http://aws.amazon.com/solutions/case-studies/spotify/ About Spotify Spotify is an online music service offering instant access to over 16 million licensed songs. The application is available for Windows and Mac OS operating systems, as well as multiple mobile device platforms. Spotify users can choose between a free, ad-supported version and a monthly subscription that allows users to enjoy music without ads on a wide range of mobile devices. Both options allow users to enjoy instant access to one of the world’s biggest music libraries, and to enjoy, discover and share music on a number of different social platforms, including Facebook. The Challenge Because the company’s goal is to help people listen to whatever music they want, whenever they want, wherever they want, Spotify faces the perpetual challenge of cataloging not only yesterday and today’s popular tracks, but also all those to be released in the future. Spotify adds over 20,000 tracks a day to its catalogue. Emil Fredriksson, operations director for Spotify, explains, “Spotify needed a storage solution that could scale very quickly without incurring long lead times for upgrades. This led us to cloud storage, and in that market, Amazon Simple Storage Service (Amazon S3) is the most mature large-scale product. Amazon S3 gives us confidence in our ability to expand storage quickly while also providing high data durability.” Why Amazon Web Services The company created Python-based backend systems to interact with its huge volume of content in Amazon S3. In addition, Amazon CloudFront delivers the Spotify application and software updates to users. Just as music trends perpetually change, Amazon Web Services (AWS) helps Spotify continuously evaluate its infrastructure in order to meet evolving business goals. Fredriksson notes, “By removing the restrictions incurred by in-house solutions, we enabled much faster development and deployment cycles.” As part of this evaluation, the company frequently examines the possibility of integrating new services into its existing AWS feature stack. Based on this practice, Fredriksson reminds other technical decision makers that resource utilization can be unpredictable. He explains, “Consequently it is very important to design your systems so that they can react to variations in performance and compensate with scaling.” The Benefits Spotify is taking its own advice to maintain a highly responsive system. While establishing new storage previously required several months of preparation, it can now be obtained instantly through AWS. While the company cannot always predict the next overnight music sensation, its infrastructure can spontaneously adjust to any alterations in user demand. Fredriksson, says, “The ability to go from a system architecture design and capacity requirements to an online and working production system in very little time is fantastic. We are very much aware of the work and lengthy preparation involved in provisioning capacity, whether it is storage, servers, or networks. Therefore, we understand how valuable it is to be able to reduce all of that to an AWS API call.”
a single frame can take up to 40 hours to render Not limited to the physical infrastructure, can expand as much as needed to meet business demands (preview the scene, change sequence, etc.) Helped them to do what they couldn’t do by their own as a company. Didn’t wanted to be in the datacenter business neither, wanted to be in the visual effects business. Can shutdown a project as easily too.
Channel 4 can analyze and model in-session data to deliver highly targeted ads to viewers – before a program ends To get closer to growing video-on-demand (VOD) audiences and match them with advertisers, Channel 4 chose a cloud-based solution to help make sense of and monetize its unprecedented volumes of platform data. Launched in 1982, Channel 4 is now established as one of the UK&apos;s premier public service TV channels. But changing consumer lifestyles and technology advances have inevitably meant that Channel 4 viewers are consuming content in new and diverse ways. The explosion of web-connected devices, such as tablets, mobiles, consoles, and Internet-enabled TVs, has accelerated the changes. Viewing audiences of 46 million now regularly access Channel 4 broadcast TV, as well as the company&apos;s ever-growing on-demand services, such as 4oD catch-up TV, Film4, and bespoke games and social channels. But business success delivers new challenges. The volume of largely unstructured web data flowing from Channel 4&apos;s proliferating video on demand (VOD) platforms grew faster than its existing business intelligence systems were designed to manage. &quot;With Amazon EMR we can analyze 100% of the data, not just a sample. Traditional analytics can&apos;t do that.&quot; - Sanjeevan Bala, Head of Data Planning & Analytics, Channel 4 So, with a business and brand mission to connect more personally with viewers and more closely with advertisers, Channel 4 identified the need for smarter, more agile decision-making in its commissioning, scheduling, and monetizing of content. Solving the business problem To realize its business objectives, Channel 4 needed a proven high-performance data-analytics solution – a flexible, integrated service capable of capturing, storing, indexing, searching, mapping, analyzing and matching high volumes of viewer and platform data on demand. They chose Amazon Elastic MapReduce (EMR) from Amazon Web Services (AWS). Amazon EMR uses Amazon EC2 instances, which feature Intel® Xeon® E5 family processors. The collaboration of AWS and Intel brings together the availability, flexibility and virtually unlimited capacity of a cloud-based solution, with the proven power to process huge volumes of data. During the Channel 4 project research phase, its teams had investigated the feasibility of using their existing data-analytics software to perform key tasks. They found it would take eight months to deliver the relevant base-data for analysts to access. By contrast, the AWS cloud-based system meant they could begin analyzing and modeling the data in just two-and-a-half days. Getting to know viewers as never before Channel 4 is now able to bring its registered VOD viewers closer than ever to both content and advertisers, with benefits for both parties. It connects more closely with viewers by better understanding their preferences and behavior, analyzing repeat viewers to schedule more relevant content and to commission enticing new content. It can even serve highly targeted ads in-session, before a viewer&apos;s program ends. Its modeling systems can even predict the socioeconomic class, age and gender of anonymous or unregistered viewers from content viewed, giving advertisers unprecedented confidence in optimizing media placements. Vast data capabilities for a modest outlay Using the AWS cloud allowed Channel 4 to avoid the need to invest in massive server infrastructure, disks, and CPUs. Instead, Channel 4 analysts can now instantly provision as much or as little capacity as they need from Amazon Elastic MapReduce, a service that enables the business to perform complex business intelligence tasks such as predictive modeling with maximum speed, cost-efficiency, and return on investment. Channel 4 has also developed a Big Data control panel (BDCP), a web-based interface that allows analysts to spin up and spin down Amazon EMR clusters, submit queries through Apache Hive and Pig, monitor job status, and extract sample or actual data to run into modeling applications. BDCP puts convenience, control, power, and speed right in the hands of the people who need it, while providing a simple solution to a complex business problem. Opening up the future AWS has helped Channel 4 build ad revenues, enhance its marketing of content, and use detailed, dynamic data to open up exciting new business opportunities. Best of all, Channel 4 can have a direct relationship with its nine million registered viewers for the first time, even modeling in-session data to deliver relevant ads before a program ends – an enviable advantage for any broadcaster. Learn how your organization can benefit from the simplicity and cost-efficiency of high-performance computing on demand.
AWS Dublin Briefing - Cool AWS Use Cases
Ian Massingham – Technical Evangelist
• SPLITS DATA INTO PIECES
• LETS PROCESSING OCCUR
• GATHERS THE RESULTS
CASE STUDY:CASE STUDY:
"WITH AMAZON EMR WE CAN ANALYZE 100% OF THE
"WITH AMAZON EMR WE CAN ANALYZE 100% OF THE
NOT JUST A SAMPLE"NOT JUST A SAMPLE"
- Sanjeevan Bala, Head of Data Planning & Analytics, Channel 4- Sanjeevan Bala, Head of Data Planning & Analytics, Channel 4
WANT TO KNOW MORE?
Join us at AWS Summit London
on 30 April and learn about the
AWS platform, new services and
how to get started.
Register for a Summit near you.