San Francisco is one of my favorite places in the world, with the constant cool breeze, tangible curiosity in the air, and a coffee shop at every corner. As I finish up my first day at DockerCon 2018, here are highlights from a few talks that particularly stood out and have me thinking about containers with a fresh perspective.
DockerCon General Session — Steve Singh, Docker CEO, and guests
I started the day off by joining the bustling crowd of people packing into the general session. Steve Singh, CEO of Docker, kicked off the event and was joined by multiple guests throughout the session. Steve and his guests consistently reiterated two major points: choice and agility; that is, the freedom to choose the technology stack you want to work with and the agility to deliver at speed. There’s no doubt that these are two areas where the Docker engine excels, and it’s great to hear that these priorities are aligned with their other products.
The general session demos focused on new capabilities in Docker for Desktop and Docker EE. I find it interesting to note that there were two Pet Store demos during the session; for those unaware, the Pet Store application was hugely popular more than a decade ago, when it served as a standard sample for Java-based enterprise applications. Modernizing legacy applications is a narrative that has clearly taken high priority at Docker.
It was also quite amusing that before guests were invited to the stage, they were first required to perform a ritual to the demo gods by cracking open a fortune cookie. Considering how well the demos went, the demo gods must have appreciated the sacrifice!
Build Secure Apps with Docker and Kubernetes — Jason McGee, IBM
My first session for the day set expectations high … and completely delivered. It was presented by Jason McGee, an IBM Fellow. Jason started right away with a powerful use-case: machine learning. He demoed a machine learning platform called Fabric for Deep Learning (FfDL, pronounced “fiddle”). It’s essentially a bundle of machine learning utilities such as TensorFlow and PyTorch available as a service on Kubernetes.
In a matter of minutes, Jason was able to upload a set of training data to the platform (running on IBM Kubernetes Service) and use that model to classify images of clothing. With the power of GPU-enabled machines on IKS, the training process was much quicker than CPU-based machines. To get started with FfDL, check out the in-depth code pattern on IBM Code.
Jason also announced a new capability that was recently pushed to IBM Kubernetes Service: the ability to have multi-zone Kubernetes cluster deployments. Essentially, this is the ability to spin up and manage multiple worker nodes on different data centers within the same IBM Cloud region, all networked together under the same cluster. This enables better fault tolerance, leading to higher availability for your applications. To learn more, check out the announcement on the IBM Cloud blog.
5 Patterns for Success for Application Transformation — Elton Stoneman, Docker
Another talk I want to highlight was clearly a popular topic, with over 700 people packed into the room to listen to Elton Stoneman highlight key patterns for transforming legacy applications for the modern cloud. He was serious about the veracity of his legacy apps: the sample apps he developed for the demos were based on .NET and Java versions from 10 years back. Major props to him for reviving an ancient tech stack for the demos!
Elton outlined the five major pieces of every application that need to be changed to “make old apps work like new apps”:
For each of these categories, Elton showed the exact code changes required to make the legacy apps run like modern Docker-based applications. To learn more, check out the videos Modernizing .NET Apps for Developers and Modernizing Traditional Java Apps for IT Pros.
IBM at DockerCon
By the way, IBM is a platinum sponsor of DockerCon! Check out our booth; you can’t miss us, we’re right in the front. If you haven’t come by yet, feel free to stop and chat with one of us about anything Docker-, cloud-, or IBM-related! We’re super excited about container technology; to prove it, check out the awesome code patterns that talented IBM developers have created.
Day 1 at DockerCon was a great success! Be sure to check back tomorrow and I’ll convey my thoughts about day 2.