Rationale 

The point of this project is to help people manage large scale systems.  We're not specifically talking about clusters, although much will be suitable for them.  We are talking about managing sets of clusters.  Without going insane.  Or spending tons of money on bad commercial software.

 Monitoring

 Metrics (aka. You're Blind, And You Don't Know it)

How to keep hundreds of metric points per process. Forever. Fast.

Alarming, how to keep your NOC workers from killing you in your sleep.

I Hate Me Some SNMP

Managment

The "Make it so" button

Why discovery is wrong headed

Virtualization: A solution in search of a problem 

CMDBs, and why you already have one.

Never Log Onto A Host (why you need rule based systems)

Death to Golden Samples (why you NEED Rule Based System!)

Single Sign-on and you.

Installation

How to remove a package: Reimage the host.

Software

Cassandra

Hive Server Framework

Func/Certmaster (Parallel Command Execution)

Puppet Recipe Manager (Config Mgmt)

Cobbler (Host Install)

Spacewalk (Package Mgmt.  Was RHN)