Cloud Native London 2019

icheikhrouhou
Nov 3 · 5 min read

News From Cloud Native London

First of all, I want to thank CloudNative London for giving me this chance to be there.

I loved the conference, The place was amazing and very organized.

There were different subjects and actually very interesting ones that fit my needs.

So following the waves of Cloud Naive solutions, Containers, Kubernetes, Observability and Cloud.

Here are some of the most trendy and also newest tech solutions and ideas presented nowadays in the Cloud and Devops communities and that were a subject for this conference.

Let’s first start by saying that all of these terms : automation, observaility, cloud native containers are actually related somehow in an Entreprise level.

For engineers that have main responsability as cloud or devops people. All of these aspects if cloud native solutions and components are necessarly something you should know about to be able to deliver a scalable and resilient in order to prepare for future business challenges.

Now let’s talk about some key points that appeared to be trendy in the Cloud Native community.

*Gitops and IAC, let’s first start by mentioning gitops that is now used by a lot of companies to successfully implement DevOps in their environment.

It’s a : Versioned CICD on top of declarative infrastructure as Kelsey Hightower said.

It’s all implemented with IAC starting from

  • Collaborative Development ( Dev & Version control ) [github or gitlab]
  • Continuous Testing ( Automation Testing : dev, test, stage) [jenkins]
  • Continuous Deployment ( Deployment automation : rollback and upgrade) , Production) [ansible, terraform, Helm]

*What is a Hybrid cloud? it’s a combination of (Public & Public Clouds) or (Public & Private Clouds) and these days a lot of entreprises follow these patterns so that they can adapt and change direction quickly.

*Vendor lock and the big dilemma of the Public and Private cloud, so many entreprises prefer staying in a private cloud scared of the idea that the public cloud’s service might go down or prices might go up or security problems but this thought of getting locked in is totally wrong since a lot of research proved that it always depends on your strategy of costs and your application needs, the prices might stay flat if you get a better management of costs and architecture.

*Serverless is the highest level of abstraction to enable focus on business value.

Why using serverless?

  • abstraction of infrastructure management
  • pay as you go
  • auto-scale
  • low maintenance
  • high availibility
  • low operational costs
  • easy deployments

*Continuous Improvement:

  • Continuous integration and testing
  • continuous delivery and deployment
  • Continuous operations

*Automating K8S with Gitops :

  • Declarative infrastructure : k8s, docker, terraform..
  • Version control your state of cluster
  • All of the config code (monitoring, dashboard) in github
  • From code change to build to updating image repo then from Git to k8s cluster.
  • Tools to use as example: Flux.

*Kubeflow:

  • Building machine learning pipeline (description of ML workflow)
  • Deployment now is simple,portable and scalable.
  • It offers a great UI and operations and an SDK for development use.

*Service Mesh :

  • service discovery and load balancing, secure service to service communication, traffic control, policies, service resilience.
  • Observability (traffic , erros ..)
  • Reliability ( Health checks, circuit breakers..)
  • Security ( ACLs, mTLS ..)

*Envoy :

  • L4 L7 service proxy
  • Traffic shadowing
  • Rate limiting
  • Access logging
  • Statistics collection.

*Istio :

  • Based on envoy proxy
  • Policy based ACL
  • Resilience, observability, traffic control
  • It offers flexible deployment options and multi cluster

*App Mesh :

  • Managed control plane
  • Buit on envoy proxy
  • Supports multiple deployment platforms (EC2, ECS, EKS..)
  • FREE to use on AWS

*Multiple mesh technologies => Super Gloo

  • Linking between two different control planes, different APIs

*Useful tools for your dev&ops cycle:

  • Forge : to build and push and apply
  • Helm : Package manager
  • Kubefwd : forwards ports
  • Telepresence : to have a local development environment for a remote Kubernetes cluster
  • Garden : makes developing and testing Kubernetes applications faster
  • BuildKit : enhancement of docker build using LLB frontends

*Continuous Delivery Build => Measure (metrics) => Learn and Analyze

*Testing in production for microservices

*Culture Patterns :

  • Learning loop
  • Avoid reinventing the wheel
  • Creative capabilities
  • Measure what matters
  • Blameless
  • Feedback

*Principle rules for high performing teams:

  • Autonomous teams
  • Lean management & product development
  • Continuous delivery
  • Psychological safety
  • Involve everyone
  • Roadmap
  • Limit WIP

*Chaos engineering : No wonder that companies’s biggest fear is the break of their system. But what if you break your own system first ?

Here it comes the chaos engineering that it’s goals are to investigate all of the weaknesses of your system and collect informations that you probably didn’t know about in your system in order to create a full plan for recovery to your system if it breaks and to maintain your system’s availability.

We can start by doing chaos by building hypothesis of a behaviour and then through:

  • Game days
  • Automated experiments

Finally, I want to say that it was a great conference when I learned a lot of what’s going on in the Cloud Native world and I will totally come back again.

And for the end :

Thanks Cloud Native London.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade