Even in big events attracting hundreds of thousands of viewers at once, such as the World Cup, we could flexibly expand the infrastructure without spending extra money on hardware, resulting in time and cost savings.
Yong Su Kim Deputy Director, Development Team, SBS I&M

SBS I&M, a Seoul Broadcasting System (SBS) subsidiary, supervises the content distribution business and new media content production and services of the SBS Media Group. Key businesses include the online video service platform business, new media content services and partnership business, and IT infrastructure services.

 

SBS I&M wants to improve the stability and lower the costs of its IT system to support rapid time to market for streaming global events such as the World Cup and the Olympic Games. SBS provides a variety of content services for viewers around the world, attracting about 700 million clicks per month. Although traffic increases only at certain times of the day, the infrastructure has to be prepared to deliver services reliably at all times. It has not yet been possible for SBS to forecast user patterns, creating limitations in preparing the infrastructure for large events. The ability to rapidly deploy new services is vital to a broadcasting company, but SBS's existing IDC took too long to prepare infrastructure, which required system purchase planning, vendor selection, payment, shipping, IDC warehousing, OS installation, network setting, source code deployment, and service launch. Even when it was possible to prepare the infrastructure in advance for a large event, there were problems with managing the infrastructure after the event ended. When SBS realized a more reliable, cost-effective solution was needed, a cloud service seemed the clear choice.

The AWS Cloud presented SBS I&M with a simple but more manageable solution to its challenges. Amazon Elastic Compute Cloud (Amazon EC2), Elastic IP, and Elastic Load Balancing source deployment and service opening were all implemented in just 10 minutes. When SBS I&M was actively considering a cloud service, AWS was the only cloud service provider with a regional office in Seoul. This was key, as it offered SBS I&M the advantage of being able to receive technical support in Korea quickly. SBS I&M also chose AWS for its extensive experience and expertise, as it has been delivering cloud services for quite some time.

SBS I&M began using AWS in April 2016 with a hybrid approach, managing its systems both on-premises and in the cloud. SBS I&M uses DBMS on-premises, in addition to a variety of cloud services, such as Lambda@Edge, Amazon Route 53, Amazon Simple Storage Service (Amazon S3), Amazon EC2, Elastic Load Balancing, and Amazon CloudWatch. For its front-end system, the company uses some of Amazon CloudFront's features, including Lambda@Edge, Amazon Route 53 and Amazon S3, to enhance system availability and facilitate deployment. Caching has also reduced the workload on the back-end system. In particular, Amazon CloudFront is well suited for irregular and unpredictable broadcasting services, as it offers a pay-as-you-go pricing model. On the back-end system, Amazon EC2, Elastic Load Balancing, Amazon Route 53, and Amazon CloudWatch were configured to deliver fast response with an asynchronous, lightweight framework. Non-disruptive deployment and intelligent system management minimize downtime, and accounts were separated from Amazon VPCs to prevent service impact. “Since May 2016, we have gradually migrated SBS’s main videos, SBS programs, the main website and SBS radio to the cloud, and we are planning to complete the migration of all services by the end of 2019,” said Yong Su Kim, deputy director of the development team at SBS I&M.

Architecture diagram configured by SBS I&M

Architecture diagram configured by SBS I&M

In order to provide reliable services through its legacy system, SBS I&M had to prepare for the maximum expected traffic. As a result, on average, only 10% of the system was used at a given time. AWS, on the other hand, presented the best choice for broadcast-related services given the unpredictable user patterns. Yong Su Kim said, “Even in big events such as World Cup soccer matches, which attract hundreds of thousands of viewers, we were able to flexibly expand our infrastructure without having to purchase additional hardware, saving time and money.”

“We’ve also been able to take advantage of AWS and its serverless system and reduce costs by applying Reserved Instances (RI) and Auto Scaling. Previously, the cost of planning a new service had to be planned within a set cost limit, and the cost of failure was significant. But in the cloud, we can create and test services with just a few clicks and pay only as much as we use, so we can flexibly develop any service that we want without fear of failure,” he explained. He added that, in the end, SBS I&M could achieve a higher quality of service for its audience. AWS has worked well for SBS I&M and its hybrid architecture because DNS services like Route 53 allow for redundancy between the cloud and the IDC, making hybrid configuration easy over both VPN and dedicated lines.

After migrating to AWS, SBS On-Air Service recorded the highest number of concurrent users since its foundation during the World Cup 2018. SBS I&M’s "SBS World Cup Special Service" page achieved the highest on-air traffic (303 Gbps) since its foundation for the match between South Korea and Sweden on June 18, 2018. It also recorded the highest number of concurrent on-air users (228,000) on June 27, 2018 during the match between South Korea and Germany. Despite the huge number of concurrent users for each soccer match, AWS services enabled SBS I&M to flexibly scale its infrastructure to deliver reliable service.

Going forward, SBS I&M plans on actively utilizing AWS and the services it continues to launch to develop various AI services, including image analysis and video content recommendation, in addition to utilizing Amazon Alexa.

To learn more, visit aws.amazon.com/media-services.