Patrick Chanezon

Announcing DockerCon 2018 Session: Cool Hacks

Patrick Chanezon

One of the most anticipated sessions at DockerCon is Cool Hacks, where we showcase a few members of the Docker community pushing the envelope on what you can achieve with Docker, in a demo heavy session, showing trends of what innovators are building on top of the Docker platform. This year, we’ll talk about Space, AI and Serverless!

Past Cool Hacks have gone to be widely used: last year Marcos Nils and Jonathan Leibiusky showed Play with Docker, a Docker playground that you can run in your browser that is now used by tens of thousands of developers and system administrators monthly to learn the basics on Docker and was applied to learning Kubernetes with Play with Kubernetes; And Alex Ellis demoed a FaaS, a portable serverless platform running on top of Swarm, that grew into the OpenFaaS project, one of the 12 installable serverless platforms mentioned in the Cloud Native Foundation Serverless Working Group serverless landscape.

This post should whet your appetite for what to expect in Dockercon 2018 Cool Hacks session.

Docker for Space, software devops in a hardware world, and how we build software to hit an asteroid

Christopher Heistand, Flight Software Lead at Johns Hopkins University Applied Physics Laboratory for the DART Mission (Double Asteroid Redirect Mission) will present how his team uses Docker to develop the flight software controlling the DART spacecraft. DART, a NASA Planetary Defense Mission, will hit the 160 meters wide asteroid Didymos-B at 6 km/s, providing tech and model validation for future asteroid impacters.

Chris will explain why space development is so difficult, and how his team has been able to make the software side of things easier using NASA’s Core Flight Executive, a new SpaceWire network architecture, and Docker. Using this hardware emulation layer his developer team has become more agile and enables them to develop the flight software on a regular laptop with a scalable CI/CD system. All of this saves hardware, time, bugs and money and helps software make the biggest impact it can.

Kubeflow: Cloud Native Machine Learning with Kubernetes

One type of containerized application that got very popular in 2018 is Machine Learning. David Aronchick, Product Manager, Cloud AI and co-founder of Kubeflow, Google will present Kubeflow, a Machine Learning Toolkit for Kubernetes designed to cover the whole lifecycle of ML applications on top of Kubernetes with three goals: composability, portability and scalability.

David will present the design and capabilities of Kubeflow, then explain and demo how Kubeflow’s portability and scalability goals can be achieved when deploying it on the Docker Platform, from a laptop to the cloud.

Data scientists can use Kubeflow to iterate on a model on their laptop leveraging the Kubernetes instance that comes with Docker Desktop on Mac or Windows.

System administrators can setup the production training and serving of that model with Kubeflow deployed on Kubernetes in the cloud.

Gloo function gateway, gluing together traditional apps, microservices and serverless, in hybrid deployments, on premises and in the cloud

Serverless architecture–decomposing your apps into functions responding to events, and deploying them to platforms where the servers are abstracted out from developers–is one of the most interesting trends in software this year and is beginning to see rapid enterprise adoption.

As Simon Wardley eloquently explains in tweet below,

One the most important value propositions of the Docker Enterprise Edition (EE) container platform  is choice; so it is no surprise that what we hear from our more than 500 Enterprise customers is that they are interested in serverless. However as they look at serverless solutions in a broader context they are trying to get the most flexibility and avoid lock-in.  That drives them to keep in mind what use cases are best targeted at a proprietary cloud serverless platform like AWS Lambda, while as a preference they want to be leveraging g one of the numerous portable serverless platforms that run on top of Kubernetes and Docker EE. The result is that they can build a clear hybrid cloud serverless strategy even when using Lamba.

The main theme for DockerCon 2018 is your journey to containerization, starting with modernizing traditional applications, adding microservices, then moving towards portable serverless. One of the most interesting development in that area in 2018 has been the announcement of Gloo by Idit Levine in March.

Idit Levine from will present Gloo, a high-performance, plugin-extendable, platform-agnostic function Gateway built on top of Envoy. Gloo is designed for microservice, monolithic, and serverless applications. By employing function-level routing, Gloo can completely decouple client APIs from upstream APIs at the routing level.

What Gloo, deployed on Docker EE, enables, is for enterprise developers to modernize a traditional application by containerizing it and then start to add functionalities to it using microservices. Then developers can leverage portable serverless frameworks on top of Docker EE, or AWS Lambda, all orchestrated and routed through Gloo to create a true hybrid cloud application. During her Cool Hacks session, Idit will demo such a modernization journey.

Join us at DockerCon to see these cool hacks around Space, AI and Serverless: wherever you are in your containerization journey, these sessions should inspire you to leverage containers to boldly go where no Enterprise has gone before!

Be the first to write a comment.

Leave a Reply