Prime Day, Amazon's summertime sale extravaganza, is nearly here. While we, and shoppers around the globe, are looking forward to the extreme sales and killer deals that this 36-hour event will bring, we're all holding our breath and hoping that we don't see a repeat of last year.
During that event, the extreme demand crashed the Amazon site, leaving thousands of us out in the virtual cold while we waited for the retail giant to fix the problem so we could continue shopping.
What does Amazon need to do to prevent this kind of data disaster from happening again?
2018's Prime Day started like any other since the shopping extravaganza started in 2015 — Amazon clicked the button that released hundreds of sale items into the internet void for savvy shoppers to snap up before they sold out.
It is easily one of the biggest shopping days of the year, rivaling both Black Friday and Cyber Monday. Amazon actually started the event to celebrate the company's 20th anniversary that year, offering sales available only to Prime members. The retail giant is on track to sell more than a million products in 36 hours.
Until disaster strikes, or what passes for disaster for an online sale — the site starts glitching. People can't access the website, and some even lose the ability to ask Alexa for things on enabled devices. Even Twitch, the streaming platform which the company purchased in 2014, started experiencing some problems.
Amazon reportedly didn't have enough servers to handle the traffic that swamped the site during 2018's Prime Day sales. During the course of the 36-hour sale, they had to manually add servers to the website to try and keep up with the demand.
While Amazon was able to get ahead of the problem and prevent it from affecting the sale too dramatically, there were plenty of people furious with the company for promising sales and delivering crashes.
This single event serves to emphasize the importance of reliable data infrastructure, especially with companies like Amazon which rely on Enterprise Resource Planning (ERP) software.
Not only did Amazon not have enough servers to support Prime Day 2018, but the lack of server support also caused cascade failures throughout the system. During the first 15 minutes of the sale, Amazon also shut down all international traffic to reduce the load on the servers.
Cloud resources, like those that Amazon provides and utilizes, are quickly becoming a staple in nearly every industry. ERP software needs to focus on three key points — infrastructure as a service, cybersecurity, and availability. With all three of these factors, the data infrastructure of any company may be doomed to failure.
Giants like Amazon don't rely on on-premise software to support their servers. Their site is entirely based in the cloud, which isn't as customizable but generally provides more stability. It's also why they were able to add new servers on the fly during the Prime Day 2018 crash.
Amazon provides infrastructure as a service to a variety of different companies around the globe, but during 2018, they didn't take the time to bolster their own servers, even though they anticipated billions of dollars in sales that year.
With Prime Day 2019 coming up on July 15th, what can Amazon do to prevent another debacle like last year? The retail giant needs to start by analyzing data from past events — how many sales were they expecting, and how many did they actually get.
With four years of data to pull from, the site's data analysts may be able to determine with relative accuracy the kind of server support that they'll need to keep the site up and running during Prime Day 2019.
IT teams with the ability to add new servers should also be on hand when the sales go live, to ensure that if there is a problem, they have someone there who will be able to fix it immediately. Prime Day is one of the biggest shopping days of the year, and the debacle during the beginning of 2018's event was embarrassing for the retail giant.
They say that proper preparation prevents poor performance, but in Amazon's case, the company needs to spend some time reinforcing its entire data infrastructure to ensure that something like 2018's Prime Day doesn't happen again.
Kayla Matthews writes about data centers and big data for several industry publications, including The Data Center Journal, Data Center Frontier and insideBIGDATA. To read more posts from Kayla, you can follower her personal tech blog at ProductivityBytes.com.