My expectations for Werner Vogels’ keynote today were very high, and I have to admit, I was not disappointed.
The keyword for Werner’s presentation was transformation. AWS is transforming the IT industry, not only technology-wise but also structurally. They’ve spent the last 10 years building a broad set of services to choose from, giving us all the building blocks we need. At the same time, AWS is helping its customers transform too, actually turning them into transformers with superpowers.
Werner focused on three parallel areas where transformation is needed:
- Development: This includes development, testing, and operations. This broad area is quickly evolving to help you reduce business risks, build smaller and more targeted apps, deliver changes much faster, and react to customer needs. It is also evolving to allow you to run many more experiments.
- Data: Given that everyone has access to the same computing power and storage capabilities today, data is truly the biggest competitive differentiator of our time.
- Compute: A very heterogeneous space where VMs, Containers, and Serverless Computing coexist.
Development is where agility lives. Developers need unconstrained access to resources. And, they need automated testing with higher fidelity to reduce time to market and achieve major improvements in terms of cost and productivity.
Here is a short list of news related to development:
- AWS OpsWorks for Chef Automate: A fully managed Chef server that is generally available today.
- Amazon EC2 Systems Manager: A collection of AWS tools for packaging, installation, patching, resource configuration, and task automation. Available today.
- AWS CodeBuild: A fully-managed build service for compiling the source code and running unit tests; it automatically scales and you pay only for the minutes of computation. Available today. Of course, it’s well integrated with the AWS CI/CD suite, which includes CodeCommit, CodePipeline, and CodeDeploy.
- AWS X-Ray: A fully managed service for analyzing and debugging distributed apps in production. AWS X-Ray is available in preview today and offers great real-time visualization of complex systems, including status codes, latencies, etc. AWS Lambda support is coming soon as well.
- AWS Personal Health Dashboard: It allows you to design personalized views of your AWS services health. Generally available today.
- AWS Shield: A managed DDoS protection service; it’s generally available today and comes in two tiers: Standard and Advanced. Shield Standard is granted by default to every AWS user and it protects from volumetric and state exhaustion attacks (network layer). Shield Advanced is additional protection against large and more sophisticated attacks at the application layer.
AWS offers a broad set of analytics capabilities, including data processing, warehousing, reporting, real-time processing, and predictive analytics. Such functionalities are supported by a whole range of data-related services, such as EMR, Athena, Redshift, Quicksight, Elasticsearch, Polly, Lex, Rekognition, and Amazon Machine Learning.
Despite such broad capabilities, 80% of what we consider analytics is not analytics at all. Indeed, most of data analytics tasks also involve discovery, indexing, data acquisition, security, storage, access, and governance operations, which contribute to less than 20% of the outcome.
Werner highlighted that Modern Data Architecture is agile, as you need to manage a constantly changing variety of users, sources, models, applications, queries, and processing needs. Sadly, the reality is that most of your data is not agile (i.e. it’s stored in “silos”).
Only three new data-related services were announced today, although they will have a big impact on your data pipelines.
- Amazon Pinpoint: It will help you understand user behavior, define who to engage with, deliver push notifications to mobile devices, and track campaign results.
- AWS Glue: A fully managed data catalog and ELT service (Extract, Load, Transform). It’s integrated with S3, RDS, Redshift, and any JDBC-compatible source. AWS Glue is supposed to cover all of the missing pieces of an agile data architecture so that your workflow will be 80% analytics and only 20% data preparation.
- AWS Batch: A fully managed batch processing service that will solve all the challenges of large-scale batch processing, such as cluster provisioning, batch software install, jobs interdependencies, job queues, jobs scheduling, and cluster scaling. AWS Batch will also perform cost optimizations through EC2 spot fleet and support priority-based queues. It is available in preview.
The current landscape of AWS Compute is where Virtual Machines, Containers, and Serverless have to coexist. Werner analyzed the many dimensions that differentiate such solutions, such as packaging, updates, execution, run time, and unit of cost.
Here are a couple of news items related to the Containers world:
- Task Placement Engine & Event Stream
- Blox: It’s a collection of open source projects for ECS. Today, only cluster-state-service and daemon-scheduler have been published. You can find out more at blox.github.io.
AWS Lambda comes with some great news as well:
- C# Lambda support
- AWS Lambda@Edge: You can now run AWS Lambda functions at CloudFront locations, which allows you to process data at the edge locations to achieve low-latency request/response. It’s only available in preview, but you can apply here.
- AWS Step Functions: Allows you to coordinate components of distributed apps using visual workflows. Indeed, you can already orchestrate multiple Lambda functions and achieve complex behavior such as concatenation, parallel execution, retry policies, long-running tasks, etc.
Further minor serverless updates have been announced during the show, but I will cover them in a dedicated post in a few days.
The Werner Vogels Keynote was a long series of exciting announcements. Although many services are still in preview, we can already start playing with most of them.
Let us know which of the new services are your favorites, and what you liked most at re:Invent 2016.