1. Overview

1.1 Introduction

1.1.1 The part of interview with client

“Last year, due to a failure in AWS Korea, the servers stopped for about an hour.

I suffered a lot of inquiries from customers who couldn't get a ticket at that time.

It was comforting that no large-scale performances have been held since then, so we didn't repair the system.

The problem is that we will be in charge of ticketing for the BTS Korean concert on June 9th.

It doesn't matter how much it costs. Please create a system that can withstand this day.

If the system doesn't get in trouble on this day, huge profits are expected!

 

1.1.2 Requirements

• Ticket sales schedule: June 9, 09:00 ~ 10:00
• The daily access volume is expected to gather between 10,000 and 30,000 people from all over the world.
• Referring to historical data, ticketing is expected to be sold out in about 15 minutes.
• The number of tickets is expected to be 300, and the number of members is expected to be 30,000.
• To be Required performance-oriented services rather than cost. 

 

1.1.3 Customer system status

• Hosting a website in AWS Seoul Region
• Serving your website with one EC2 instance in one Availability Zone (AZ)
• Two MySQL DBs (Master-Slave structure) are in operation
• No system developer and the current system can only be maintained and repaired

 

1.1.4 System building goals

• It can withstand up to 30,000 concurrent users, and the operation should be simple and automated enough to allow the customer to operate it independently.
• Build a system that can maintain high availability even after the end of the project.

 

1.1.5 Benefit

• Avoiding situations such as server outages
• It should be helpful in mission-critical situations, such as when the number of users increases.
• Configured for easy operation by customers through automated systems

 

 

1.1.6 System overview

1. Existing customer systems are not scalable, so they cannot cope with the increase in traffic.

 

 

 

2. The newly introduced automated system can respond flexibly to traffic through Auto Scaling, etc.

 

3. Even if a failure occurs in the Seoul Region, It is possible to deliver the service properly through the failover system prepared in the Virginia Region.

 

1.2 Roles

 

1.3 WBS