+ All Categories
Home > Software > Introduction to Typesafe ConductR

Introduction to Typesafe ConductR

Date post: 15-Jul-2015
Category:
Upload: typesafeinc
View: 566 times
Download: 1 times
Share this document with a friend
Popular Tags:
24
A Reactive Application Manager that lets Operations conveniently deploy and manage distributed systems
Transcript

A Reactive Application Manager that lets Operations

conveniently deploy and manage distributed systems

Finally, a tool to help Ops deploy

Reactive applications

3

What is Typesafe ConductR?

• A Reactive Application Manager that lets

Operations conveniently deploy and manage

distributed systems

• ConductR helps Operations orchestrate their

Reactive application deployment to a cluster,

quickly scaling in and out while ensuring

responsiveness, resilience and elasticity of the

system during the rollout.

5Go Reactive

6

7

Go Reactive 8

DevOps friendly

Go Reactive 9

How it works

DeployAutomate

Instantiate Discover

Go Reactive 10

Automate

• Works with your existing infrastructure &

provisioning automation tools:

• Amazon EC2, Docker, Puppet, Chef, Ansible, etc.

• Orchestrates deployment of application bundle

instances on provisioned nodes

• Enables rapid scaling in and out

• Ensures responsiveness, resilience and

elasticity of the system during the rollout

Go Reactive 11

• Bundles application and configuration, with

isolated config parameters

• Distributes bundle instances to wherever

resources are known to be available

• Asynchronous libraries with back-pressure

detection mechanisms

• Uses proven technologies, like HAProxy

Deploy

Go Reactive 12

Proxying removes the need for developers to care

about

the physical location of services (location transparency)

Go Reactive 13

ConductR has support for load balancing built in

through HAProxy

Go Reactive 14

• Isolated and redundant bundle instances

• Because distributed, any node can take on any

responsibility

• Multiple instances provide redundancy

• Failure remains isolated within each bundle

• Parts can fail and recover without compromising

the system as a whole

Instantiate

Go Reactive 15

• Reactive applications are loosely coupled, so

location information is not hardcoded

• Dynamic service discovery finds the service

name and provides an address

• Lives long enough to carry out the operation,

ensuring all critical services are responsive

Discover

Final thoughts

Go Reactive 16

Go Reactive 17

The 6 Benefits of ConductR

4. Designed to manage failures gracefully• Maintains responsiveness,

resilience even during failures

5. Wiser use of time and resources• Homegrown solution

defocuses core competencies

6. Improves company culture and employee retention• Encourages DevOps

approach using shared platforms and modern tools

1. Puts convenience first• ConductR focuses on

providing convenience for Operations dealing with Reactive application clusters

2. Made for microservicesand clusters

• Constant visibility and awareness of services, applications and nodes

3. Non-disruptive integration with existing tools

• Integrate with existing infra automation tools via RESTful API

Appendix

The DevOps Landscape

Go Reactive 20

Orchestration

Scheduling

Container

Provisioning / Configuration

Operating system

Virtual infrastructure

Physical infrastructure

Layer 7

Layer 6

Layer 5

Layer 4

Layer 3

Layer 2

Layer 1

ConductR, Marathon

Mesos, Kubernetes

Docker

Ansible, Chef, Puppet

Ubuntu, Linux, CoreOS

EC2

Compute, Network, Storage

Go Reactive 21

Key terms

• Bundle

• A bundle is a discrete component of your Reactive

system, such as a collection of web services built

with Play Framework

• An application is bundled as a zipped tar file

or a Docker image

• A bundle is then deployed to Linux and launched

with the CLI

Go Reactive 22

Key terms

• Container

• An additional layer of abstraction and automation of

operating-system-level virtualization on Linux

• Docker implements a high-level API to provide

lightweight containers that run processes in isolation

• Uses features of Linux to provide that level of

abstraction without starting a full-blown virtual

machine

Go Reactive 23

Key terms

• Cluster

• Nodes of machines (think IPs, whether physical or

virtual) make up a cluster of machines

• Akka Clustering is a good example, this removes the

need for Akka developers to point requests to specific

IPs, but rather tell/ask the entire cluster

• Clusters are notoriously difficult to manually configure

and inventory in production

©Typesafe 2015 – All Rights Reserved


Recommended