Deployment to large server farms using Fabric and BitTorrent.

In this day and age of web services, there are applications that run out of huge server farms having hundreds or even thousands of servers.

Deployment to such a set of thousands of servers can be time consuming and resource intensive.

The aim of this talk is to show how Python can be leveraged for configuration management, machine management and extend it with P2P technologies like BitTorrent, operating in tandem with Fabric to update servers in a huge server farm with minimal time, effort and resource.

Deploying to servers in a large server farm, possibly geographically distributed, can be time consuming.

Centralized deployment options do not scale in such scenarios and for businesses, time is money. Using the peer-to-peer BitTorrent file sharing system, deployment to a large number servers can be executed in a decentralized manner resulting in a steep decrease in the time and resources spent.

BitTorrent will become the standard for updating servers in such huge server farms and using Fabric and the power of Python is the way to go.

1) Introduction to Fabric. 2) BitTorrent for distributing files in a production environment. 3) Fabric configuration and BitTorrent tracker, seeders and peers setup. 4) Demonstration of the setup. 5) Real world usage examples of similar systems ( Twitter, Facebook server farms deployment)

Shyam Sundar CS is a senior build and deployment consultant at ThoughtWorks. He has used Python extensively for system administration, build and deployment automation.

Manoj S Mahalingam has been a build and release engineer for the last one year at ThoughtWorks Technologies (India) Pvt Ltd. He is experienced in developing with Django, Juno and TurboGears. A PyS60 aficionado and has developed multitude of applications for Symbian smartphones.