing bandwidth, then holding bandwidth fixed and decreasing latency. He found that additional bandwidth initially improved page load, but above 5 Mbps, it had diminishing returns. On the other hand, decreasing latency provided linear improvement in page load, from 200 ms down to 20 ms—improving latency eventually improved page load time by 10X!

App Perf Page Load Times Graph compressed

How is Page Load Typically Optimized?

* This rough evaluation of performance impact is based on our experience and the data shown above from Mike Belshe. Of course the actual performance impact of each element will depend on your use case, for example on media-rich websites like Instagram or YouTube, bandwidth is a prime concern.


Silver Bullets in Web Application Performance

Performance engineering is hard. Improving performance for large scale websites and web applications requires careful analysis, architectural changes, scaling servers vertically and horizontally, automated optimization of images and other media, testing and constant monitoring.

Are there silver bullets in web application performance? Something you can easily add to your application that will change the game and improve performance with low effort? We believe the next-generation DNS system can be exactly that.

DNS is the global infrastructure that translates human-readable addresses (like www.examples.com) into machine-readable IPs.

DNS can be an effective tool for improving application performance, increasing the effectiveness of whatever Application Performance Monitoring (APM) and Network Performance Monitoring (NPM) strategies are already in place. Because DNS is already a required element of the application stack, using DNS to improve application performance is effectively “cost free.” It does not require buying and deploying any additional tools or services, but it requires using DNS services that were designed to help you optimize application performance.

DNS doesn’t have to be static, always sending the user to the same location, or bound by rigid geography-based routing rules. In a more advanced DNS setup, you can leverage DNS to distribute traffic between different servers, cloud instances or edge nodes using a chain of logical rules that look at availability, latency, the last resource that served the user, and even real-time Real User Measurement data so that users are routed to the best resource every time.


How Can Basic DNS Routing Improve Performance?

The first step in connecting a user to your web application is the DNS lookup. Hosting your DNS on a fast, global service can significantly reduce connect time for your users.

Many basic DNS services also offer geo-routing. Geo-routing directs users to the application points of presence that are geographically closer to them. This can reduce network latency between users and your application and is generally superior to techniques such as simple round-robin DNS.

In summary, a basic DNS service can improve application performance by:

Reducing application connect time by delivering sub-50ms response to DNS queries

Lowering average latency using geo-routing


How Can Next-Generation DNS Improve Application Performance?

Because DNS is the first point of contact between a user and an application, there is an opportunity to leverage that position, using real time data to deliver an optimal DNS answer.

The best that basic DNS can do is use geographic proximity as a proxy for latency. New methods available in next-generation DNS use actual performance to route users to the best-performing resource based on rules that you specify. Next Generation DNS consistently delivers better DNS answers, optimized for performance.

By leveraging visibility into actual conditions in the network and at the application delivery endpoints (servers, data centers and cloud instances), next-generation DNS can optimize based on network latency, as well as other criteria that affect the end-user experience. These criteria can include network congestion, available bandwidth, server load and capacity, as well as business logic such as relative cost of serving applications from different points of presence.

Why DNS Response Quality is Paramount to Application Performance

NS1 Platform

Infrastructure Aware DNS

Maintains real time view of the network and data center infrastructure to route users to the optimal end point.

Leverages:

Real User Monitoring (RUM)

End-point load and capacity

Cloud capacity

Congestion

Bandwidth

End-point availability

End-user IP prefix and ASN


Filter Chain ™

The filter chain is a drag and drop logic engine that makes it easy to create traffic management rules that meet the needs of your business.

Filter chain enables:

Active failover

Advanced geo-routing

Sticky routing

Weighted load balancing

Real time GSLB

Load shedding

Multi-cloud, multi-CDN optimization

Which Application Performance Elements Can NS1's DNS Accelerate?


The Impact of Smart Traffic Routing: Shaves 20-38% off Round Trip Time

NS1 ran a comprehensive benchmark which compared our own high performance DNS network, but without smart traffic routing, to an identical setup with Pulsar, our smart traffic management technology. What is the impact of smart traffic routing?

The data shows that adding smart traffic management boosts round trip time (RTT) by 25% for the average user. For the 95th percentile, users experiencing the worst performance, the improvement was much more dramatic, decreasing RTT by almost 40%.

Our data measures RTT, not full page load. In a full page load there can be dozens to hundreds of resources and multiple connections between client and server, and the effect of lower RTT is multiplied.


Performance Features

Server Capacity, Load and Availability

NS1 provides an open API, which can receive data about server availability, load, capacity, concurrent connections, connectivity, bandwidth, cost, and more, and use all of these in the routing decision.

Network Latency Optimization

NS1’s anycasted DNS network and advanced traffic steering lets you respond in real time to network conditions, and redirect each user to the server, cloud resource or CDN that will provide the lowest latency—dramatically improving page load times.

Traffic Engineering Filters

NS1’s DNS filters are little programs that run inline for every DNS query and automatically select the best response. Define rules with advanced operations like shedding load, stickiness, shuffle / weighted shuffle, ASN, real user monitoring and server capabilities.

Real User Monitoring

NS1's DNS solutions have more than a dozen filters that can be used to create custom routing rules based on your goals. In addition, NS1 is the only DNS provider that can incorporate Real User Measurement data into the DNS traffic routing rules. NS1’s Pulsar technology can route traffic based on your own network data, or data that NS1 securely gathers billions of measurements a day from real users around the globe using major cloud providers and CDN. NS1 uses this real-time data to direct each user to the optimal destination.

Featured Resources