KubeCon + CloudNativeCon North America 2018

Knative, a Kubernetes native serverless frame, was just made public. Although Knative is powerful and flexible, very little is known about how to integrate, use, and extend it.

In this talk, we present a deep learning inference demo to walk through Knative end-to-end.

Specially, we explain the following concepts and process:
- How to extend Knative Eventing by showing a new Knative Eventing Source: Ceph Rados Gateway (RGW) PubSub.
- Knative Serving concepts, such as route and configuration. We illustrate these concepts by providing a Knative route that is able to invoked by RGW PubSub event and trigger a deep learning inference task.
- Knative build concepts, such as build and buildtemplate. We substantiate these concepts by a real-world source to image pipeline that builds the deep learning task.

Jupyter notebooks, deployed on top of k8s, are finding rapid adoption in universities and industry. While k8s can spawn new pods for each notebook session, launching high performance computing (HPC) jobs during dynamic workflows is complicated. Fn is a serverless solution that sits on top of k8s. Individual simulations are deployed as Fn functions that are called dynamically from workflows run within Jupyter notebooks. A notebook running on a lightweight k8s cluster can burst HPC workloads via Fn serverless calls to another cluster running on a bare metal HPC/GPU cloud. Using Fn, we have constructed a distributed access, authentication and accounting layer around a shared object store. This adds security and accountability, thereby making it easy for end users to manage complex multi-cloud workflows involving lots of different types of hardware from within a easy-to-use Jupyter notebook.

This talk will cover T-Mobile's serverless journey with the recently announced Knative project. The session will go into depth on how build and serving components of Knative project is leveraged to deliver store locator experience which consists of two functions indexing and query. Query function leverages geo distance search query capabilities provided by elastic search which is also deployed on kubernetes

Knative Serving builds on Kubernetes and Istio to provide request driven compute that can scale to 0. It supports a variety of pluggable scaling strategies and tools for managing concurrency and trade-offs between cost and latency.

This presentation reviews the challenges of designing and building a system that can rapidly scale serverless workloads to 1000's of pods. And scaling to and from 0 when idle. It presents the scaling capabilities and extension points of Knative Serving and how to configure it to take advantage of both horizontal and vertical scaling. And the concurrency controls to configure Knative Serving as a single-threaded function framework or a multi-threaded application framework (or something in-between).

Knative is an OSS project to build high-quality serverless components on Kubernetes with a "voltron effect": the pieces work separately, but come together to work even better. Jessie Zhu (Google) will facilitate a discussion on how we build and release Knative and the challenges of decoupled and pluggable interfaces.

This talk explores how a bank in Paraguay struggled with fire-fighting issues with their monolithic enterprise service bus code and embarked on a journey of digital transformation with serverless functions and Cloud Native technology. For Vision Banco SAECA Serverless with OpenFaaS gave an easy on-ramp to start making use of technology like Docker, Kubernetes and Prometheus within days through user-friendly abstractions and workflows. Patricio Diaz who leads the effort at the bank is now running two key parts of the home banking service (HomeServe) with OpenFaaS functions.

Due to the sensitive nature of the workloads - monitoring dashboards are important for the team and these were set up with the existing Prometheus data and Grafana. Given the initial success Patricio is working on moving more of the existing system functionality over to functions.