See what Amazon learned from this year’s huge Prime Day numbers
Most everyone is likely familiar with Black Friday and Cyber Monday, but what about Prime Day? If not, it’s a day on which Amazon Prime members gain access to hundreds of thousands of deals. And because of Amazon’s scale, the company must methodically plan for this annual event to ensure the AWS infrastructure that supports it is up to the task.
This year’s Prime Day set a variety of records, topping those for Amazon on Black Friday and Cyber Monday, which makes it the largest day in the company’s retail history.
Key metrics from Prime Day 2017
Below are a few of the metrics that Amazon shared after their various teams were able to tally the results:
Stack creation – Almost 31,000 stacks were created using AWS CloudFormation for Prime Day, which facilitated bringing additional AWS resources online as needed.
Storage – Amazon Elastic Block Store (EBS) usage grew by 40% compared to last year’s Prime Day, with data transfer coming in at 52 petabytes, which was an increase of 50%, and I/O requests coming in at 835 million, which was a 35% increase.
Database – Instead of using an SQL database, Amazon relies on its own DynamoDB technology to power all of AWS. Requests to Amazon DynamoDB totaled 3.34 trillion during Prime Day.
API Calls – Using Amazon CloudTrail, more than 50 billion events were processed and over 419 billion calls were made to different AWS APIs during Prime Day.
How did Amazon prepare for Prime Day?
This is actually the third Amazon Prime Day, so the Seattle giant has already had a few test runs to ensure they’re ready for this level of demand. But, as is showcased in these numbers, each year continues to be bigger than the last, so it’s always important to properly prepare for such a monumental event.
Two of the best practices used to prepare for this year’s Prime Day included Auditing and GameDay:
Auditing – This is just a straightforward way to identify risks, as well as track preparations and progress towards objectives. Each AWS team is required to respond to both operational and technical questions to gauge their readiness.
GameDay – This process introduces simulated failures, helping team members identify and respond to issues in a practice environment. In addition, failover and recovery mechanisms are tested, which can often expose latency issues that may have not otherwise been seen.
What can your company learn from Prime Day?
The key takeaway from these metrics is the sheer power of AWS, and also the significant amount of planning that is required to pull off an event of this size. Amazon is ‘best in class’ when it comes to cloud knowledge and infrastructure, but that doesn’t mean they can avoid the intense preparation required to ensure they’re prepared for the next Prime Day. As is the case in our own companies, the work is never complete. Next year’s Prime Day will be bigger than the last, and so Amazon’s personnel and infrastructure will again have to prove they can meet the demand.
Is your organization up to the challenge of meeting its required demand? If not, it may be time to look at your IT infrastructure to find areas where gains can be made. At CloudHesive, we take a comprehensive approach to helping our clients leverage the power of the cloud through our partnerships with companies like Amazon. To learn more about what we do and how we can help you reach your goals, feel free to reach out to us today at 800-860-2040 or through our online contact form.