17 Media: Meeting surging demand for live streaming

About 17 Media

17 Media, a subsidiary of M17 Entertainment Group, runs a live video streaming application and social networking platform. The business invites experts, celebrities, artists, and others to stream video and record shows that viewers can share and interact with in real time. Founded in 2015, 17 Media reached 10 million downloads in just 248 days. The application is popular in Asia, particularly in Taiwan, Japan, and Hong Kong.

Industries: Media & Entertainment, Gaming
Location: Taiwan

Using Google Cloud Platform, 17 Media gained the scalability, availability, and responsiveness to support its fast-growing video streaming services, as well as analyze user data in real time for more informed decision making.

Google Cloud Results

  • Obtains the flexibility to ingest 1TB of data per day for analysis to support decision-making
  • Reduces latency from 300 milliseconds to 200 milliseconds and eventually 50 milliseconds
  • Lowers time to deploy new services from 30 minutes to 3 minutes

Supports move to daily release cycles from weekly or fortnightly cycles

Live video streaming revenues in Asia are growing rapidly. In Taiwan, 17 Media – a subsidiary of M17 Entertainment Group – is at the vanguard of monetizing streaming and content sharing in the region through its application and social networking platform. The business invites users such as experts, celebrities, and artists to stream video that viewers can share, access, and interact with in real time. Founded in 2015, 17 Media reached 10 million downloads in just 248 days. The application is popular in Asia, particularly in Taiwan, Japan, and Hong Kong.

“We’ve grown quickly over the past three years,” says Sammy Lin, Engineering Manager, SRE (Site Reliability Engineer), 17 Media. “We started with 30 employees in 2015 and now have about 700 employees worldwide.”

17 Media emerged from the cloud and operated as a traditional mobile application for two years. As the business grew, a growing desire to diversify its income, minimize the risk of downtime, and reduce latency prompted it to review its infrastructure. Furthermore, the business wanted to reduce a one-to-two-week cycle for software releases to deliver more products and services faster.

“At 17 Media, we extensively use containerization for fast development and testing. With Google Kubernetes Engine, our development cycle experienced an 80 percent increase in overall speed. Google Kubernetes Engine fits our requirements perfectly. Moreover, Google Cloud data processing products such as BigQuery and Cloud Dataflow enable us to better understand our user base, as well as monitor, analyze, and review the quality of our services.”

Kevin Li, Senior Vice President, Product and Engineering, 17 Media

“Our 30-minute deployment times forced us to invest a lot of engineering time on deployment processes, including identifying and resolving errors,” explains Lin. “On some occasions, fixing a single error would require attention from the entire engineering team, delaying the release of new features.”

BigQuery for analysis

17 Media was already using BigQuery to analyze up to 1TB of data per day, having determined in 2015 that no other product could perform a similar function. Its decision to move its infrastructure to Google Cloud Platform was made easy by the quality of its experience with the cloud-based analytics solution.

“BigQuery provided a highly scalable, cost-effective option that enables us to capture and analyze data in real time,” says Randy Huang, Data Lead, 17 Media. “We could see what times users logged in and how long they logged in for, allowing us to determine what content compelled them to stay longer and what content was less interesting.

“This information enables more than 100 business unit heads and operations team members to make informed decisions about products and services, including recommending the best type of streaming for individual users.”

“We moved our services to Google Kubernetes Engine as it allows us to deploy containerized applications, monitor and control our environments from a dedicated console, and scale according to user demand.”

Sammy Lin, Engineering Manager, SRE (Site Reliability Engineer), 17 Media

Reliability and flexibility

17 Media was particularly impressed by the reliability that Google could deliver through its subsea cable network and load balancing. Furthermore, a pricing model based on usage gave the business the flexibility to consume Google Cloud Platform infrastructure as needed rather than be locked in to set usage periods.

The business migrated its infrastructure to Google Cloud Platform earlier this year. The simplicity and intuitive nature of Google products, combined with support from a Google customer engineer, enabled 17 Media to complete the project on time and within budget. “The customer engineer would refer us to the correct expert within Google or to management if we needed an improvement or information about a product roadmap,” says Lin. “The engineer also provided us with recommendations and guidelines for the use of Google Cloud products. This allowed us to better understand those products and their designs.”

Google Kubernetes Engine delivers control and scalability

17 Media now runs its services, as well as data processing and analysis, in an infrastructure comprising a range of Google Cloud Platform products. As well as BigQuery, these include Google Kubernetes Engine, Cloud Storage, Stackdriver, Cloud CDN and Cloud Pub/Sub, with some mutable – or highly changeable – services running in Compute Engine. “We moved our services to Google Kubernetes Engine as it allows us to deploy containerized applications, monitor and control our environments from a dedicated console, and scale according to user demand,” says Lin.

Cloud CDN enables 17 Media to cache load-balanced content close to its users via Google’s edge points of presence, keeping content serving costs and delivery times low. The business uses Stackdriver to monitor and diagnose infrastructure issues, while Cloud Pub/Sub ingests event streams and delivers them to BigQuery for analysis. 17 Media uses Cloud Storage for log backups, website asset storage, and some infrastructure metadata storage, while Cloud Dataflow undertakes data processing and Google Data Studio operates as a dashboard and visualization tool.

“Using Google, we’ve been able to transform our release cycles; we now release daily rather than once every one to two weeks. The decrease in deployment times enables us to revert immediately to previous, stable versions if we experience a problem.”

Sammy Lin, Engineering Manager, SRE (Site Reliability Engineer), 17 Media

In moving from a different cloud provider to Google Cloud Platform, 17 Media deployed its infrastructure in the Google Cloud Platform region in Oregon in the United States. However, the business is progressively moving to the Google Cloud Platform region in Taiwan. After reducing latency from 300 milliseconds to 200 milliseconds by moving from the previous cloud provider, 17 Media expects to cut latency to just 50 milliseconds when it completes the transition to Taiwan.

Time to release new services slashed

With Google Cloud Platform, 17 Media has reduced the time to deploy new services from 30 minutes to just 3 minutes. “Using Google, we’ve been able to transform our release cycles; we now release daily rather than once every one to two weeks,” says Lin. “The decrease in deployment times enables us to revert immediately to previous, stable versions if we experience a problem.

“Our three teams use this hectic release cycle to regularly deliver new features and services that satisfy our customers around the globe.”

“At 17 Media, we extensively use containerization for fast development and testing,” says Kevin Li, Senior Vice President, Product and Engineering, 17 Media. “With Google Kubernetes Engine, our development cycle experienced an 80 percent increase in overall speed. Google Kubernetes Engine fits our requirements perfectly. Moreover, Google Cloud data processing products such as BigQuery and Cloud Dataflow enable us to better understand our user base, as well as monitor, analyze, and review the quality of our services.”

17 Media is now ideally positioned to support continued growth and diversification into new services. “Because the Google Kubernetes Engine deployment model allows us to clearly separate our microservices, we can more create separate branches to develop different features,” explains Lin. “By combining Google Kubernetes Engine with Stackdriver, we can easily work with and analyze our features and services to determine which are most popular.”

“We’re extremely excited by the possibilities that Google is helping open up for us,” Lin concludes.

About 17 Media

17 Media, a subsidiary of M17 Entertainment Group, runs a live video streaming application and social networking platform. The business invites experts, celebrities, artists, and others to stream video and record shows that viewers can share and interact with in real time. Founded in 2015, 17 Media reached 10 million downloads in just 248 days. The application is popular in Asia, particularly in Taiwan, Japan, and Hong Kong.

Industries: Media & Entertainment, Gaming
Location: Taiwan
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE