Sign In

Three Keys to Getting Amazon Level Web Resiliency for your Own Prime Day


EVP, Product


The volume of traffic Amazon gets on a normal day is tremendous – estimates put their monthly visitors at more than 200 million, which translates to roughly 6.5 million visits per day and peak hour in the million visitor per hour range. While Amazon was mum on its 2015 Prime Day visitors (and sales), they did concede that last year’s Prime Day sales exceeded 2014 Black Friday.

While even big retail brands only get a fraction of Amazon’s traffic on their best day, there are valuable lessons to be learned from Amazon’s infrastructure that can help any online business mitigate the risks of traffic spikes and provide the best quality digital experience for their customers. After all, whether you aspire for your own “Prime Day” or it happens as a result of an unexpected traffic surge, your infrastructure should be planned to shoulder the load.

Here are three smart steps to improve Web resiliency.

1. Employ Secondary DNS

Amazon has many DNS layers to prevent outages – it’s possible Amazon’s redundancies are so thickly layered that even a billion users couldn’t crash their site. But secondary DNS is helpful for much smaller sites as well to guard against outside threats (DDoS and other attacks), outages of any kind or internal update or service errors that are accidental and unexpected.

Secondary DNS works simultaneously with to your primary DNS. When a DNS query is made, Secondary DNS participates along with primary DNS providing the fastest most reliable answer for every user; in the case of a latency, DDOS attack or outright outage, Secondary DNS will ensure business continuity. With secondary DNS your business will never see the large revenue loss and loss in customer confidence that happens as a result of a large latency or outage.

2. Use Multiple Cloud Providers and CDNs

A brief review of Amazon’s infrastructure indicates they have tens of thousands of geo-located servers, thousands of cloud instances and a complex network of hubs that service their massive online business. Not even Wal-Mart can compete with their scale and international reach (though we’re sure they’re trying).

But even much smaller businesses can receive enormous benefits from using more than one cloud provider or CDN. Given the unpredictable nature of the internet, using just one cloud or CDN leaves your business vulnerable to latencies and outages, which can seriously impact brand reputation. Relying on just one cloud opens your business up to threats, whereas a multi-cloud and CDN configuration provides multiple routing options, the ability to manage traffic based on load and to better budget based on performance characteristics.

3. Adopt Real-Time Monitoring and Optimization Tools

Most businesses monitor their internal assets – data centers, applications, databases and internal networks- but too few have the same insight and control into their cloud and internet-based assets, and specifically, how these internet-based assets look to the user.. If your single cloud is buckling under load, how do you monitor this to know in real-time that there is strain on your network? And if you’re not monitoring to maintain and improve performance, how can you accurately assess the digital experience your customers are receiving, both locally and in different locations around the world?

Real-time tools can help you improve performance, help build a more flexible and resilient infrastructure and guard against performance issues and notify an IT admin if there are anomalies to consider.

You might not be Amazon, but by employing simple solutions to bolster your network, you can markedly improve your business performance online and mitigate against the unpredictability of the public internet.

Scott Hilton is the EVP of Product at Dyn, the world leader in Internet Performance Solutions that delivers traffic management, message management, and performance assurance. Follow him on Twitter at @wscotthilton and @Dyn.