Amazon CTO Werner Vogels Debuts 4 New AWS Tools At NY Summit

The offerings include a serverless event bus, an open-source software development framework to model and provision cloud application resources using familiar programming languages, the general availability of the AWS Toolkit for Visual Studio Code and the preview of managed spot training for Amazon SageMaker

Amazon Web Services rolled out several new tools at its AWS Summit New York, including a serverless event bus and an open-source software development framework to model and provision cloud application resources using familiar programming languages.

Amazon chief technology officer Werner Vogels introduced the new offerings, which also include the general availability of the AWS Toolkit for Visual Studio Code and the preview of managed spot training for Amazon SageMaker, AWS’ fully managed cloud machine learning platform.

CRN sat down with Aaron Kao, senior manager of AWS product marketing and Deepak Singh, director of computer services at AWS, for a deeper dive into some of the developments.

Amazon EventBridge

Amazon EventBridge is a serverless event bus for ingesting and processing data across AWS services and software-as-a-service (SaaS) applications.

“All these integration pieces -- people have written a lot of custom code for that,” Werner said at the summit. “And the goal for us was really to provide a much simpler programming model, where you can stitch both your code, AWS code, AWS services and services from third parties. It makes it easier for integration with all these different services, so the event bus is a very powerful programming model.”

Customers have been really interested in event-driven architectures and have been using Amazon CloudWatch Events -- which delivers a near real-time stream of system events that describe changes in AWS resources -- for information technology automation use cases, according to Kao.

“One of the things that customers told us is we've increasingly been building SaaS applications and trying to integrate it with our applications as well, so it would be great to have those resources and all of those events put in the CloudWatch events,” Kao said. “Up until now, a lot of what customers have been doing is they build essentially software and run infrastructure to pull the APIs of the SaaS vendors, and then bring it down and integrate with their applications. But the problem with that is it's hard to scale your fleets. You have to manage capacity, you have to manage the software. A lot of what we call ‘undifferentiated heavy lifting’ happens from that. With the launch of EventBridge, we've essentially taken CloudWatch events, then added 10 SaaS vendors to it and broke it out as a separate service. This is super good for customers that want to build event-driven applications and be able to use their SaaS vendor events in it.”

EventBridge supports 10 SaaS application partners and more than 90 AWS services as event sources.

Launch integration partners are Zendesk, PagerDuty, Datadog, Whispir, Saviynt, Segment, SignalFx, SugarCRM, OneLogin and Symantec, and many more are in the pipeline to be added to the event environment, Werner said.

“There's a whole new class of applications that you're going to be able to build with EventBridge,” Kao said. “It's all serverless, so there's no management that you have to do there. We're super excited to see what customers will do with it.”

Zendesk essentially provides a customer service support software ticketing system. “If a new customer ticket comes in, you can send it from Zendesk into EventBridge, and EventBridge can route it to something like SageMaker, where you can do a sentiment analysis, then tag it back to the ticketing system,” Kao said.

PagerDuty offers an incident-response notification platform.

“What you can do with PagerDuty is you essentially can send all these incidents events into something like Redshift and be able to do analysis later on what's the remediation velocity or how much time do my engineers spend on operational issues,” Kao said.

Communications workflow platform Whispir worked with AWS to increase its internet of things functionality. With EventBridge, Whispir users now can easily implement communication workflows that trigger a series of responses based on real-time changes in geolocation data, Whispir said.

“Whispir customers can automate communication responses to triggers – if this happens, then send this – without needing to write connection code,” Whispir chief technology officer Jonathan Swift said in a statement. “Access to the full portfolio of Whispir’s event types is available for users to browse and subscribe using an interactive interface.”

Amazon EventBridge is generally available in these AWS regions: US East (Ohio and N. Virginia), US West (Oregon and N. California), Canada (Central), EU (Stockholm, Paris, Ireland, Frankfurt, and London), Asia Pacific (Mumbai, Tokyo, Hong Kong, Seoul, Singapore, and Sydney) and South America (Sao Paulo). Support for AWS regions in China and the Asia Pacific (Osaka) local region is coming.

AWS Cloud Development Kit

The new AWS Cloud Development Kit (CDK) is an open-source software development framework to model and provision cloud application resources using familiar programming languages.

“That really makes it easier for you to write in high-level constructs in an environment that you're familiar with, to really describe how your infrastructure should look,” Werner said. “This is sort of the ideal destination of infrastructure as code, because it is really code that you're writing.”

Four programming languages will be supported: TypeScript and Python are available now, and .NET and Java are in preview.

“Increasingly in the cloud, we've been seeing customers build their applications more in a distributed manner and then using a lot of different services to support it,” Kao said. “So provisioning that infrastructure has become hard for our customers. They have to do a lot of manual work, or they have to do a bunch of scripts, or they have to learn a domain-specific language of some sort. (With CDK), you can just use the programming language that you're programming with for your code for modeling your infrastructure as well. So there's no context switching that will happen. You can use the exact same tooling that you use.”

“The other thing that CDK has is it's more of an abstracted-level language,” Kao explained. “In these things called constructs -- they're just high-level abstractions -- we've pre-built and pre-filled out and configured parameters that normally would be just a lot of work and heavy lifting. If you don't like those, you can actually just sub it in with your own parameters and customize it however you want. With CDK, you're going to be able to get started a lot quicker...because writing a few lines of code, you're able to do something that normally would require a lot more.”

To build a container application, one line of code sets up a virtual private cloud, a second line deploys an Amazon Elastic Container cluster and, with a third line, a customer is able to spin up an elastic load balancer, attach a Fargate service and then plumb it with CloudWatch.

“That was normally 700 lines…and that was done only in a few lines,” Kao said.

Presenting best practices on top of standard AWS infrastructure is why Singh believes CDK can make it much easier for customers to get started, because they’re not thinking through the best way to run a service behind a load balancer.

“You have a better-established pattern that does it for you,” Singh said. “You can just get going and then, over time, you can tweak it. These are patterns that we can provide, but because it's an open-source project, customers can provide patterns as well.”

“People can say this is the pattern that we're going to use to deploy an application of this type, and you can then just pack it up in a library, and then somebody can reference it,” Singh said. “You don't have to get stuck in configuration code, which most people can't parse through very easily.”

Customers also can bake their security, compliance and governance into their own construct and distribute it to their company, according to Kao.

“One of the key things of computer programming is you just write the code once, and then you don't have to write it again, and everyone can share that,” he said. “With CDK, customers are going to be able to have higher development velocity and also just be able to have the expressive-power programming languages.”

AWS Toolkit for Visual Studio Code

The AWS Toolkit for Visual Studio Code, which provides an integrated experience for developing serverless applications, now is generally available.

The open-source plug-in, previously in developer preview, was designed to make it easier to create, step-through debug, build and deploy applications, according to AWS. The toolkit supports .NET, node.js and Python applications and serverless applications. Other AWS services are planned for the future. The toolkit has built-in project templates that leverage the AWS Serverless Application Model to define and configure resources.

“One of our goals is we want to meet customers where they are, especially in their tooling,” Kao said. “It has the SDKs (software development kits) and all those things built in and allows you to build service applications faster.

Managed Spot Training for Amazon SageMaker

AWS is now previewing managed spot training for its fully managed Amazon SageMaker cloud machine learning platform.

“You can use spot instances to run your training over, which basically means a reduction up to 80 percent in cost of your training,” Werner said. “And all of this capacity is fully managed by SageMaker. You can pick any of the algorithms that you normally would use, whether those are built-in algorithms, whether it's the custom algorithms that your data scientists have created for you or one of the 150 algorithms that currently sits in the SageMaker marketplace.”