AWS re:Invent 2017

4K video has resulted in a huge uptick in resource requirements, which is difficult to scale in a traditional environment. The cloud is perfect to handle problems of this scale. However, many unanswered questions remain around best practices and suitable architectures for dealing with massive, high-quality assets. We define problem cases and discuss practical architectural patterns to handle these challenges by using AWS services such as Amazon EC2 (graphical instances), Amazon EMR, Amazon S3, Amazon S3 Transfer Acceleration, Amazon Glacier, AWS Snowball, and magnetic Amazon EBS volumes. The best practices we discuss can also help architects and engineers dealing with non-video data. Also, Amazon Studios presents how, powered by AWS, they solved many of these problems and can create, manage, and distribute Emmy and Oscar Award-winning content.

AWS Metering provides customers with detailed usage information (down to a specific Amazon EC2 instance or Amazon S3 bucket used in a single hour), enabling them to gain deep insights into their utilization of cloud resources. However, this level of transparency is not available across most customers' traditional IT infrastructure, making it difficult to understand what resources are being used, when, and by whom. Join us in this session to learn how to meter, measure, and understand your usage from AWS, on-premises data centers, containers, serverless compute, even other clouds across your IT infrastructure. We show you how to meter your non-AWS resources to make smarter decisions about your business and investment in the cloud.

WebGL has made great improvements over the past years. However, it still can't provide photorealistic experiences alone. In order to provide products with the best look and feel, we decided to use server-side 3D rendering. In this session, we show you how we built our real-time 3D configurator stack using Amazon EC2 Elastic GPUs, RESTful microservices, Lambda@Edge, Amazon CloudFront and other services.

When customers across the globe place orders on Amazon.com, those orders are processed through many different backend systems, including Herd, a workflow-orchestration engine developed by the Amazon eCommerce Foundation team. A mission-critical system used by more than 300 Amazon engineering teams, Herd executes over four billion workflows every day. Beginning in 2013, Herd's workflow traffic was doubling year-over-year, and scaling its then dozens of horizontally-partitioned Oracle databases was becoming a nightmare, and this number kept increasing. To support Herd's increasing scaling needs, and to provide a better customer experience, the Herd team had to re-architect its storage system and move its primary data storage from Oracle to Amazon DynamoDB. In this session, we discuss how we moved from Oracle to Amazon DynamoDB, walk through the biggest challenges we faced and how we overcame them, and share the lessons we learned along the way.

SaaS presents developers with a unique blend of architectural challenges. While the concepts of multi-tenancy are straightforward, the reality of making all the moving parts work together can be daunting. In this session, we move beyond the conceptual bits of SaaS and look under the hood of an SaaS application. Our goal is to examine the fundamentals of identity, data partitioning, and tenant isolation through the lens of a working solution and to highlight the challenges and strategies associated with building a next generation SaaS application on AWS. We look at the full lifecycle of registering new tenants, applying security policies to prevent cross-tenant access, and leveraging tenant profiles to effectively distribute and partition tenant data. We intend to connect many of the conceptual dots of an SaaS implementation, highlighting the tradeoffs and considerations that can shape your approach to SaaS architecture.

Real-time bidding applications are designed for very high scale and performance. A typical RTB deployment needs to be designed to handle at least a million queries per second with TP99 query processing latency of 25 ms. In this session, we feature Bidder-as-a-Service™ by Beeswax and discover how AWS enables their core technology. We will begin by examining the end to end architecture of a real-time bidder application on AWS. Next, we will talk about the challenges and best practices for implementing a durable and high-performing system. Finally, we will conclude the talk with some recommendations on minimizing infrastructure cost while operating a RTB platform at a very large scale.

In this session, you'll learn how AdTech companies use AWS services like Glue, Athena, Quicksight, and EMR to analyze your Google DoubleClick Campaign Manager data at scale without the burden of infrastructure or worries about server maintenance.  We'll live-process a click stream so you can see how Machine Learning can help maximize your revenue by finding the most optimal path of a campaign and we'll look at a real world demo from A9's Advertising Science Team of how they use the data to build Look-alike Model in their projects.

From CloudFront to ElastiCache to DynamoDB Accelerator (DAX), this is your one-stop shop for learning how to apply caching methods to your AdTech workload: What data to cache and why? What are common side effects and pitfalls when caching? What is negative caching and how can it help you maximize your cache hit rate? How to use DynamoDB Accelerator in practice? How can you ensure that data always stays current in your cache? These and many more topics will be discussed in depth during this talk and we'll share lessons learned from Team Internet, the leading provider in domain monetization.

Interested in learning how to integrate the Internet of Things into your advertising platform and combine it with AWS Greengrass, AWS Lambda, Amazon DynamoDB, and Amazon API Gateway to send context-aware advertisements to users at the point of buying? In this session, Mobiquity, the leader in digital engagements servicing the world's top brands, and their Innovation Partner Flomio discuss how they've been able to use AWS to create compelling digital experiences for their clients. We deep-dive on the technology behind Mobiquity's innovative shopping system that uses RFID, Bluetooth, captive Wifi, and a mobile app to provide real-time context for understanding how and where your customers interact with your products and services, allowing you to better tailor your ads to their particular preferences.

Join us for an overview and demonstration of Amazon Connect, a self-service, cloud-based contact center based on the same technology used by Amazon customer service associates worldwide to power millions of conversations. The self-service graphical interface in Amazon Connect makes it easy to design contact flows for self and assisted call-handling experiences, manage agents, and track performance metrics – no specialized skills required. In this session, you will hear from Capital One and T-Mobile on how they are using Amazon Connect to provide their customers with dynamic, natural, and personalized experiences.  See how quickly you can get started with Amazon Connect and build your contact center.