Elevator Pitch
Serverless platforms promised to manage apps infra. on behalf of the devs. The providers ensure scalability & availability of infra. at their own terms. In spite of quick gains, there are many players & lacks open standard. This talk addresses how to standardize serverless deployment using Knative.
Description
The evolution of Public Cloud Platforms made infrastructure provisioning and delivery of applications/services faster than ever before. Applications are traditionally deployed on compute instances and offers full control of infrastructure. However, teams are indebted to maintain/rehydrate instances in addition to paying full usage of resources in spite of low utilization. On the other hand, Cloud vendor’s serverless platform such as AWS Lambda has created huge buzz lately however it is good for trivial applications (small workloads) and often requires massive redesign for porting existing applications. Though the vendor’s serverless platform charges only for the usage of resources, it creates a vendor lock-in when build and design applications tailored made to run on the platform.
In parallel, Kubernetes (K8s) emerged as the first choice container orchestration cluster solution. It is matured, cloud-agnostic, battle-tested and planet-scale platform, born out of 15 years of Google’s experience of running production workloads. While k8s opened up an array of new possibilities of managing applications, we are yet again surprised with Knative, open source serverless add-on, run atop of K8s.
In this technical session, demonstrates application deployment simplified through Knative, abstracting physical infrastructure details from users. The k8s cluster nodes spread across cloud platforms (AWS, GKE, Azure) will be used for the demo which ensures no vendor lock-in and proves any containerized application can be deployed on a shared infrastructure yet isolated by namespace and promotes optimal resource utilization. Also, the session deep dive into Knative ecosystem and how to leverage built-in support for metric collection, log monitoring and distributed tracing in no extra time.
Notes
Mahesh Veerabathiran is a Full-stack engineer, specialized in building Microservices at scale, Big Data processing & DevOps automation. He is a Certified Kubernetes administrator (CKA), AWS developer and SysOps administrator associate. He has a depth of experience in cloud application delivery, built CI pipeline and deployment solution for the enterprise. One of the early adopters of Kubernetes in Capital One enabled 50+ event-driven big data loads and other microservices that powers credit-card decision engines. He is active in Kubernetes & Knative community, recently proposed a solution to build Enterprise Serverless platform @ Capital One in an internal hackathon that made it to Top 5 solution among 150+ shortlisted submissions (See link below).
This talk will be benefited to participants who are able to relate themselves in at least one of the following.
- Anyone who deploys services on a cloud platform or closely work with the team who does the deployment e.g. Product owner, Manager, Developer.
- Anyone who is interested in Serverless platform to increase efficiency and low maintenance but no vendor lock-in.
- Anyone who advocates/interested in a solution that leverages multi-cloud platform to achieve more resiliency as well as build and deploy services that are cloud platform agnostic.
Enterprise serverless platform | hackathon deck
https://drive.google.com/open?id=1vx0G59T_YLZoh3g1JxwWyL2NSPpMBOCp